---
title: "How to Build a Scalable B2B Content Engine with AI (Without Losing the Fundamentals)"
episode: 61
podcast: "The LeanScale Podcast"
publisher: "LeanScale"
guest: "Benjamin Hoehn"
guest_title: "Head of Marketing, LightSource"
date_published: 2026-05-15
date_modified: 2026-07-22
duration: 00:32:20
word_count: 5964
topics: ["demand-generation", "ai-in-gtm", "gtm-strategy", "brand-positioning", "enterprise-sales"]
canonical_url: https://leanscale-knowledge-hub.netlify.app/podcast/benjamin-hoehn-b2b-content-engine/
source: "LeanScale Podcast Knowledge Hub — https://leanscale-knowledge-hub.netlify.app"
license: "Free to quote and cite with attribution to The LeanScale Podcast."
---

# How to Build a Scalable B2B Content Engine with AI (Without Losing the Fundamentals)

_Benjamin Hoehn on ICP-first content, founder-led marketing, and using AI to accelerate — not replace — the fundamentals_

**Episode 61 · The LeanScale Podcast**  
Benjamin Hoehn, Head of Marketing, LightSource · Hosted by Anthony Enrico  
Published May 15, 2026 · Updated July 22, 2026 · 00:32:20  
Canonical: https://leanscale-knowledge-hub.netlify.app/podcast/benjamin-hoehn-b2b-content-engine/

**Topics:** Demand Generation · AI in GTM · GTM Strategy · Brand & Positioning · Enterprise & Public-Sector Sales


## Executive summary

Everyone wants to point AI at their marketing and print pipeline. Benjamin Hoehn — a developer-turned-marketer now Head of Marketing at LightSource, an AI-native sourcing and procurement platform — thinks most B2B teams are skipping the step that makes AI worth anything: the fundamentals. In this episode with LeanScale co-founder Anthony Enrico, Ben lays out a no-hype, tactical blueprint for building a content engine that actually scales, and where AI belongs in it (at the end, as an accelerant, not at the start, as a substitute for knowing your customer).

Ben's origin story is the whole thesis in miniature. He joined LightSource openly admitting he knew SaaS and startups but nothing about procurement. Instead of guessing, he spent his first 30–45 days doing three to five 'buy-you-a-coffee' calls a week — manual LinkedIn outreach to real procurement people, no pitching — then fed those transcripts, plus analyst research, into a Google NotebookLM to distill a full ICP and persona. Only then did he build messaging, positioning, and even revise product vision on top of that bedrock. His warning: you can't automate on a bad foundation, and most companies who think their ICP is 'dialed in' are nowhere near deep enough.

The middle of the conversation is a clinic on the fundamentals themselves. Ben maps the real B2B buying committee — 11 to 15 people in a mid-market-to-enterprise deal: the executive sponsor, the director-level champion tasked with evaluating vendors, the end users, and the technical implementers who hold veto power over your integration. You have to develop content for every one of them, then read website intent signals to see which personas are actually resonating. He walks through his highest-leverage content play: founder-led 'admiration content,' where he researches a target account's public supply-chain data, the CEO refines it in his own voice and posts it, and the AE references it live in the deal — a virtuous cycle that once pulled an inbound from SharkNinja mid-draft.

On the topics operators obsess over, Ben is refreshingly blunt. Attribution: stop spinning your wheels — once it hits pipe, interview the reps, have RevOps build a model everyone agrees on, and just ask customers 'how did you hear about us?' (asked multiple times across the form, disco, and demo, it becomes a genuine multi-touch signal). LinkedIn: post a few times a week, trendjack through the ICP lens (his Taylor Swift–Travis Kelce engagement post reframed through procurement — 'how could he not at least do three bids and a buy?' — landed with sourcing pros), and lean into authentic, almost-Facebook-style content as the noise rises.

Finally, the AI stack: three 'bedrock' NotebookLM knowledge centers (competitive research on ~90 competitors, product/release/analyst docs, and ICP/persona plus Gong pain-point calls), and a pain-point-to-content waterfall that crosses 82 personas by ~50 topics into hundreds of monitored prompts, each seeding a blog post or white paper. The throughline Anthony draws out: start with the foundation, understand whose pain you solve, build content that speaks to them — and then, and only then, use AI to accelerate and ideate. Who should listen: B2B SaaS and AI-startup marketers, founders doing founder-led sales, and RevOps leaders building attribution and intent instrumentation.


## Key takeaways

1. **Do the fundamentals before you touch AI** — A year and a half ago the advice was 'start experimenting'; now, if you're not using AI you're behind. But before you feed anything into AI, you still have to do the old-school work of figuring out who your audience is and nailing your ICP. AI on a bad foundation just scales bad marketing faster.
   _Why it matters:_ Sequence matters: ICP and persona research first, AI acceleration second. Teams that invert this produce more content, not better content, and erode trust with the exact buyers they're trying to win.
   _For:_ Marketing Leaders, Founders, RevOps Leaders

2. **Learn a market you don't come from by talking to humans, not pitching** — Ben joined LightSource knowing nothing about procurement, so he ran 3–5 non-pitch 'buy-you-a-coffee' calls a week for 30–45 days via manual LinkedIn outreach — asking what makes people tick, why they got into the field, what they value. People happily shared, because breaking into a market with bad, impersonal marketing is exactly what they don't want to see.
   _Why it matters:_ Manual, unscalable discovery is the raw material AI can't manufacture. The transcripts of real practitioner conversations become the training data for every persona, message, and piece of content that follows.
   _For:_ Marketing Leaders, Founders

3. **Map the entire buying committee — all 11 to 15 of them** — A mid-market-to-enterprise deal typically has 11–15 people: an executive sponsor, a director-level champion who researches and evaluates vendors, the end users, and the technical implementers. Implementers hold veto power — if your tech doesn't integrate with the stack, they boot you out — so you need content and messaging for every persona in the room.
   _Why it matters:_ 'We go after Fortune 500' is a vertical, not a target. Meaningful messaging requires knowing each role in the deal and building materials for each, then watching which personas show up on which pages.
   _For:_ Sales Leaders, Marketing Leaders

4. **ICP depth is a spectrum — go as deep as true one-to-one when it pays off** — The shallowest acceptable depth is knowing the main persona in the buying committee. The deepest, now achievable with tools like Clay, is genuinely one-to-one: identify a website visitor, enrich them, pull public psychographics (even their Instagram), and send a personalized message. It takes rigging, but it's possible.
   _Why it matters:_ Most companies that think their ICP is 'dialed in' are operating several layers too shallow. Deciding deliberately where you sit on the depth spectrum determines how personalized — and how effective — your outreach can be.
   _For:_ Marketing Leaders, RevOps Leaders

5. **Founder-led 'admiration content' creates a virtuous cycle** — Ben researches a target account's public supply-chain data, drafts an insight-rich breakdown, the CEO refines it in his own voice and posts it, and the AE then references it live in the deal ('we just did some research on you guys'). Once, mid-draft on a SharkNinja piece, the company sent an inbound — and the AE could point to the admiration content to accelerate the cycle.
   _Why it matters:_ In Series A/B founder-led sales, the founder still has to market. Turning genuine account research into founder voice, posted 2–3 times a week, compounds brand and warms specific deals at the same time.
   _For:_ Founders, Marketing Leaders

6. **Stop over-engineering attribution — get to pipe and just ask** — Attribution is an argument you can spin your wheels on endlessly. Most leaders now care less about exact source than about pipe and rev. MQLs and early signals are near-vanity metrics for judging campaign health; once a deal hits pipe, interview the reps, have RevOps build a model everyone agrees on, and report at that level.
   _Why it matters:_ Time spent adding algorithmic permutations to an attribution model is time not spent getting more deals. Keep it simple (first touch / last touch) for sales-led growth and reinvest the energy into better content and ICP work.
   _For:_ RevOps Leaders, Marketing Leaders

7. **'How did you hear about us?' — asked multiple times — is real multi-touch attribution** — AI makes unstructured self-reported data usable. Ask the question on the request form, in the disco script, and again in the first or second demo. Because the champion may have found you on a podcast while another stakeholder came through a different channel, asking repeatedly reveals distinct touchpoints per persona — telling you where to double down.
   _Why it matters:_ A cheap, human question outperforms elaborate models for most B2B teams. Instrument it across the funnel and let the pattern of answers, not an algorithm, guide channel investment.
   _For:_ RevOps Leaders, Sales Leaders

8. **Start your content engine where your ICP congregates, and monitor it** — Persona research should surface the influencers and communities your buyers gather in. At LightSource, ~50% of the persona was on Reddit, so Ben runs a weekly deep Perplexity report on it to surface trending topics, complaints, and questions — which become blog FAQs and LinkedIn posts. Founder-led podcasts and influencer engagement are the easiest, most measurable ways to scale from Series A to B.
   _Why it matters:_ Don't guess at channels. Let the ICP dictate them, then stand up lightweight monitoring so community pain becomes a renewable content supply rather than a one-off brainstorm.
   _For:_ Marketing Leaders, Founders

9. **Win LinkedIn by trendjacking through the ICP lens** — Ben posts a few times a week with the founder and internal SMEs, and runs cultural moments through his persona notebook before reacting. For the Taylor Swift–Travis Kelce engagement, instead of a generic post he asked 'what would a procurement person think?' — breaking down the ring's bill of materials and joking 'how could he not at least do three bids and a buy?' Sourcing pros called it the perfect post.
   _Why it matters:_ Generic trendjacking gets ignored; trendjacking filtered through a persona bedrock resonates deeply with a narrow audience. The persona notebook lets you produce that in ~20 minutes.
   _For:_ Marketing Leaders, Founders

10. **Build three 'bedrock' knowledge centers and run a pain-point-to-content waterfall** — Ben runs three NotebookLM knowledge centers: competitive research (~90 competitors' value props), product/release notes/analyst reports, and ICP/persona plus Gong pain-point calls. From there he mines pain points, converts them into topics, crosses ~82 personas by ~50 topics into hundreds of monitored prompts, and each prompt seeds a blog post or white paper.
   _Why it matters:_ Curated LLM knowledge centers turn scattered research and call data into an on-demand content factory — but every output still needs human refinement before it ships. As an 'army of one,' this is how a lean team produces enterprise-grade assets like a full competitive matrix overnight.
   _For:_ Marketing Leaders, RevOps Leaders

11. **AI accelerates, humans finish — do the fundamentals, then abstract into delight** — The workflow is foundation first (know whose pain you solve), then AI to accelerate and ideate — and every AI output gets a human pass into readable copy. Once the blocking-and-tackling is done and you know what's working, you 'abstract': tap into B2C-style, distinctive, delightful campaigns to break through the noise (Ben's team ran one and was the most popular booth at an event).
   _Why it matters:_ AI is the accelerant on top of fundamentals, not a shortcut past them. The endgame isn't more content — it's earning the right to do something authentic and distinctive that actually cuts through.
   _For:_ Marketing Leaders, Founders


## Frameworks

### Fundamentals Before AI (01:15)

**Definition:** Before feeding anything into AI, do the old-school work of identifying your audience and nailing your ICP and personas; AI is an accelerant layered on top of that bedrock, never a substitute for it.

Ben's core thesis: automating on a bad foundation just scales bad marketing. The sequence is ICP/persona research first, then messaging and positioning, then AI to accelerate — invert it and you get more content, not better content.

### The 11–15 Person B2B Buying Committee (05:09)

**Definition:** A mid-market-to-enterprise deal has 11–15 stakeholders: the executive sponsor (C-suite), the director-level champion who discovers and evaluates vendors, the end users, and the technical implementers who hold integration veto power. You must speak to every one.

Identifying the buying group and each role's behavior is the key output of persona research. It also tells you which content to build and lets you read intent signals — e.g., IT hitting the integrations page vs. an exec attending a C-suite webinar.

### The ICP Depth Spectrum (04:32)

**Definition:** A range from shallow (know the main persona in the buying committee) to deep (true one-to-one: identify a website visitor, enrich them in Clay, pull public psychographics, and send a personalized message).

Ben argues you can almost never go too deep, and most companies that think their ICP is dialed in are several layers too shallow. Choosing where you sit on the spectrum sets the ceiling on how personalized your marketing can be.

### Founder-Led Admiration Content (The Virtuous Cycle) (08:38)

**Definition:** Research a target account's public data, draft an insight-rich 'we admire you' breakdown, have the founder refine it in their voice and post it, then let the AE reference it live in the deal to accelerate the cycle.

It compounds brand and warms specific deals simultaneously. The SharkNinja story — an inbound arriving mid-draft — shows the play in action; do it 2–3 times a week to sustain founder-led scale from Series A to B.

### Multi-Ask 'How Did You Hear About Us?' Attribution (14:37)

**Definition:** Rather than over-engineering an attribution model, ask 'how did you hear about us?' on the request form, in the disco script, and again in the first or second demo — capturing self-reported touchpoints across the funnel.

Because different personas in the same deal found you through different channels, asking multiple times becomes a genuine multi-touch signal that tells you where to double down. AI makes the unstructured answers easy to synthesize.

### Three Bedrock Knowledge Centers (25:23)

**Definition:** Three curated NotebookLM knowledge bases: (1) competitive research (~90 competitors' value props), (2) product docs, release notes, and analyst reports, and (3) ICP/persona research plus Gong pain-point call transcripts.

Each acts as a mini knowledge center you can query for a deep dive on any topic — from an overnight competitive matrix to persona-specific messaging — provided a human refines every output before it ships.

### The Pain-Point → Content Waterfall (26:36)

**Definition:** Pull close-won/close-lost call transcripts into an LLM, extract pain points, convert them to topics, cross ~82 personas by ~50 topics into hundreds of monitored prompts, and let each prompt seed a blog post, white paper, or ad/email copy.

It turns real customer language into a renewable, industry- and persona-sliced content supply. Ben feeds pain points directly into web and email copy, with a human pass to make it readable, and monitors visibility across every prompt permutation.

### Fundamentals, Then Abstract (30:32)

**Definition:** First do the blocking-and-tackling (know your customer, produce content, know your channels and measurement); only then 'abstract' into distinctive, delightful, B2C-style campaigns that break through the noise.

As the market gets noisier, authenticity and distinctiveness win — but you earn the right to run a delight-driven contest or campaign only after the foundation is in place. Ben's team ran one and became the most popular booth at an event.


## Quotes

_Speakers inferred from an undiarized transcript — verify before attributing._

> "Before you begin marketing, before you start feeding anything into AI, you really do need to still do the fundamentals of figuring out who your audience is and figuring out your ICP."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (01:15)

> "I know marketing, I know SaaS, I know startups. I don't know anything about procurement. So I did a lot of manual outreach — I literally just need to talk to people."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (01:56)

> "People, especially if you're breaking into a new market, they don't want to see bad marketing. They wanted to see someone actually trying to get to know them."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (02:57)

> "You can't just start automating things on a bad foundation."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 61 (03:30)

> "There are like 11 to 15 people typically in a deal, mid-market to enterprise. So what you really need to do is find out who's in that room, who you're talking with, and what their role is in that room."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (05:09)

> "The implementers have veto powers. If your tech doesn't play nice or integrate with the current tech stack, they're going to boot you out. So you need to be able to speak to every single one of them."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (06:08)

> "This stuff is super nerdy and kind of boring for anyone outside of supply chain and procurement, but it's gold for our ICP."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (09:02)

> "As I was drafting that article, we actually got an inbound from Shark Ninja. It created a virtuous cycle."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (10:06)

> "Attribution is one of these arguments that you can get into and just spin your wheels with everyone almost endlessly."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (11:55)

> "Things like MQLs and searching for super early signal, they're almost vanity metrics that marketers can use internally to just judge the health of their campaigns."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (12:30)

> "If you can ask 'how did you hear about us?' multiple times, that is actually a pretty effective form of attribution."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (15:15)

> "He only went to one supplier. How could he not at least do three bids and a buy? It was love through the lens of procurement."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (22:21)

> "Trendjacking is back. It's 2009 all over again."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 61 (23:30)

> "I'll put something really thoughtful, do a lot of research, and then I post a picture of our offsite and that gets like 10 times the engagement."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 61 (24:09)

> "If you're not pulling down those call transcripts and feeding them into a mini LLM, do that tomorrow."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (28:06)

> "Start with the foundation, understand who you're solving a pain for in the first place — and then, and only then, use AI to accelerate it and help with the ideation."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 61 (29:22)

> "These are the blocking and tackling pieces of scaling content. Once you have these things done, then you get to abstract."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (30:32)

> "The noise is out there so much that you have to be authentic and do something really distinctive to make an impact."
>
> — Benjamin Hoehn, The LeanScale Podcast Ep. 61 (31:10)


## Practical advice by role

### Marketing Leaders

- Do the fundamentals before AI: run manual, non-pitch discovery calls with real buyers, feed the transcripts (plus analyst research) into a NotebookLM, and build your ICP, personas, messaging, and positioning on that bedrock first.
- Build content for every persona in the 11–15-person buying committee — sponsor, champion, end users, and veto-holding implementers — then read website intent signals to see which personas are actually resonating.
- Stand up three knowledge centers (competitive, product/analyst, ICP + Gong pain points) and run a pain-point-to-content waterfall, but give every AI output a human pass before it ships.
- Trendjack through the ICP lens: run cultural moments through your persona notebook and post the version your narrow audience will love, not the generic reaction everyone else posts.

### Founders

- In founder-led sales, you still have to market — publish curated, in-your-voice 'admiration content' about target accounts 2–3 times a week to compound brand and warm specific deals at once.
- Start your content engine where your ICP congregates (for LightSource, ~50% on Reddit) and lean into founder-led podcasts and influencer engagement — the easiest, most measurable way to scale from Series A to B.
- Experiment with authentic, almost-Facebook-style content and, once fundamentals are in place, abstract into a distinctive, delightful campaign to break through the noise.

### RevOps Leaders

- Don't over-engineer attribution — once a deal hits pipe, interview the reps and build a model everyone agrees on rather than adding algorithmic permutations; keep first-touch/last-touch simple for sales-led growth.
- Instrument 'how did you hear about us?' on the request form, the disco script, and the first or second demo — asked repeatedly, it surfaces distinct per-persona touchpoints and doubles as multi-touch attribution.
- Stand up website intent instrumentation (visitor de-anonymization plus analytics and enrichment) so you can see who's arriving from published content almost overnight and which personas they map to.

### Sales Leaders

- Know every role in the deal room — champion, sponsor, end users, and the implementers whose integration veto can kill you — and equip reps to speak to each.
- Use founder admiration content in live deals: have the AE reference the research ('we just did a breakdown of your model') to build rapport and accelerate the cycle.
- Ask buyers how they found you in disco and demos; the answers reveal which channels created the deal and where to concentrate.


## AI takeaways

**Thesis:** AI is an accelerant you layer on top of marketing fundamentals — not a shortcut past them. Do the ICP, persona, and buying-committee work first; then use AI (NotebookLM knowledge centers, pain-point mining from call data, enrichment) to scale content and ideation. Automate on a bad foundation and you just scale bad marketing.

- **Fundamentals first, AI second** — Before feeding anything into AI, nail your audience and ICP. AI on a bad foundation produces more content, not better content — and erodes trust with the buyers you're trying to win.
- **Turn call data into content** — Pull close-won/close-lost transcripts from Gong or Fathom into a mini-LLM, extract pain points, and cross personas by topics into hundreds of prompts — a renewable, persona-sliced content supply.
- **Curated knowledge centers beat one-off prompts** — Three NotebookLM bases (competitive, product/analyst, ICP + pain-point calls) let an 'army of one' produce enterprise assets like a full competitive matrix overnight.
- **AI makes messy data usable** — Self-reported 'how did you hear about us?' answers become real multi-touch attribution once AI can synthesize the unstructured responses across form, disco, and demo.
- **Humans still finish the work** — Every AI output needs a human pass into readable, on-brand copy; AI accelerates the fundamentals and ideation but doesn't replace judgment, voice, or the manual discovery that seeds it all.

**Agent & automation ideas**

- A weekly community-listening agent that runs a Perplexity/Reddit sweep of your ICP's channels and outputs trending topics, complaints, and questions as ready-to-use blog FAQs and social hooks.
- A pain-point-to-content pipeline: ingest call recorder transcripts, extract and cluster pain points by persona and industry, then draft persona-specific copy for human refinement.
- An intent-signal router that watches website de-anonymization (RB2B) plus analytics, maps visitors to buying-committee personas, and flags which content is pulling which persona.
- A trendjack agent that runs a cultural moment through the persona notebook and drafts the ICP-lens version of the post before the moment passes.


## Operations takeaways

### Revenue operations

- **Don't over-engineer attribution.** Once a deal hits pipe, interview reps and build one model everyone agrees on; adding algorithmic permutations is time not spent winning deals.
- **Instrument the 'how did you hear' question.** Ask it on the form, in disco, and in demos — repeated asks surface distinct per-persona touchpoints and function as multi-touch attribution.
- **Stand up intent instrumentation.** Visitor de-anonymization plus analytics and enrichment lets you see who arrives from published content almost overnight and which personas they map to.
- **Keep it simple for sales-led growth.** First-touch/last-touch at the deal or opportunity level beats elaborate models for most B2B teams that aren't drowning in deals.

### Pipeline & marketing ops

- **Content warms specific deals.** Founder admiration content can pull inbound (SharkNinja) and gives the AE a live rapport-builder to reference in the cycle.
- **Read intent by persona.** Watch which buying-committee roles hit which pages — IT on integrations, execs at the C-suite webinar — to see what's resonating.
- **Vanity vs. real signal.** MQLs and early signals judge campaign health internally; click-through and conversion by campaign/channel tell you where to double down or cut.
- **Community as a pipeline source.** Monitoring where the ICP congregates (Reddit, influencers) turns real buyer pain into a renewable topic supply that feeds inbound.


## Metrics mentioned

| Value | Metric | Context |
| --- | --- | --- |
| 3–5 calls/week for 30–45 days | Discovery cadence to learn a new market | Manual, LinkedIn-sourced 'buy-you-a-coffee' conversations Ben used to learn procurement before writing any marketing. |
| 11–15 people | B2B buying committee size | Typical mid-market-to-enterprise deal, higher on the enterprise side — every persona in the room needs content. |
| 2–3x per week | Founder content cadence | How often the founder should publish curated, in-his-voice admiration articles/posts to sustain founder-led scale. |
| ~90 | Competitors tracked in one notebook | A single NotebookLM knowledge center holding ~90 competitors' value propositions for instant competitive deep-dives and an overnight competitive matrix. |
| 82 personas × ~50 topics → hundreds of prompts | Content permutation engine | Pain points converted to topics crossed with personas produce hundreds of monitored prompts, each seeding a blog post or white paper. |
| ~50% of persona | Community coverage on Reddit | Roughly half the ICP congregates on Reddit, which Ben monitors weekly with a Perplexity report. |
| ~20 minutes of research | Trendjack turnaround | The Taylor Swift–Travis Kelce procurement post took ~20 minutes because the persona notebook already existed. |


## Entities mentioned

- **LightSource** (company) — Ben's employer; an AI-native sourcing and procurement platform where he heads marketing. He joined knowing SaaS and startups but nothing about procurement, and rebuilt the ICP, personas, messaging, and product vision from manual discovery. · https://leanscale-knowledge-hub.netlify.app/company/lightsource/
- **SharkNinja** (company) — A consumer-products company Ben was drafting an admiration/breakdown article about (fascinated by their influencer-led, user-base go-to-market) when SharkNinja sent an inbound — his proof of the founder-content virtuous cycle. · https://leanscale-knowledge-hub.netlify.app/company/sharkninja/
- **Retention.com** (company) — Cited via the RB2B founder's story: he says he 'scaled the wrong way' at Retention.com by hiring ~50 people, versus growing RB2B through podcasts, influencers, and founder-led content. · https://leanscale-knowledge-hub.netlify.app/company/retention-com/
- **Benjamin Hoehn** (person, guest) — Head of Marketing at LightSource; a developer-turned-marketer who builds ICP-first B2B content engines and accelerates them with AI. · https://leanscale-knowledge-hub.netlify.app/guest/benjamin-hoehn/
- **Anthony Enrico** (person, host) — Co-founder of LeanScale and host of The LeanScale Podcast. · https://leanscale-knowledge-hub.netlify.app/guest/anthony-enrico/
- **Google NotebookLM** (tool, AI Research / Knowledge) — Ben's core 'bedrock' knowledge centers (referenced as Google/Gemini notebook). He runs three — competitive research, product/analyst docs, and ICP/persona plus Gong pain-point calls — and pulls deep dives on demand; also where he distilled discovery-call transcripts into a full procurement persona.
- **Clay** (tool, GTM Data / Enrichment) — Cited as how you reach true one-to-one B2B messaging — identify a website visitor, enrich them, pull public psychographics, and send a personalized message.
- **Gong** (tool, Revenue Intelligence) — Named (with Fathom) as the call recorder whose close-won/close-lost pain-point transcripts you should feed into a mini-LLM; a core input to Ben's ICP/persona knowledge center.
- **Fathom** (tool, Call Recorder / Notetaker) — Named alongside Gong as a call recorder — pull the transcripts and feed them into an LLM to mine customer pain points.
- **Perplexity** (tool, AI Search / Research) — Ben runs a weekly deep Perplexity report on Reddit to surface trending procurement topics, complaints, and questions, which become blog FAQs and social posts.
- **Reddit** (tool, Community / Social) — Roughly 50% of Ben's persona congregates on Reddit, so he monitors it weekly (via Perplexity) as a content-sourcing channel.
- **RB2B** (tool, Website Visitor Identification) — Website-visitor de-anonymization tool both Ben and Anthony use to see who's arriving from published content almost overnight; paired with Google Analytics and enrichment. Also the subject of the founder's content-vs-headcount scaling story.
- **Google Analytics** (tool, Web Analytics) — Paired with RB2B and enrichment to instantly gauge inbound traffic and which personas are arriving from newly published content.
- **LinkedIn** (tool, Social Platform) — The main B2B social channel; Ben is spinning up a founder/SME posting program (a few times a week) and runs the 'trendjack through the ICP lens' play here.
- **Instagram** (tool, Social Platform) — Used in the deepest one-to-one play — enrich a website visitor, check their public Instagram for psychographics, then send a personalized message.


## FAQ

**Q: Should you build your ICP before using AI for marketing?**

A: Yes. Benjamin Hoehn's core rule is fundamentals before AI: identify your audience and nail your ICP and personas first, then layer AI on top to accelerate. Automating on a bad foundation just scales bad marketing faster and erodes trust with the buyers you're trying to win.

**Q: How many people are in a B2B buying committee?**

A: Typically 11 to 15 in a mid-market-to-enterprise deal, and often more at the enterprise level. The group includes a C-suite executive sponsor, a director-level champion who evaluates vendors, the end users, and the technical implementers — who hold veto power if your product won't integrate with the existing stack. You need content for every one of them.

**Q: How do you learn a market you don't come from?**

A: Talk to real practitioners before you market to them. Hoehn ran three to five non-pitch 'buy-you-a-coffee' calls a week for 30–45 days via manual LinkedIn outreach, then fed the transcripts and analyst research into a NotebookLM to distill a full ICP and persona. The manual, unscalable discovery is the raw material AI can't manufacture.

**Q: How should you handle content attribution when the data isn't clean?**

A: Stop over-engineering it. Once a deal reaches pipe, interview the reps and have RevOps build one model everyone agrees on rather than adding algorithmic permutations. Treat MQLs and early signals as internal health checks, and simply ask buyers 'how did you hear about us?' on the form, in disco, and in demos — asked repeatedly, it becomes real multi-touch attribution.

**Q: Where should an early-stage B2B company start its content engine?**

A: Where your ICP already congregates. Persona research should surface the communities and influencers your buyers follow; monitor those channels (Hoehn runs a weekly Perplexity report on Reddit, where ~50% of his persona is) and turn the pain points into content. Founder-led podcasts and influencer engagement are the easiest, most measurable way to scale from Series A to B.

**Q: What is founder-led admiration content and why does it work?**

A: It's content where marketing researches a target account's public data, drafts an insight-rich 'we admire you' breakdown, the founder refines it in their own voice and posts it, and the AE references it live in the deal. It compounds brand and warms specific deals at once — Hoehn's team once received an inbound from SharkNinja while still drafting a piece about them.

**Q: How do you use AI to scale B2B content without losing quality?**

A: Build curated LLM knowledge centers — Hoehn runs three in NotebookLM (competitive research, product/analyst docs, and ICP/persona plus Gong pain-point calls) — then mine pain points, cross personas by topics into hundreds of prompts, and let each seed an asset. Critically, every AI output gets a human pass into readable, on-brand copy before it ships.

**Q: What is trendjacking through the ICP lens?**

A: Instead of reacting to a cultural moment with a generic post, you run it through your persona knowledge base and publish the version your narrow audience will love. Hoehn reframed the Taylor Swift–Travis Kelce engagement through procurement — breaking down the ring's bill of materials and joking about doing 'three bids and a buy' — and sourcing professionals called it the perfect post.


## Timeline

- **00:00** — Intro: a developer-turned-marketer at LightSource
- **01:15** — Start with fundamentals before AI: nail your ICP
- **01:56** — Learning procurement: manual LinkedIn 'coffee' calls
- **04:32** — How deep is deep enough on ICP?
- **05:09** — Mapping the 11–15 person buying committee
- **07:55** — Building a content engine on top of ICP work
- **08:38** — Founder-led admiration content (the SharkNinja inbound)
- **11:10** — Content attribution without clean data
- **14:37** — Just ask 'how did you hear about us?' — multiple times
- **15:54** — Where to start a content engine: channels and timelines
- **16:37** — Monitoring communities: Reddit + Perplexity
- **19:36** — Optimizing LinkedIn for founder-led sales
- **21:23** — Trendjacking through the ICP lens
- **23:30** — Authentic vs. engineered content
- **25:23** — The AI stack: three NotebookLM knowledge centers
- **26:36** — Turning pain points into content at scale
- **29:22** — Foundation first, then AI to accelerate
- **30:32** — Closing: abstract into delight; where to find Ben


## Related episodes

- **Ep. 93: How HubSpot Built a Media Empire** (Jonathan Hunt) — The owned-media, treat-content-like-a-media-property companion to Ben's scalable content engine. · https://leanscale-knowledge-hub.netlify.app/podcast/jonathan-hunt-hubspot-media-empire/
- **Ep. 90: Why Your Niche Isn't Niche Enough** (Gary Frazier) — The ICP/positioning foundation that must exist before AI-accelerated content — 'Fortune 500' is a vertical, not a target. · https://leanscale-knowledge-hub.netlify.app/podcast/gary-frazier-niche-brand-positioning/
- **Ep. 33: Why AI Won't Save Your Bad Brand** (Mario Paganini) — The sibling 'fundamentals before AI' argument applied to brand — automation can't fix a weak foundation. · https://leanscale-knowledge-hub.netlify.app/podcast/mario-paganini-ai-wont-save-brand/
- **Ep. 1: Why Is Multi-Touch Attribution So Hard? And Is Anyone Actually Doing It?** (Bernardo Alves) — Deep counterpart to Ben's 'stop over-engineering attribution, just ask' stance on content ROI. · https://leanscale-knowledge-hub.netlify.app/podcast/bernardo-alves-multi-touch-attribution/
- **Ep. 46: Why Your GTM Strategy Is Broken—And How AI Can Fix It** (Neel Kamal) — The broader AI-plus-GTM-strategy frame around Ben's content-engine tactics. · https://leanscale-knowledge-hub.netlify.app/podcast/neel-kamal-gtm-strategy-ai/
- **Ep. 18: Our Fastest Growing Customers are Measuring These 3 Marketing Metrics** (Bernardo Alves) — Pairs with the vanity-metrics vs. pipe-and-rev discussion on what marketing should actually measure. · https://leanscale-knowledge-hub.netlify.app/podcast/bernardo-3-marketing-metrics/


## Full transcript

_Machine-transcribed and not diarized; speaker attribution is inferred._  
_Transcript only, as a separate file: https://leanscale-knowledge-hub.netlify.app/podcast/benjamin-hoehn-b2b-content-engine/transcript.md_

### 00:00 — Intro: a developer-turned-marketer at LightSource

**[0:00]** Today we have Ben Hain, developer turn marketer who has always operated at the intersection of technology and go-to-market. Ben has an impressive background that started in B2C agency work, then B2B agency, and then took his experience and talents to scale some fast-growing startups. Ben has currently had a marketing at LightSource, an AI-native sourcing system aiming to modernize procurement processes and supply chain management, and has a unique way of helping startups scale that starts with fundamentals and then accelerates with technology.

**[0:39]** Ben, super, super excited to have you here, and I know we have two really relevant topics right now that we're going to be diving into. A lot revolving around content and AI and how those can really, really help accelerate your brand. And just to kick it off, I think there's just so much hype about AI, of course, and everybody's wondering, "Am I using it right? Am I doing it right? How's so-and-so using it over there?" But when you're getting into a new organization, before you even start talking about layering in AI, where do you as a marketing leader like to start?

### 01:15 — Start with fundamentals before AI: nail your ICP

**[1:15]** Well, first of all, thanks for having me on. I'm also excited to be here, Anthony. A lot of fun to be in this kind of cool studio environment as well with your audience. So the question you're asking is, where is the hype within AI? A year and a half ago, I would have said start experimenting with it, start playing around with it, start feeding some things. And at this point now, if you're not using it, you're certainly kind of behind from a certain aspect. Before you begin marketing, before you start feeding anything into AI, though, you really do need to still do the fundamentals of figuring out who your audience is and figuring out your ICP.

### 01:56 — Learning procurement: manual LinkedIn 'coffee' calls

**[1:56]** Some of that is old school. For instance, at LightSource, it is a procurement tech platform. My background is not in procurement. Even during my interview process, I was telling everyone, "Hey, I know marketing. I know SaaS. I know startups. I don't know anything about procurement." So what I ended up doing was highly manual. I actually did a lot of manual outreach through my LinkedIn network and just said, "Hey, anyone out there, if you know someone within procurement, like I know the job titles I'm talking to, not trying to pitch, I literally just need to talk to people."

**[2:26]** I think I spent like the first 30 to 45 days doing three to five of these calls a week and just asking them like, "Hey, I'll buy you a coffee. Tell me what makes you tick. Why are you getting into this? What was your degree? How did you get into this?" You know, what values do you have? What do you look at? And surprisingly, people were really happy to tell their stories. People were really happy to kind of just share their expertise because I think in general people, especially if you're breaking into a new market, they don't want to see bad marketing.

**[2:57]** They don't want to see like less personalized marketing. So they wanted to see someone actually trying to get to know them with the effort of improving marketing. Obviously, during those conversations, if I like stopped and like started pitching the platform, I would have done poorly to begin with, though they're too short, but also like they would have really turned off right during that process. And then from there, I was able to kind of distill all of those things down, leveraging AI. I would take the transcripts. I fed it all into like a Gemini notebook and I turned it into this entire, you know, building a persona of procurement.

**[3:30]** Obviously, there's a lot more research that I did with analyst firms and folks within procurement, but those one-on-one conversations supplemented all of the research that I did to kind of round out a full ICP and persona. And then from there, I was able to build out messaging, positioning, revise our product vision and align everything toward the bedrock of that persona research that I had done. Yeah, I think that's that's really, really important to know. Like you can't just start automating things on a bad foundation.

**[4:01]** I think when people hear the terms ICP and personas, I think a lot of companies think they have their ICP and personas dialed in and tuned in, but there's so many levels and layers deeper that you really need to get to have meaningful messaging, meaningful outreach. How how would a company know if they've gone deep enough? And what are some depth examples that would be a signal to say, hey, yeah, I really understand my ICP persona really well?

### 04:32 — How deep is deep enough on ICP?

**[4:32]** Great question. So I don't know if you can ever go deep enough until the deepest you can go now and you can easily do this now with, you know, systems like clay and in our B2B is a truly one-to-one sense of messaging, where you can and I've seen this done before, where you can identify a visitor on your website and then enrich them like in clay and then like go and stalk their Instagram and other things if it's public and then literally put together their one's desires and psychographic stuff and then send them a one-to-one message. You can totally do that. It takes some rigging, but like that would probably be as deep as you can go.

### 05:09 — Mapping the 11–15 person buying committee

**[5:09]** The the shallowest you should be going is at least figuring out, you know, their main persona within like the buying committee of your software. And this is obviously from a B2B perspective from a B2C side like that previous example I gave was probably better. But from a B2B perspective, what we typically do is we understand that there are like 11 to 15 people typically in a deal mid-market to enterprise and it probably goes higher on the enterprise side at this point. So what you really need to do is find out who's in that room, who you're talking with and like what their role is in that room.

**[5:42]** And that previous example I gave you around putting together the personas, one of the key things I did was identifying the buying group and their behavior. So you have your C-suite executive that's usually the executive sponsor. You have like your champion, maybe their director level, and they're the one tasked with both discovering and researching and evaluating all the different platforms and vendors and solutions out there. And then you have your end users and your implementers.

**[6:08]** Now those end users may or may not get pulled into the buying committee that they're going to be using product and then the implementers usually they're on the tech side. I always say that they have like veto powers, right? If they hate it or if like your tech doesn't play nice or integrate with the current tech stack, they're going to boot you out. So you need to be able to speak to every single one of them. How do you know that you're speaking to them correctly? Well, you should be developing content and materials around them. And then what you can also do is once you start to identify people coming to your website, you're monitoring them.

**[6:39]** And then you're saying, OK, hey, look, we're seeing more domains or more visitors from the specific domain, you know, in a cross-section of that different buying committee. Now I'm seeing the IT guys show up to our like integrations page. Oh, I'm seeing the executive sponsor attend, you know, the C-suite webinar or this dinner or breakfast or something that we schedule. So there's signals that you can capture from there and you can kind of distill that into figuring out who's really resonating with the content, obviously, that you need to put out that is aligned to each persona.

**[7:13]** I love it. I think a lot of people don't really estimate the complexity of some of these deals to the level that they can be and how many different types of personas you may be working within a deal. Because I work with a lot of companies like, oh, yeah, I'm going after Fortune 500 companies. And it's like, OK, that is so unbelievably broad, like that's it's a vertical, you know, what are the people, the personas, like that way you can start to create meaningful messaging. So I know I know a big part of your strategy is investing in content. I know a lot of people are doing what they can to invest in content.

### 07:55 — Building a content engine on top of ICP work

**[7:55]** What's your take on it? Obviously, it's a big part of Lean Scale strategy. It's it's been a massive part of our growth engine. But how how do you think companies, especially if they're in like B2B, SaaS, B2B, AI, how should they approach content and layer in that ICP persona work that you did before? So what we're starting to see now are a couple of different things, especially if you're like an early SaaS startup like Series A, Series B, like around that point that you're still you have like one leg in the founder led sales bucket and like the other like in scalability. You still need your founder to do quite a bit of marketing.

### 08:38 — Founder-led admiration content (the SharkNinja inbound)

**[8:38]** So one example that I started to see that we are playing with at Light Source still was having our founder start to write specific articles that we'd kind of curate for him and then he'd refine in his own voice. And what these articles did were it was me doing kind of online research for key companies that we might want to target or that were interesting that have interesting products, right?

**[9:02]** I'd actually go online, research some of their Leiden data that you can see publicly online. The Leiden data shows how many different shipments and suppliers they have, what frequency, how much cost, what geography is going to and even a breakdown of different suppliers. This stuff is super nerdy and kind of boring for anyone else outside of supply chain and procurement, but it's gold for our ICP because it shows that we can cover a specific company. We can break down what their supply chain looks like and then we can pull specific insights coming from our CEO's perspective as to why and why not this was successful.

**[9:38]** So we were seeing some engagement on some broader topics like, oh, you know, EV in China, we did a breakdown like on on Fendi and their media work rise. And then we actually did a breakdown that I was starting to draft on a company called Shark Ninja that we were really fascinated with because they had this really unique kind of go to market model of leveraging their user base. It's like influencers. As I was drafting that, we actually got an inbound from Shark Ninja, which was amazing, right? So he's super excited.

**[10:06]** He's booking the meeting and I'm like, hey, you know what? We have an article that we're going to publish and our CEO went in, kind of spruced it up in his own voice. And then he posted that on his social. So you need to be doing stuff like that at least two to three times a week based on what I've seen from kind of the scalability factor. But it created a virtuous cycle in that instance, because right now that AEE is meeting with this target account and now he can point to, hey, you know what? We just did some research on you guys and like this really cool coverage article breaking into your model.

**[10:35]** We just admire you. And it's kind of building up, you know, that we're fans of them and then accelerating the deal cycle. That's just one instance of a lot of different content that you can create. Once you know your ICP, you know where to research, you know what data is going to be meaningful for them. And then you have a really good channel that, you know, is going to work. Yeah, I think something a lot of people struggle with. You let me know if you run into this. A lot of times your content efforts, they don't have a super clean demand and style attribution visibility.

### 11:10 — Content attribution without clean data

**[11:10]** So, you know, we have a ton of people watching our videos on YouTube. We have a lot of people looking at our content on our website. It's tough to say, hey, was that the initial source? Was that the source that pushed him over the deal? When you're having discussions with CEOs, CFOs, how are you justifying the investment and giving them enough data to feel comfortable having faith that it's working even when you don't have as clean of data as other sources you might. That's a great question. I think attribution is one of these arguments that you can get into and just spin your wills with everyone that you mentioned, like almost endlessly.

**[11:55]** Recently, I was at like a CMO group and there was a leader there that was just asking, you know, who still has the split level of like Piper revenue, like sales versus marketing versus other stuff. And a lot of like mixed hands were still up. But the general consensus, and this was, I would say, two years ago even, was that people don't really care that much anymore about like where stuff is coming from as long as there's pipe and rev. Obviously, you need to be able to measure and attribute probably more at the deal level, like where we think stuff came from.

**[12:30]** But things like MQLs and searching for like super early signal, they're almost vanity metrics that marketers can use internally to just judge the health of their campaigns. Obviously, if you're looking at things like click through rates and conversion rates on a per campaign and per channel basis, that's going to give you the signal to like double down on an investment or discontinue an investment or a specific creative.

**[12:56]** But once it hits the pipe stage, really just, you know, interview the folks that are creating the deals, have your rev ops team do their best to kind of build a model that everyone agrees upon to build attribution there and just reported at that stage. It's what we do at LightSource. We literally just look at, you know, the meetings that wash up and then the the actual AEs are saying, yeah, this came from an event, this came from marketing, this came from sales, this was a referral. People generally know and you can, again, have your rev ops folks like double check looking at the model. But I wouldn't really spend too much time or energy on it.

**[13:33]** Like that's time and effort that you really should be spending if you're a marketer or seller or a GTM person going to get more of those deals. Like that's the name of the game. Enriching your content, making it better, researching the ICP, like those things are going to provide better results rather than like, oh, let's let's like dissect this attribution model and, you know, add all these algorithmic permutations to it. You know, maybe there are some like fully automated PLG emotions out there that can negate this mindset. I'm not working in that space.

**[14:03]** But for a typical B2B SaaS, I wouldn't really spend a ton of time on it unless you're just drowning in deals to the extent that you, you know, have the time and energy to invest in complex attribution. Yeah, I think that makes a ton of sense. And usually we have really similar guidance at LeanScale because we're oftentimes on the hook of implementing attribution methods for these type of companies. And I agree for sales led growth, looking at the deal levels great or opportunity level and saying, hey, first, maybe first touch, last touch, keep it simple.

### 14:37 — Just ask 'how did you hear about us?' — multiple times

**[14:37]** But also just I like especially now that you can leverage AI to manage unstructured data a little bit better, just like asking the customer, hey, how'd you hear about how'd you hear about us? And they'll tell you, well, ever was most important in their mind, like maybe they did see a YouTube video, maybe they did go to your website, maybe they did see an ad on LinkedIn, but they were like, oh, one of your customers referred us. OK, then that's probably the most important component of it. And it's amazing how much that comes up and how easy that question is to ask when you're, you know, doing the disco or getting into a deal flow.

**[15:15]** The interesting thing about that, though, I found is if you can ask that question multiple times, that is actually a pretty effective form of attribution, because what you might find is like the champion, right, found you on a podcast and then some of the other folks found you in one of those different channels. So that's going to tell you different points to double down. So adding it to, you know, your request form, adding it to the disco script and then adding it to the first or second like demo script, not a bad play. Yeah, absolutely. I love it for companies that are maybe just thinking about starting their content engine.

### 15:54 — Where to start a content engine: channels and timelines

**[15:54]** And I'll preface with I assume it's going to start with your ICP and persona and how they would want to consume content. But are there any general buckets or media categories that you think of that would be good starts for those early companies? Is it podcasts? Is it YouTube? Is it, you know, social like Instagram, TikTok or like how do you how do they get started on building a content engine? And for this, and I know it took us a while, when should they start to see some of the results of that investment? Great questions. So starting with the content engine, assuming you've done, you know, the persona and ICP research,

### 16:37 — Monitoring communities: Reddit + Perplexity

**[16:37]** that should include figuring out the influencers in your space and where people are kind of congregating. So then you need to start monitoring what those channels are saying. For instance, that light source, I identified that a lot of folks were using Reddit, maybe not 100 percent of our persona, but let's say like 50 percent around there. And so I literally just started monitoring Reddit from a weekly basis, doing like a deep perplexity run report. And it would spit out trending topics, what people were talking about, complaining about, had issues with.

**[17:13]** And I would turn that into FAQs at the bottom of my blog content or just like leveraging that for the social media posts. So in general and by social media on the B2B side, it's typically linked in. Right. So in general, leveraging things like that, finding out what people are talking about, look at the influencers that people are listening to, see if you can even connect with them. There's an entire play out there. It's more of a G-team engineer sort of thing where you can start to follow, comment and engage with influencers in the space that you're playing in.

**[17:46]** And then if you can get their attention enough, have them echo out some of the content that you might be either dropping on their feed or you can drop a mention. If you do something like on the heels of some content that they published, that's effective. It's very interesting, though, like we live in this world now where, you know, there's two ways to scale, like the series, series A to B. I was just watching like the YouTube video of the RB2B founder, right? He was saying, oh, wow, like Retention.com, I scaled the wrong way. I like hired 50 people and then like we were dashed upon the rocks.

**[18:19]** But he's like, RB2B, I did everything, just podcasts, influencers, founder led marketing on the content side. And I think that's the easiest, most measurable way to go. We just started our podcast at Light Source. We've already seen some inbound coming from there three years ago when, you know, we were running podcasts at a different company. Like the observation was tough to see, but people are really starting to engage with podcasts a lot more, which, you know, awesome. Awesome choice here, Anthony, on doing your own podcast at Linkscape. So, you know, just find out what's resonating with folks. Start to put the content out there.

**[18:54]** But in terms of seeing results, you've got to plug in something that's going to monitor, you know, intense signals in your website as soon as possible. I'm not trying to sell RB2B, but it has been effective for us. And like quickly gauging, like inbound, you know, that paired with Google Analytics and some enrichment. Like we can instantly see people coming in from stuff that we're publishing almost overnight. Yeah, we're leveraging RB2B as well, and it has been really helpful to see like when we do something, what type of traffic is coming in and who they are, what the personas are. I want to take a minute. It made this is a little selfish.

### 19:36 — Optimizing LinkedIn for founder-led sales

**[19:36]** I want to take a minute to talk about LinkedIn and especially for founder-led sales companies, series A, series B, where you're expecting that founder to be on LinkedIn. What have you seen? I've seen a lot of different flavors of LinkedIn. I've seen people do it really, really poorly. There's some people that seem to have it really dialed in with engagement and everything they post just gets quite a bit of attention. What are some general guidelines you'd give and how you optimize LinkedIn if you're a founder? Great question.

**[20:10]** I feel like I'm still learning this because LinkedIn has shifted in the form of content that is being echoed out within I would say the last 18 months. I haven't perfected it and I'm still a learner, but I always proceed with the kind of scientific method of let's test, measure and scale what's working. What I do see that is working is posting at least a few times a week with your founders or your subject matter experts internally. And that's kind of what I when I say I'm learning, it's because I'm literally spinning up this program.

**[20:47]** We have a lot of thought leaders internally, so we're kind of testing with our CEO first and then kind of spreading this out to the other thought leaders. But posting a few times a week, that first play that I kind of told you about earlier, where we're looking at a company that we kind of just admire. But supplementing it with relevant data for our industry, which is supply chain and procurement, right? Like telling a cool story around that, that resonates a lot. The other thing is also filtering anything that you're going to post from the company page through the ICP to make sure it's going to resonate with them.

### 21:23 — Trendjacking through the ICP lens

**[21:23]** One story that we did recently, I think everyone was abuzz around the Taylor Swift and Travis Kelsey engagement on the B2B and B2C side a while ago. Yep, everyone took advantage of it. Everyone took advantage of it, right? Everyone said, all right, we got to post something and just get some eyeballs. And even my team, they're like, oh, we got to do something around this. Let's just post, you know, something fun. And I was like, wait a minute, wait a minute, like we could do something kind of just fun and easy. But let me run this through like the lens of the procurement person, right?

**[21:50]** That's that entire notebook that I had built with all this research. And I was like, what would your own at first think about this engagement? What I ended up finding out was, you know, what what ring and jewelry provider Kelsey went to, the cost of it, the breakdown of the bill of materials, all that stuff. I did all this deep research online and then I composed a post through the lens of procurement and sourcing and saying like, oh, my God, he only went to one supplier. And here's the bill of materials for this ring. And how could he not at least do three bids and a buy?

**[22:21]** And the best practice of sourcing and procurement to get the best deal, he really must have been in love. Right. And so like it was, you know, love through the lens of procurement and any other person kind of looking at the poster like, it's kind of nerdy. Like why are you talking about this? But we had so many people come to us that are deep in sourcing and procurement and they're like, this is the perfect post. This is like exactly how we think about everything. I even used a eye to kind of generate this overlay and like a breakdown of the bill of materials and the ring and stuff like that.

**[22:51]** But that was like a quick I'm saying 20 minutes of research thinking about that. But I had, again, this bedrock of all this persona stuff that I did and I was able to augment, you know, with this post and we were able to rapidly respond. So if you're going to do stuff on social media again, think about the ICP first, but then also hit up your subject matter experts first, get them in the rhythm and start to scale with posts that are making sense. Yeah, I love it. I think it's like trendjacking whatever is going on, but then tailoring it to your ICP. And trendjacking is back. It's 2009 all over again.

### 23:30 — Authentic vs. engineered content

**[23:30]** Yeah. So I think that's a good way to just like stay relevant, but also relevant to your persona, like modern, but also the people you're talking to. And I don't know, I've done all kinds of different posts on LinkedIn and I'm pretty serious about my LinkedIn presence for the past three, four years or so. And I think it's like I've spent a lot of time engineering posts or like putting something really thoughtful, doing a lot of research, something I think my, you know, people that I'm trying to speak to would really enjoy. And then I post like a picture of our offsite and that gets like 10 times the engagement.

**[24:09]** So I don't know. So those like maybe I should just treat it like Facebook. I'll just like post pictures of my kids and stuff. Maybe it'll be bad. That's the thing that I'm starting to see. It's like this is off. It's like authentic content, but almost to the point that it's like Facebook is starting to become the trend. So I'm actually starting to experiment with that a lot more because I was seeing the same thing. I used to post a ton of podcast snippets and even like blog posts. And I haven't posted as much recently because I've just really been curating a lot of, you know, the light sort of stuff.

**[24:43]** But I'm starting to experiment more with those just little snippets of authenticity. Right. I'm not a big fan of putting my kids up there, but just, you know, fun little anecdotes that are related to B2B life or what I'm starting to test with. So, you know, I'll come back in a couple of weeks and tell you how it's going, but we'll see how it plays. Amazing. Amazing. Last thing I just want to go into when we're talking about content, persona, ICP, top ways to leverage AI. What are some practical ways you're leveraging AI to enhance your content engine, refine your ICP and personas?

### 25:23 — The AI stack: three NotebookLM knowledge centers

**[25:23]** What are the best tips and tricks as of November 13th, 2025, because I'm sure it'll change in the next week or two. It all came through. So there's kind of stages right now that I leverage AI for. So the first is, you know, what I'm calling, you know, these bedrock pieces where developing a small LLM to curate knowledge around specific examples. So I am a huge fan of Google notebook. I've been using it from like the very beginning, actually, of a webinar that I did with another practitioner where we tested the bounds of Google notebook and the the audio overview.

**[26:02]** Like we did like 20 experiments and just broke it and now it's like a lot more a lot more robust. But those are kind of like mini knowledge centers where all used to curate things like competitive research. I'm tracking like 90 competitors in one value propositions. I'm loading up all of our release notes, documentation, analysts reports in one. And then like I said, ICP and persona research, I'm just loading up everything I know, including interviews, customer pain point calls from Gong, all of those things.

### 26:36 — Turning pain points into content at scale

**[26:36]** Like with those, you know, three central knowledge centers, I can now kind of go in and get like a deep dive on almost any of those topics that I want. So I'll give you three examples from those three notebooks. So we did have an offsite in which I had to provide like a full competitive matrix. I'm an army of one, I don't have a dedicated product marketing manager that's doing competitive analysis at this time. I'm hiring for one if there's anyone out there.

**[27:04]** But in order to get this, I actually pull all of this data out, fed it into modus and said, hey, you know, enrich this data and normalize all of these competitive points and then spit out, you know, a standard competitive matrix. I was able to do that overnight. Obviously, some human intervention and refinement was needed to kind of get the slides to work. But this data was definitely good enough to present to the company and give everyone an overview of where we lived feature by feature in terms of, you know, what's publicly released against our competitors

**[27:36]** and kind of stack rank by, you know, different revenue bands, different industries, things like that. So this was great data and you could just kind of continually refresh it. The other two pieces that I was talking about, the persona and pain points. Pain points, I think, are huge to leverage, especially if you have gone or fathom or any call recorder. If you're not running like at pinpoint or, you know, close one close lost report and pulling down those transcripts and feeding them into like a mini LLM, do that tomorrow.

**[28:06]** Like you need to start doing that. I now feed paid points into copy as a developing web copy, like I'll just have it spit out pain points. And then again, a human needs to go and like look at stuff and turn it into readable, you know, format, email copy, add copy, anything you can think of. Like I'm leveraging what's being mentioned from like a popularity sense and then slicing and dicing it again by industry vertical, which persona is saying what that is literally feeding a lot of the content that's out there. So it is a waterfall from there, though, you know, you can then feed it into some some other content engines and get ideas.

**[28:47]** I'm working with like an LLM monitoring service where I have all of the pain points converted into topics and then I have all of those personas. We have like 82 personas, like 50 topics, and then it permeates through all of these different prompts. So there's like hundreds of prompts that we're monitoring and each one of those prompts and topics then necessitates like a blog post, a white paper, all these other things that you can start to build upon. We're like just getting started, but you can literally see the visibility between you, your editors, the entire internet for each of these like different permutations on a topic prompt.

### 29:22 — Foundation first, then AI to accelerate

**[29:22]** So it's a lot to unpack. I feel like it's going to be a whole different webinar, but like that's why I'm using AI today. Now, I think that makes a ton of sense. It's start with the foundation, understand who you're solving a pain for in the first place, who's feeling the pain the most. How do you communicate to them, build content that really speaks to them and then and only then use AI to accelerate it and help with the ideation. But you really have to have those foundational elements in play and a process for storing storing data as well and making sure that you have those examples.

**[29:59]** So I think that makes a ton of sense. Well, Ben, this has been super, super insightful. Thanks for sharing everything. I love especially that it comes from like an engineering background. You've seen B2B, you've seen B2C, you've kind of synthesized all of this together in a way where we can digest it and leverage it in really practical and tactical ways, which is what I love. That's really what this podcast is about. Make sure that people can walk away and do something tangibly different tomorrow. So I just want to thank you so much for being here. Thanks for sharing everything you've done.

### 30:32 — Closing: abstract into delight; where to find Ben

**[30:32]** And like I said, I know we're sitting here on November 13th, 2025, a week from now, I'm sure you'll have different strategies and new things. So I can't wait to follow what you do and see what happens next. Thanks so much for having me on. And for the listeners out there, a lot of what I've unpacked, it might feel like foundational pieces. This is what I'm explaining. These are like the blocking and tackling pieces of scaling content and stuff like that. Once you have these things done, then you get to abstract. And by abstract, I mean you get to tap into a little bit of like the B2C campaign stuff now because the worlds are colliding, right?

**[31:10]** The noise is out there so much that you have to be authentic and do something really distinctive to make an impact. So by abstracting, what I'm saying is you know your customer, you've pumped out a ton of content, you know how to measure like how things are coming in. You're at a good place and you're like knowing what channels are working. Come up with like something really fun. That's what we're focused on now is like, hey, how do we break through the noise and come up with like a contest or a feature or like delight, right?

**[31:37]** A really delightful campaign that's going to break through. We started to experiment that at a recent event and we were like the most popular booth there. So definitely tap into that after you get these like basic steps done. I'll stop talking. You can follow me at lightsource.ai or mojo-mogul.com. Mojo-mogul is my personal brand where I'll be blogging and unpacking some of the stuff and hey, look for light source. If you know someone in procurement or you know someone in content and product marketing, I could use some help. Amazing. Well, I'm sure there's plenty of listeners that would be interested.

**[32:10]** Ben, thanks again and hope to have you again in the future. I'd love to come back. Thanks so much, Anthony.


---

_LeanScale Podcast Knowledge Hub. Free to quote and cite with attribution to The LeanScale Podcast (https://www.leanscale.team)._