---
title: "Your Website Isn't Built for AI"
episode: 56
podcast: "The LeanScale Podcast"
publisher: "LeanScale"
guest: "Kevin White"
guest_title: "Head of Marketing, Scrunch AI"
date_published: 2025-11-03
date_modified: 2026-07-22
duration: 00:34:26
word_count: 6207
topics: ["ai-in-gtm", "demand-generation", "brand-positioning", "gtm-strategy"]
canonical_url: https://leanscale-knowledge-hub.netlify.app/podcast/kevin-white-website-built-for-ai/
source: "LeanScale Podcast Knowledge Hub — https://leanscale-knowledge-hub.netlify.app"
license: "Free to quote and cite with attribution to The LeanScale Podcast."
---

# Your Website Isn't Built for AI

_Kevin White on AI bot traffic, AEO, and the future of digital marketing_

**Episode 56 · The LeanScale Podcast**  
Kevin White, Head of Marketing, Scrunch AI · Hosted by Anthony Enrico  
Published November 3, 2025 · Updated July 22, 2026 · 00:34:26  
Canonical: https://leanscale-knowledge-hub.netlify.app/podcast/kevin-white-website-built-for-ai/

**Topics:** AI in GTM · Demand Generation · Brand & Positioning · GTM Strategy


## Executive summary

For most of the last two decades, a company's website had exactly one audience: humans. Kevin White — a B2B tech marketing leader who built growth at Segment, Retool, and Common Room and now leads marketing at Scrunch AI — argues that assumption is already obsolete. Buyers increasingly default to an LLM (ChatGPT, Claude, Gemini) to do their research, compare products, and shortlist vendors, which means an AI bot, not a person, is now the thing crawling your site, interpreting it, and relaying an answer to the buyer behind the prompt. That 'bot traffic' is 10–20% of visits today at a typical company and, Kevin believes, is heading toward 50% or even 90%.

The reframe at the heart of the episode is that AI bot traffic isn't garbage — it's people. In the old world, a marketer ignored crawlers; a Google index bot carried no intent. But a retrieval bot is different: it's taking a real-time prompt from a human, fetching content from your site, and pulling the answer back into the conversation. That human is often deep in a buying state, mid-way through a long stream of comparison prompts. Studies and Scrunch's own first-party data suggest that when someone does click a citation through to your site, that traffic converts higher and carries higher deal value than ordinary organic or direct visits. Anthony's analogy lands it: the buyer has sent their executive assistant to go read your website for them.

So what do you actually do about it? Kevin's playbook is concrete. First, measure — get a baseline of which bots (retrieval vs. training vs. index) are hitting which pages, something he says 90% of companies still don't do. Second, speak the AI's language: the lingua franca of these bots is markdown, JSON, and schema markup, and heavy JavaScript, animations, and multi-column pricing pages routinely trip them up (the AI reads the back-end code, not the rendered page, and mixes the Pro plan's price with the Enterprise plan's features). Third, consider serving a parallel, AI-optimized version of your site — clean structure plus supplemental FAQs and content — alongside the pretty human version. He is blunt about the trap on the other side: mass AI-generated 'slop' pages may work as a short-term arbitrage but will get hit with a Panda-style penalty, because the models can tell when content was machine-made and adds no value.

On the market itself, Kevin sees AEO/GEO (AI Engine Optimization) tooling proliferating — incumbents like SEMrush and Ahrefs plus a wave of startups — and expects monitoring to commoditize while philosophies about how to act on the data diverge. His contrarian, hopeful note: the winning move is at the bottom of the funnel, answering very specific, long-tail questions only your brand can credibly answer, rather than fighting expensive top-of-funnel keyword wars. The back third turns personal, as Kevin traces his own path — from being handed Marketo at Gigya and told to 'figure it out' to leading whole marketing orgs — and offers a candid warning: the climb to CMO can pull you away from the hands-on work that made the job fun, and in today's small, flat AI companies, a senior IC role may be the more rewarding path.

Who should listen: marketers, founders, and growth leaders trying to understand where their traffic is really going, RevOps and demand-gen teams instrumenting a new channel before it becomes the channel, and any operator deciding whether to chase a title or stay close to the craft. The through-line is Kevin's own operating principle — curiosity and learning by doing. The best way to understand AI as a growth channel isn't to listen to someone describe it; it's to go measure it on your own site and start experimenting today.


## Key takeaways

1. **Your website now has two audiences: humans and AI bots** — Buyers are shifting research from your site to an LLM, so an AI bot fetches and interprets your content on their behalf. The site has to work for a person who wants it to look nice and for a machine that wants clean, structured content it can accurately relay.
   _Why it matters:_ Stop designing purely for human eyeballs. Treat AI legibility as a first-class requirement of your web experience, not an afterthought.
   _For:_ Marketing Leaders, Founders, RevOps Leaders

2. **AI 'bot traffic' isn't garbage — it's high-intent people** — A retrieval bot takes a real-time prompt from a human, crawls your site, and pulls the answer back into the chat. The person behind it is frequently deep in a buying journey, so this is some of the highest-intent traffic you get — the opposite of a Google index crawl you used to ignore.
   _Why it matters:_ Reclassify AI crawler activity from noise to signal. What the bots read, and how accurately they can read it, directly shapes whether a buyer shortlists you.
   _For:_ Marketing Leaders, RevOps Leaders, Revenue Executives

3. **AI is 10–20% of site traffic today and heading to 50–90%** — Kevin estimates most companies already see 10–20% of visits from AI sources, and expects that share to keep climbing as more buyers route research through LLMs. The inflection began when ChatGPT added citations and sources around January 2025.
   _Why it matters:_ This is not a someday problem. The channel is compounding now, and the teams that instrument and optimize for it early will have a durable head start.
   _For:_ Marketing Leaders, Founders

4. **Start by measuring — 90% of companies don't** — Before any optimization, get a baseline: which bots are hitting your site, which pages they crawl, and whether the mix is shifting toward high-intent retrieval bots. You can integrate at the CDN (Cloudflare, Vercel, Akamai, WordPress) via a product like Scrunch, or start manually in Google Analytics.
   _Why it matters:_ You cannot optimize a channel you can't see. Baseline measurement is the cheap, obvious first move most marketers are still skipping.
   _For:_ Marketing Leaders, RevOps Leaders

5. **Speak the AI's language: markdown, JSON, and schema markup** — The lingua franca for AI bots is markdown and JSON. Structured, technical content plus schema markup (author, definition, and the like) helps the model understand what a page means and how to reference it accurately back to the buyer.
   _Why it matters:_ Bake structured content and schema into your site as a standard, not a one-off. Clean structure is what lets an LLM cite you correctly instead of guessing.
   _For:_ Marketing Leaders, RevOps Leaders

6. **Know the failure points: heavy JavaScript and complex pricing pages** — AI can't reliably parse fancy JavaScript and animations yet, and multi-column pricing pages are a notorious trap — because the model reads the back-end code rather than the rendered page, it will mix the Pro plan's price with the Enterprise plan's features.
   _Why it matters:_ Audit the pages that matter most (pricing, product, FAQ) for how a machine reads them. If the AI garbles your pricing, buyers get wrong answers about you at the exact moment of evaluation.
   _For:_ Marketing Leaders, Founders

7. **Consider a parallel, AI-optimized version of your site** — The emerging pattern is one site that looks great for humans and a separate, parallel version served to AI bots — clean markdown/JSON plus supplemental content like extra FAQs that reinforce a page's intent. Scrunch intercepts the bot traffic and serves this AI-native experience.
   _Why it matters:_ You don't have to compromise a beautiful human site to feed the machines. Serving both audiences their preferred format is becoming a core part of technical marketing.
   _For:_ Marketing Leaders, RevOps Leaders

8. **AI-generated 'slop' content is a short-term arbitrage that will get penalized** — Spinning up hundreds of AI-generated pages to chase untapped keywords may work briefly, but Kevin expects a Panda-style update to wipe out that traffic — the models can tell when content is machine-generated and adds no real value, and they'll stop surfacing it.
   _Why it matters:_ Don't build your AEO strategy on generated volume. Invest in genuine insight, credible stats, and named authors — the signals models reward over time.
   _For:_ Marketing Leaders, Founders

9. **Win AEO at the bottom of the funnel, not on top-of-funnel keywords** — Unlike SEO's expensive, hyper-competitive keyword races, AI optimization rewards long, specific prompts — bottom-funnel questions only you (or a couple of players) can truly answer. No one games a narrow prompt, and the AI will route interested buyers straight to your content.
   _Why it matters:_ Point your content at your buyers' precise pain points and the questions your thought leaders uniquely own. Specificity, not scale, is the edge in the AI channel.
   _For:_ Marketing Leaders, Founders, RevOps Leaders

10. **AEO tooling will commoditize on monitoring and diverge on philosophy** — Kevin expects a wave of AEO/GEO players — incumbents like SEMrush and Ahrefs plus new startups — where baseline AI-traffic monitoring becomes table stakes and differentiation comes from what you do with the data (rewriting site content vs. content workflows vs. other approaches).
   _Why it matters:_ When evaluating tools, look past the dashboard to the point of view on action. And don't over-fear a single AI super-app monopolizing the channel — the market stayed plural across Google, Amazon, Facebook, and Shopify, and likely will here too.
   _For:_ Marketing Leaders, Founders, Revenue Executives

11. **Don't just read about a new channel — go run the experiment** — Kevin's strongest advice is to get hands-on: put prompts in a spreadsheet and query the models yourself, or set up a product to do it, then start iterating. The learning comes from doing, not from listening to anyone (including him) talk about it conceptually.
   _Why it matters:_ Treat AEO as a live experiment on your own site. Baseline first, then test changes and let the feedback loop teach you what actually moves visibility and conversion.
   _For:_ Marketing Leaders, RevOps Leaders, Founders

12. **Growth marketing is the path to leading marketing — but be careful what you wish for** — Kevin credits curiosity and a learn-by-doing growth mindset — merging the analytical, data-driven side with creative offer- and channel-building — for taking him to the top of marketing. Yet climbing to a leadership seat filled his days with board updates, internal politics, and SDR drama that killed his joy, and at Common Room he deliberately stepped back into a senior IC role.
   _Why it matters:_ The title everyone chases can pull you away from the hands-on work that made marketing fun. In today's small, flat AI companies, a senior IC path is a legitimate and often more rewarding choice.
   _For:_ Marketing Leaders, Founders, Revenue Executives


## Frameworks

### The Three Kinds of AI Bot Traffic (07:37)

**Definition:** AI crawlers split into three types: retrieval bots (take a live user prompt, fetch your content, and pull the answer back — RAG-style), training bots (crawl broadly to train a model over time), and traditional index bots (classic search indexing like Google).

The distinction matters because retrieval bots carry real, real-time buyer intent while training and index bots do not. Watching the mix — and whether retrieval traffic is rising — tells you how much high-intent AI demand is flowing to specific pages.

### Bot Traffic Is People (High-Intent Retrieval) (09:08)

**Definition:** Reframing AI retrieval bots not as noise but as human buyers acting through a machine: the person sends the LLM to read your site on their behalf and receives the answer inside the chat.

Because the buyer is often mid-evaluation, this is among the highest-intent traffic a site sees, and citation click-throughs convert higher and at higher deal value than typical organic or direct visits. The mental model flips a decade of 'ignore the crawlers' instinct.

### The Parallel AI Site (14:05)

**Definition:** Serve two versions of your website: the visually rich experience for humans and a separate, machine-optimized version for AI bots — clean markdown/JSON, schema markup, and supplemental content like extra FAQs — delivered by intercepting bot traffic at the CDN.

It resolves the tension between a beautiful human site and an AI-legible one by giving each audience its preferred format, so the model can accurately represent and cite you without forcing design compromises.

### AEO / AI Engine Optimization (vs. SEO) (19:44)

**Definition:** AI Engine Optimization (also called GEO) is the practice of optimizing how LLMs retrieve, represent, and cite your brand — a successor discipline to SEO that targets AI answers rather than search-result rankings.

Where SEO rewards ranking for competitive keywords via paid spend or content/backlink armies, AEO rewards structured, credible content that answers specific prompts. Kevin expects monitoring to commoditize and vendors to differ on how to act on the data.

### Bottom-of-Funnel Prompt Strategy (23:04)

**Definition:** Win the AI channel by creating content for long, specific, bottom-funnel prompts that only your brand (or a very few players) can credibly answer — instead of chasing high-awareness, top-of-funnel keywords.

Narrow prompts are hard to game and map to real pain points, so the model routes genuinely interested buyers to your content. It inverts the SEO logic of competing for broad, expensive terms.

### Baseline, Then Experiment (24:48)

**Definition:** A two-step operating loop for a brand-new channel: first establish a measurement baseline (which bots, which pages, which prompts, cited or not, versus competitors), then start making and testing changes to learn what moves visibility and conversion.

Kevin likens the baseline to Google Analytics — you need the reference point before you can infer anything — and insists the real learning only comes once you start doing and iterating, not from theory.

### Growth Marketer as the Path to Leading Marketing (29:22)

**Definition:** A growth-marketing foundation — merging the analytical, data-driven side (attribution, reporting, infrastructure) with the creative side (offers, channels, audience) — as the strongest stepping stone to running all of marketing.

Every business needs growth and distribution, so proximity to revenue performance plus command of the creative and analytical halves prepares an operator to lead the whole function. Kevin pairs it with a caution that the leadership seat can trade craft for management.


## Quotes

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

> "I really view these LLM platforms like ChatGPT or Claude or Gemini as almost the new growth channel for marketing teams — first to understand them, and then to figure out how to wield them to promote your brand."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (03:04)

> "Companies are seeing 10 to 20% of traffic from AI sources right now. In the future, I think it's going to be more like 50%, or maybe 90%, because that bot traffic is going up and up and up."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (06:15)

> "In the old world of marketing, bot traffic was garbage — I didn't want to know when Google was crawling my site. But with the AI bots crawling your site now, there could actually be real intent behind it, and oftentimes there is."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (06:37)

> "This is almost 100% guaranteed a user chatting with ChatGPT, and that traffic is visiting our site on behalf of the user — so the user is not actually coming to our site."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (08:38)

> "That bot traffic isn't actually bot traffic. It's the people behind the bot. They're sending their executive assistant to go check out your website and get the information they need."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 56 (09:08)

> "The lingua franca for AI bots is markdown and JSON. If you can optimize your site to feed that AI traffic with that kind of structured, technical content, it understands the way it should reference and serve that information back."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (12:26)

> "A pricing page is a great example — the AI isn't looking at the site like a human, it's looking at the back-end code, and it'll mix up the pricing from the pro plan with the features of the enterprise plan."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (13:31)

> "You have this nice site that looks great for humans, but then also a separate, parallel site for AI — which is also super important traffic."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (14:05)

> "That type of content is AI slop. It might work in the near term, but long term it's going to get hit with a Panda-like update and all that traffic is going to go away."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (18:31)

> "The way to optimize for AI traffic is to focus more on bottom-of-funnel queries — long prompts that answer very specific questions. No one's going to try to game that prompt."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (23:04)

> "Right now it seems like focusing on the bottom of the funnel — answering users' true pain points and making it specific to something only your brand or your thought leaders can answer — seems to be working actually pretty well."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (24:15)

> "The best way to learn a new channel like this is to go out there and do it, versus listening to people like me talk about it conceptually."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (24:48)

> "They said, hey, we bought this thing called Marketo, go figure it out. And that essentially just got my curiosity building — a learn-by-doing mentality."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (28:19)

> "On paper that career trajectory looks really great, but be careful what you wish for. My day-to-day was just filled with board updates, internal marketing, and dealing with drama from the SDR team — all this stuff that was just killing my joy."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (30:30)

> "Especially nowadays, with AI companies being small and pretty flat, that IC role is actually something that, knowing what I know now, I might lean into more."
>
> — Kevin White, The LeanScale Podcast Ep. 56 (31:16)


## Practical advice by role

### Marketing Leaders

- Instrument AI bot traffic before you optimize anything — integrate at the CDN (Cloudflare, Vercel, Akamai, WordPress) or start in Google Analytics, and separate retrieval bots from training and index bots so you can see high-intent demand.
- Rewrite your highest-stakes pages (pricing, product, FAQ) for machine legibility: clean markdown/JSON, schema markup, named authors, credible stats, and chunked content — and de-risk heavy JavaScript and multi-column pricing layouts the AI misreads.
- Aim your content at bottom-of-funnel, long-tail prompts only your brand can credibly answer, rather than fighting expensive top-of-funnel keyword battles.
- Resist the AI-slop shortcut: mass-generated pages may spike briefly but will get penalized when the models discount machine-made, low-value content.

### Founders

- Treat AI retrieval as a real, compounding growth channel now — it's already 10–20% of traffic and climbing since ChatGPT added citations in early 2025.
- When you evaluate AEO/GEO tools, look past the monitoring dashboard (which will commoditize) to the vendor's philosophy on what to actually do with the data.
- Don't assume a single AI super-app will own the channel; the market stayed plural across Google, Amazon, Facebook, and Shopify, so build for a diverse retrieval ecosystem.

### RevOps Leaders

- Add AI-source traffic to your reporting as its own high-intent segment and watch how it converts relative to organic and direct — citation click-throughs tend to convert higher and carry higher deal value.
- Baseline first, then experiment: stand up the measurement layer like you would Google Analytics, then run controlled content and structure changes and read the feedback loop.
- Partner with marketing to make structured content and schema a repeatable standard in the site build, not a manual, one-off task.


## AI takeaways

**Thesis:** AI has quietly added a second audience to every website — the LLM bot that reads your site on a buyer's behalf. Winning in this new retrieval channel is less about generating more content and more about making genuine, structured, credible content legible to machines, then measuring and iterating like it's a brand-new growth channel (because it is).

- **Bots are buyers** — Retrieval bots carry real-time human intent — the person sent the LLM to read your site. Citation click-throughs convert higher and at higher deal value than ordinary organic or direct traffic.
- **Structure is the moat** — The models' lingua franca is markdown, JSON, and schema. Heavy JavaScript and multi-column pricing pages get misread because the AI parses back-end code, not the rendered page — so it swaps Pro pricing for Enterprise features.
- **Serve a parallel AI site** — Keep a beautiful human site and serve bots a clean, AI-optimized version with supplemental FAQs and content, so the model represents and cites you accurately without design compromises.
- **Don't ship AI slop** — Mass-generated pages are a short-term arbitrage that a Panda-style penalty will erase; the models detect machine-made, low-value content and stop surfacing it.
- **Win the bottom of the funnel** — AEO rewards long, specific prompts only you can answer — a cheaper, less gameable game than SEO's expensive top-of-funnel keyword wars.
- **Measure, then do** — Baseline which bots hit which pages and whether you're cited versus competitors, then experiment. Most companies (~90%) still don't measure this at all.

**Agent & automation ideas**

- A monitoring agent that watches CDN logs to classify AI crawlers (retrieval vs. training vs. index), attributes them to pages and inferred prompts, and alerts when high-intent retrieval traffic spikes on a key page.
- A page-legibility agent that audits pricing, product, and FAQ pages the way an LLM reads them (back-end code, not rendered view) and flags where the model would garble plans, prices, or features.
- A parallel-site generator that emits a clean markdown/JSON + schema version of each page, enriched with supplemental FAQs, served only to AI bots at the edge.
- A bottom-funnel prompt agent that mines real buyer questions, tests them across ChatGPT/Claude/Gemini, and drafts brand-owned answer content for the prompts where you're absent or misrepresented.


## Metrics mentioned

| Value | Metric | Context |
| --- | --- | --- |
| 10–20% of visits | AI bot traffic today | Kevin's estimate of the share of a typical company's website traffic already coming from AI sources. |
| 50–90% of visits | Projected AI traffic | Where Kevin expects AI-driven bot traffic to head as more buyers route research through LLMs. |
| ~90% of companies | Baseline measurement gap | The share Kevin estimates still don't even measure their AI bot traffic — the cheap, obvious first move most marketers skip. |
| January 2025 | AI-traffic inflection | When ChatGPT introduced citations and sources, which Kevin marks as the point AI-referred traffic started coming online in volume. |
| Team of one | Scrunch marketing team size | Kevin runs marketing at Scrunch AI solo — his example of staying hands-on and the flat, small shape of modern AI companies. |


## Entities mentioned

- **Scrunch AI** (company) — Kevin's current company, where he leads marketing; an AI-first customer-journey / AEO platform that intercepts AI bot traffic at the CDN, shows which LLM bots crawl which pages in real time, and serves a parallel AI-optimized version of a site. · https://leanscale-knowledge-hub.netlify.app/company/scrunch-ai/
- **Segment** (company) — Where Kevin led growth marketing; part of the B2B marketing career (Segment, Retool, Common Room) that shaped his operator's view. · https://leanscale-knowledge-hub.netlify.app/company/segment/
- **Retool** (company) — Where Kevin led marketing after Segment, continuing his growth-to-marketing-leadership path. · https://leanscale-knowledge-hub.netlify.app/company/retool/
- **Common Room** (company) — Where Kevin led marketing and then deliberately stepped back into a senior IC role after board updates and internal politics 'killed his joy' running a 30-person team. · https://leanscale-knowledge-hub.netlify.app/company/common-room/
- **Gigya** (company) — The SaaS company where Kevin got his start and was handed Marketo and told to 'figure it out,' sparking his learn-by-doing growth mindset; its CEO Patrick Salyer is now a Scrunch investor. · https://leanscale-knowledge-hub.netlify.app/company/gigya/
- **RunPod** (company) — A Scrunch customer whose AEO implementation correlated with performance and conversion improvements — a customer story Kevin cites while heavily caveating correlation vs. causation. · https://leanscale-knowledge-hub.netlify.app/company/runpod/
- **Google** (company) — The old index-bot crawler marketers used to ignore, and the SEO incumbent whose expensive keyword 'vig' Kevin contrasts with bottom-funnel AEO. · https://leanscale-knowledge-hub.netlify.app/company/google/
- **OpenAI** (company) — Cited as unlikely to be 'the only company that exists' in AI — evidence the retrieval channel will stay plural rather than collapse into one super-app. · https://leanscale-knowledge-hub.netlify.app/company/openai/
- **Amazon** (company) — Named alongside Google, Facebook, and Shopify as proof the market stayed plural — a reason not to fear a single AI platform 'pulling up the ladder.' · https://leanscale-knowledge-hub.netlify.app/company/amazon/
- **Meta** (company) — Referenced as 'Facebook' among the durable independent platforms (with Google, Amazon, Shopify) that show markets specialize rather than consolidate into one winner. · https://leanscale-knowledge-hub.netlify.app/company/meta/
- **Shopify** (company) — Cited among the durable independent platforms as evidence the AI retrieval market will support many specialized players, not a single monopoly. · https://leanscale-knowledge-hub.netlify.app/company/shopify/
- **Kevin White** (person, guest) — Head of Marketing at Scrunch AI; growth-marketing leader (ex-Segment, Retool, Common Room) focused on AEO and optimizing websites for AI retrieval. · https://leanscale-knowledge-hub.netlify.app/guest/kevin-white/
- **Anthony Enrico** (person, host) — Co-founder of LeanScale and host of The LeanScale Podcast. · https://leanscale-knowledge-hub.netlify.app/guest/anthony-enrico/
- **ChatGPT** (tool, AI Assistant) — The canonical LLM research platform; its retrieval bot is shown hitting Scrunch's site live, and its addition of citations/sources (~January 2025) is what opened the AI-traffic channel.
- **Claude** (tool, AI Assistant) — Named (transcribed as 'Quad') among the LLM platforms buyers now default to for research; Claude training bots are shown crawling Scrunch's site.
- **Gemini** (tool, AI Assistant) — Cited alongside ChatGPT and Claude as one of the dozen-plus LLM platforms that have become the new research-and-growth channel.
- **Cloudflare** (tool, CDN / Edge Network) — The CDN Scrunch integrates with to intercept and surface AI bot traffic; named with Vercel, Akamai, and WordPress as supported integration points.
- **Marketo** (tool, Marketing Automation) — The marketing-automation tool Kevin was handed at Gigya and told to 'figure it out' — the spark for his learn-by-doing growth career.
- **SEMrush** (tool, SEO / Marketing Intelligence) — An SEO incumbent building AEO/AI-visibility features and a Scrunch competitor, cited as evidence the monitoring layer will commoditize.
- **Ahrefs** (tool, SEO Toolset) — Named with SEMrush as an SEO incumbent moving into AI-search optimization amid a wave of new AEO/GEO startups.


## FAQ

**Q: What is AEO (AI Engine Optimization) and how is it different from SEO?**

A: AEO — AI Engine Optimization, sometimes called GEO — is the practice of optimizing how large language models like ChatGPT, Claude, and Gemini retrieve, represent, and cite your brand when a buyer asks them a question. It's the successor to SEO: where SEO competes for keyword rankings through paid spend and content/backlink volume, AEO rewards structured, credible content that accurately answers specific prompts, and it favors long, bottom-of-funnel questions over broad, expensive top-of-funnel terms.

**Q: Is AI bot traffic on my website actually valuable?**

A: Yes — much of it is high-intent people, not noise. A retrieval bot takes a real-time prompt from a human, crawls your site, and pulls the answer back into the chat, so the person behind it is often mid-evaluation in a buying journey. First-party data and studies suggest that when someone clicks a citation through to your site, that traffic converts higher and carries higher deal value than typical organic or direct visits. That's the opposite of the old Google index crawler marketers used to ignore.

**Q: What percentage of website traffic comes from AI today?**

A: Kevin White estimates most companies already see roughly 10–20% of their website traffic from AI sources, and he expects that share to climb toward 50% or even 90% over time as more buyers route their research through LLMs. The channel began coming online in volume around January 2025, when ChatGPT introduced citations and sources.

**Q: How do you make your website AI-friendly?**

A: Start by measuring which AI bots hit which pages. Then speak the AI's language: structure content in markdown and JSON, add schema markup (author, definitions, and the like), use named authors and credible stats, and chunk content clearly. Avoid heavy JavaScript, animations, and complex multi-column pricing pages, which AI misreads because it parses back-end code rather than the rendered page. Many teams also serve a separate, parallel AI-optimized version of the site to bots via their CDN.

**Q: What is the difference between a retrieval bot and a training bot?**

A: A retrieval bot takes a live user prompt, fetches content from your site in real time, and pulls the answer back into the buyer's conversation — so it carries genuine, immediate intent. A training bot crawls broadly to gather data that trains a model over time, similar to how early ChatGPT was built, and doesn't represent a buyer in the moment. There are also traditional index bots (like classic Google search indexing). Watching whether your retrieval-bot share is rising tells you how much high-intent AI demand you're capturing.

**Q: Does using AI to generate lots of website content help your AEO?**

A: Not durably. Spinning up hundreds of AI-generated pages to chase untapped keywords may work as a short-term arbitrage, but Kevin expects a Panda-style penalty to wipe out that traffic — the models can tell when content is machine-generated and adds no real value, and they'll stop surfacing it. The lasting strategy is genuine insight, credible statistics, and named authors answering specific questions your brand can uniquely address.

**Q: How should a marketer get started with AEO?**

A: Baseline, then experiment. First establish measurement — which bots crawl your site, which pages they hit, and whether you're being cited versus competitors — either manually by putting prompts in a spreadsheet and querying the models, or with a product like Scrunch that integrates at your CDN. Once you have the baseline, start making and testing content and structure changes. Kevin's core advice is to get hands-on, because the real learning comes from doing and iterating, not from theory.

**Q: What is Scrunch AI?**

A: Scrunch AI is an AI-first customer-journey platform where Kevin White leads marketing. It integrates at your CDN (Cloudflare, Vercel, Akamai, WordPress) to intercept AI bot traffic and show, in real time, which LLM bots are crawling which pages and for what. Beyond monitoring your AI presence and comparing it against competitors, Scrunch serves a parallel, AI-optimized version of your website — clean markdown/JSON plus supplemental content — so LLMs represent and cite your brand more accurately.


## Timeline

- **00:00** — Intro: Kevin White's B2B marketing career
- **01:14** — The seismic AI shift facing marketers
- **04:52** — Bot traffic vs. human traffic (10–20% today)
- **06:37** — Why AI bot traffic is high-intent
- **08:38** — Inside the Scrunch dashboard: who's crawling you
- **11:16** — Making your website AI-friendly
- **12:26** — The AI's language: markdown, JSON, and broken pricing pages
- **14:05** — Building a parallel site for AI
- **15:20** — Results, and correlation vs. causation
- **18:01** — The AI-slop content trap
- **19:44** — Is AEO the new SEO? Can you game it?
- **23:04** — Winning at the bottom of the funnel
- **24:48** — Getting started: measure, then experiment
- **27:01** — Kevin's career: curiosity and growth marketing
- **30:30** — Be careful what you wish for: the IC path
- **33:05** — Staying hands-on + wrap


## Related episodes

- **Ep. 93: How HubSpot Built a Media Empire** (Jonathan Hunt) — The demand-gen companion — as discovery shifts to AI, HubSpot's owned-media bet is the other half of surviving the AI disruption of the top of the funnel. · https://leanscale-knowledge-hub.netlify.app/podcast/jonathan-hunt-hubspot-media-empire/
- **Ep. 33: Why AI Won't Save Your Bad Brand** (Mario Paganini) — Pairs directly with Kevin's warning against AI slop — AI amplifies real substance and punishes empty content, in brand and in AEO alike. · https://leanscale-knowledge-hub.netlify.app/podcast/mario-paganini-ai-wont-save-brand/
- **Ep. 18: Our Fastest Growing Customers Are Measuring These 3 Marketing Metrics** (Bernardo Alves) — The measurement discipline underneath Kevin's 'baseline first' rule — you can't optimize a channel you don't instrument. · https://leanscale-knowledge-hub.netlify.app/podcast/bernardo-3-marketing-metrics/
- **Ep. 92: Agents That Run Outbound While You Sleep** (Mica) — The agentic-GTM sibling — where Kevin arms the website for AI retrieval, Mica automates the outbound motion with agents. · https://leanscale-knowledge-hub.netlify.app/podcast/mica-ample-market-outbound-agents/
- **Ep. 24: AI Is Breaking Sales — Here's How to Fix It** (Mustafa Saeed) — A broader take on AI reshaping GTM motions, complementing Kevin's marketing-specific view of the same seismic shift. · https://leanscale-knowledge-hub.netlify.app/podcast/mustafa-saeed-ai-breaking-sales/


## Full transcript

_Machine-transcribed and not diarized; speaker attribution is inferred._  
_Transcript only, as a separate file: https://leanscale-knowledge-hub.netlify.app/podcast/kevin-white-website-built-for-ai/transcript.md_

### 00:00 — Intro: Kevin White's B2B marketing career

**[0:00]** Kevin White has had one of the most accomplished B2B tech marketing careers that we've had on the show, having led marketing at Segment, Retool, Common Room, and now paving the way for the AI-first customer journey with Scrunch AI. Kevin is not only a leader and innovator, but loves diving deep into the weeds as a proper operator and unbundling bottlenecks to help keep startups moving at the breakneck speed they need to.

**[0:31]** I'm super excited for today. Kevin's going to share how to transform your website for the best possible AI customer experience, not AI for your customers, but building the best experience for an AI bot to interact with your website and brand. Kevin, so stoked to have you here. This is such an exciting topic. Thank you for being on the show and excited to hop in. Yeah, awesome. This is one of the best interests I've got, maybe the best intro I've had so far. I've got to steal that from my LinkedIn bio or something like that. I really appreciate it, and you're far too kind, so hopefully I can live up to that preamble there.

### 01:14 — The seismic AI shift facing marketers

**[1:14]** No worries, no worries. I have no doubts, no doubts. I think, given your career, you've really been at the forefront of modern marketing techniques, approaches, and strategies, and everyone knows it. It's old news by now at this point, but AI is completely making a seismic shift in everything, but I think in some ways that people didn't necessarily expect. So from your perspective, what are the major shifts that you're seeing that are demanding marketers shift their strategy and tactics?

**[1:52]** Yeah, aside from just doing different types of work or creating different types of offers, I feel like this is something we could get into too, but the world of creating white papers and ebooks and all this kind of stuff is like that content is so commoditized that I feel like there's almost like a new way of looking at what kind of offers can we create.

**[2:13]** Can we vibe code something? Can we create an interactive image generation thing? So on that side of things, I'm happy to get into and I've done some of that stuff in the past, but the other thing is, I think we're going to try and talk more about today, is this kind of shift from people and humans consuming your website content and then defaulting to an LLM platform to do their research, to ask questions, to do evaluations of products.

**[2:41]** And so this is kind of left marketers and myself, I was definitely in this camp and still am in some way, is just okay, if all of our traffic of our website is shifting away to this other platform, that seems like both like a scary type of thing that is existential that we need to figure out, but also an opportunity if you can bridge that and do it in the right way.

**[3:04]** And so I really view these LLM platforms like Chat GPT or Quad or Gemini or there's a dozen of them or so, they are almost like the new growth channel for marketing teams specifically to figure out how to understand them and then after you understand them, how to wield them to promote your brand. And so that's one of the reasons why I joined Scrunch is I saw that problem that I was having, I was like, oh, we're seeing traffic coming from these sources, we know that traffic is actually doing like a lot more research behind the scenes.

**[3:45]** This is like human traffic and so there must be something here and how do I figure out even how to tap into this new channel and then I just got curiosity peaked and I just kept on going further and further down and pulling the thread on that and that's what kind of led me to Scrunch too and Scrunch is kind of solving this problem. And so that starts with monitoring, but also gets into how do we intercept that AI bot traffic and serve up a different version of our site and how we make that easier for marketers to create that experience for users and so on and so forth.

**[4:19]** So yeah, I think it's a really cool space. It's very early learning stage and I'm still learning and it's really only taken off in the last year I would say and so I'm definitely not an expert so take everything I say today with a grain of salt, but I'm happy to share what I've learned and seen so far. Yeah, I don't know if there are any experts yet just because it's so new and it's changing so quickly and exactly how the behavior looks like, but I think something that was really interesting to me and I don't know if you have any stats on this or just rough guidance.

### 04:52 — Bot traffic vs. human traffic (10–20% today)

**[4:52]** I don't think I realized how much more bot traffic and AI driven bot traffic there is than human traffic. You have a breakdown roughly of like take your run-of-the-mill site. I'm sure it differs a little bit, but like what percentage breakdown is AI traffic versus organic? Yeah, I think most companies are like probably in the 10 to 20% range actually. We have a product actually that I can show. This is kind of a fun thing to show. So let me go to our product if you don't mind and so what it does is it'll intercept that bot traffic and it'll kind of like show you what that bot traffic looks like on your site in real time.

**[5:39]** And so what I can do is kind of give you a sense of like for us at scrunch and because you have to own the website, of course, and so you have to like actually integrate with a CDN, we use CloudFlare to find that bot traffic and then service it up in analytics. And so I can share, sorry, I'm trying to like talk and do things at the same time, kind of like what that looks like to give people a sense of like the volume. To answer the question, you know, I feel like companies are seeing 10 to 20% of traffic from AI sources right now.

**[6:15]** In the future, I think it's going to be more like 50% or maybe like 90% because, you know, that bot traffic is going up and up and up. If you think about what's going on behind the scenes there, it's a human on the other side of things that's prompting an AI and then the AI is going and retrieving that traffic, retrieving your site content and crawling your site and then pulling that response back to the user.

### 06:37 — Why AI bot traffic is high-intent

**[6:37]** So in the old world of things in marketing, I was like bot traffic is garbage, like I don't want to know when Google's crawling my site, but now with the types of AI bots that are crawling your site, there could be like actually real intent behind that and oftentimes there is. And so that concept of, you know, this we call it like a retrieval type bot. It's essentially a bot that is taking a real time prompt from a user, going and fetching that data from your website and then pulling it back in the response.

**[7:07]** That's actually someone who's in a buying state. A lot of times they have a lot of intent and so that traffic is actually going to become super important. And so this is a look at our dashboard here. It's using our scrunch website because again, we are the only ones that have first party data that can like actually surface this. But you could do it for any site. You just have to integrate with your CDN like Cloudflare, Purcell, Akamai, WordPress, all these things that we integrate with and you can kind of like see this bot traffic on your own site.

**[7:37]** So we break it down by a few things like retrieval bots. These are the bots that are like doing that rag type of search, a user prompt something, the bottle go fetch something from your site and pull it back. There's like the training bots that you can see, which are more like long term type of like we're going to use this, we're going to crawl this site as an LLM and then we're going to use it to train our model. Those are a little bit more like what the initial version of chat GBT and others were like. And then there's like your typical like Google index search bots.

**[8:06]** And so you can see at the top bot agents visiting our site here, you can see the different pages that are being crawled the most by these bots. And then what's even interesting in our conversation here in the last since like 11 o'clock, you can see chat GBT is hitting up our site. Cloud training is hitting up our site and you can see the type of bot and the pages that are that are being visited. So someone on this one could have said, you know, what is scrunch is with scrunch is funding. What is stretching scrunch is pricing model and this chat GBT retrieval bot will go hit our site and then pull that answer back to the user.

### 08:38 — Inside the Scrunch dashboard: who's crawling you

**[8:38]** So this is almost 100% guaranteed a user chatting with chat GBT and then that chat GBT traffic is visiting our site on behalf of the user and so the user is not actually coming to our site. So we think I think in the future like this is what we're going to be building for markets. We're going to be building a site experience that works on the human end when someone's visiting your site and it looks nice. But then also we're feeding this AI traffic the right kind of content that it wants to surface up the right information to the people behind the prompt.

**[9:08]** Yeah, and I think I think that's an interesting part that a lot of people don't think of it was a nuance for me that hey, that bot traffic isn't actually bot traffic. It's it's people behind the bot. So they're sending their executive assistant to go check out your website and get the information that they need, but there's an actual person who's receiving that information. So that's actually I would argue in a lot of ways probably some of the highest intent traffic that you might be getting because you might get random people coming to your website on accident.

**[9:41]** But if somebody is going into an L I'm asking a very specific question and your website has relevant information for that question and then you're getting positioned in front of that person, that's really, really important interaction. Yeah, the behavior that we kind of infer and some data that we've seen there's been studies on this where if you're getting human traffic from an LM like they click on the citation and actually get to your sites.

**[10:09]** That traffic is actually converting a lot higher and higher has usually a higher value with the visit compared to like, you know, organic or direct traffic. And if you think about the reason why is it usually the person is prompting the GPT or the LM with many like a long stream of questions like compare this product versus this other product. I want to learn about this. I have this problem. I have this pain point and they're getting a feed of information. And then once they get to a point where it's like, oh, I want to make a decision now.

**[10:42]** That's when they would click the link and come over to your site and then convert. And so we you kind of see that behavior play out with your traffic from LMS. But it's even like but by capturing that bot traffic, you get a sense of what's happening before that as the bots are doing the work with as as the person's chatting, if that makes sense. Yeah, totally makes sense. And I know it's early days. So I know there might not be a full formula for this. And I know the LMS may behave a little bit differently in the near future, who knows.

### 11:16 — Making your website AI-friendly

**[11:16]** But current state, are there any practical things a company can do to make sure that their website is set up to accommodate the bot traffic and give it the information it needs and do it in a way that gives them what they want efficiently and passes it back to the person that's working with it? What are some practical things people can do? Yeah, well, the one thing to start is just starting to measure this traffic in general. And so whether you're going to measure it through Google Analytics or product like scrunch, I feel like just getting a baseline of, you know, this is the bot traffic that's visiting my site.

**[11:58]** And like, why and trying to interpret like, why is this bot traffic there? And what is it doing? And like, what is the intent behind it gives you like a baseline and then you can kind of like measure like, oh, is our, our, is our bot traffic going up over time? Is it going up for the right, the right types of retrieval bot versus a training bot, etc. So just like baseline measurement, which is like what I would say 90% of people like don't even do right now is probably a good starting point.

### 12:26 — The AI's language: markdown, JSON, and broken pricing pages

**[12:26]** From there, it's kind of like there are the bots have a way of crawling and indexing a site. And so they really speak the language like the lingua franca for for AI bots is markdown and JSON. So if you can optimize your site to feed this AI traffic with that type of structural like technical content, also like schema markup, like, this is an author post, this is a, this is a definition whatever like Google has this whole like, you know, this is the schema that you can use for your content.

**[13:03]** AI is really good at like consuming content, it wants to like consume content in that way. And so some broad strokes things are like if you have a bunch of like fancy JavaScript and animations and stuff like that on your site. AI is not really good at parsing and understanding that JavaScript yet, it might get there but that is one thing that, you know, you might want to look out for other things that we've seen kind of AI get trapped tricked up with is the structure of like a pricing page is a really

**[13:31]** good example where there's multiple columns, and there's different pricing structures and different features for like four or five different plans, for example, and if you look at the code behind it because the AI is not looking at the site, like, like a human would it's looking at the back end code. It's oftentimes like structured in a way where it mixes up the pricing from the pro plan to the features of the enterprise plan and all this kind of stuff. That is like a good example of, you know, structuring your content in the right way for AI to consume it by putting in like nice markdown nice format helps it understand what the

### 14:05 — Building a parallel site for AI

**[14:05]** way it should be referencing and servicing that information back to the user behind the prompts. I don't want to plug scrunch too much but but that is like the product that we are building as well which is not only intercepting this AI site traffic is in giving you analytics behind it, but servicing up a separate version of your website a parallel version of your website that includes things like the AI needs like markdown JSON clean format but then also maybe additional content like more FAQs about this topic or different content to supplement what that page is intent is and so

**[14:38]** that kind of separates like you have this nice site that looks great for humans but then also like a separate parallel site for AI which is also super important traffic. And so we see that as kind of like the future or the way of things happening with marketing is to build this kind of site that is that is optimized for AI to consume and that's how you communicate with this channel. Yeah, and I'm wondering, in the work you've done with customers, do you see any specific results so once you start to optimize or the LMS to receive the information well teed up to the people there that are getting prompted.

### 15:20 — Results, and correlation vs. causation

**[15:20]** How, what type of tip and performance you typically see once they've implemented some of these practices. And this is the part where like heavily caveat things is that it's definitely correlation, but it's, it's impossible to know the causation because the AI like LMS platforms don't open up like this is how we score algorithm on things. But we, when we, when we roll out this product, you can see that the content that you, the first thing is just showing that it works. So it's like, we fed AI this content, your site was actually formatted in this way before, and after the AI is now

**[16:03]** servicing an accurate representation of that pricing page, that FAQ page that whatever so that's like the first thing is just getting the getting the accuracy of the content that you want to communicate that over to the AI in the right way. And then the other thing that we see is definitely performance improvements in terms of like traffic and prompt visibility and citations and stuff like that that's actually the first part of our product is just baseline monitoring your AI presence comparing it versus competitors

**[16:32]** and different stages of the funnel and all these other sorts of dimensions, and then like getting insights to optimize that but then the third part is this like, you know, build a site that's built for AI consumption. But we do see when best practices around, I hate the word best practices shouldn't be used best practices but that's okay. When people are taking those insights and implementing them, they see, you know, one of our customers has like

**[16:59]** this X performance improvements and conversions called the run pod. There's a customer story up on our site. And I would say, this is, it's, I put a huge caveat on this because I'm not going to say like overnight, you're going to get 6x more traffic and conversions and all this kind of stuff. Sure. It's a rising tide lifts all boats so you're also getting more people using AI and more traffic from AI in general, but we do see heavy correlation with if you implement these things, like,

**[17:28]** using authors on your site, chunking out content, putting stats that are credible, like those types of things tend to help the LLMs surface up the right kind of content to the end user. And as a way of doing that, then you tend to see a correlation with higher traffic and higher conversion rates. Yeah, makes a ton of sense. And I know I, early days, who knows what best practices right now. I think there's a currently this is what some people are doing and I think a lot of people aren't even aware of some of this stuff going on.

### 18:01 — The AI-slop content trap

**[18:01]** There is, there is one thing, sorry, there is one thing that I caution kind of against, which is a lot of, a lot of folks out there will say, hey, you need to find untapped content opportunities and then use AI to generate content, which, to me, ends up you create these workflows that like generate a bunch of pages like anyone can create a page nowadays with just a prompt and it takes zero time and zero effort.

**[18:31]** And so I feel like that type of content is AI sloth and it might work in the near term, but I feel like in the long term, it's going to be hit with a Panda like update and like all that traffic is going to go away. So maybe now's the time to do lean into that arbitrage if you're open to the risk. But I feel like over time, like that strategy is going to be a losing strategy and the LMS know if something's generated by an LMS. So at some point, I think they'll turn off any kind of visibility or any kind of traffic to that if it's not actually helping servicing the user, you know.

**[19:07]** Yeah, I think that makes a ton of sense, which I'm excited about. I'm excited for, hey, people who have real content to share, organic stuff to share, actual insights. I mean, I've seen it a million times. You go on somebody's website and they have hundreds of articles. Oscillary? Yeah, a bunch of definitions of things, yeah. It's like, come on, nobody wrote these and nobody put those together. But in your opinion, there's such an industry around SEO and, you know, there's so many companies and tools and everything just like built just to do that.

### 19:44 — Is AEO the new SEO? Can you game it?

**[19:44]** Do you think we're going to see the same thing or what it's been branded as, AEO right now? Or do you think the nature of the LMS themselves will kind of make it difficult to gain the system? Yeah. Interesting question. I feel like we're already starting to see all these types of AEO, GEO companies spin up. I mean, Scrunch certainly has dozens, if not maybe hundreds of competitors already. Some of it's the incumbents like SEMrush and Ahrefs that are building something similar and they're big in the SEO space. And then there's other like startups that spin up, you know, almost every day, every week.

**[20:27]** So I actually think that that market of monitoring AI traffic, well, I think our product is, you know, the best one out there. And I'm biased. I do think that that's going to be a commoditized feature. And where we'll start to see diversions in the market is the different philosophies around like, oh, we have this philosophy of changing the content of your website to service this AI bot traffic. But others might have a philosophy around, you know, building content workflows or something like that or some other divergent market differentiator there.

**[21:07]** So the market is definitely shaping out. It's still early days. I feel like all this traffic from AI started coming more online when chat GPT introduced citations and sources. And that seemed to be like the January 2025 period. And we started seeing like more and more traffic over time. And then people are like, oh, I need to measure this traffic. And then I need a tool for that. And so that's how the market kind of got started. But I feel like over time, it's going to evolve quite a bit. And there will be more science around it, I would imagine. The other question around is, are the AI platforms just going to pull up the ladder on us?

**[21:50]** Maybe, I mean, that's totally possible. They could just make a super app and there's nothing else that exists. I don't think that that plays out in the market so far. Like there's Facebook, there's Google, there's Amazon, and there's also Shopify. There's like so many different other alternatives to things out there that I feel like there's going to be a market across the board and specializing. I don't think OpenAI is going to like just be the only company that exists, if that makes sense.

**[22:23]** Oh, it totally makes sense. Well, it makes sense. I agree. I think that's probably a hyperbolic view of how it might end up. But I think, I'm just curious how that industry is going to shape up and how much you can, I don't want to call it gaming because I know some of it's just like, hey, we're going to create, put out information that is helpful. But I'm just wondering how much of it you can game versus what the SEO industry seems to have done. What I actually really like about some of the differences that I'm seeing that were, you know, again, this is all a black box.

### 23:04 — Winning at the bottom of the funnel

**[23:04]** So but like some of the correlation that we're seeing is with the way to optimize for AI traffic is actually to focus a little bit more on like bottom funnel type of queries, like long prompts that answer like very specific questions. And if I, you know, I think there's a world where, you know, you're the only company out there that can answer this question, or there's maybe like two or three out there and like no one's going to try and game that prompt. And so you, you can create content around this like very specific thing and like AI will find people who are interested in that and then service your content to them.

**[23:45]** I feel like we're seeing more of that than than compared to something like SEO, which is like rank for this keyword. It has all this awareness. It's a super competitive game. And it's like really, it's either expensive to pay Google the vig to show up there, or like you have to have like a content army and backlink army to like rank number one, you know, so I feel like it is still early days. You know, it's like, you know, over time, it might become harder and harder as people figure out, you know, new opportunities and it gets known and spread and marketers ruin everything.

**[24:15]** But right now it seems like focusing on what matters for that bottom of funnel like answering users like true pain points and making that specific to something like your brand or your, your thought leaders can only answer like that seems to be working actually pretty well. I think these insights are incredible. I'm learning a ton right now. I'm just curious. Is there anything else as a business we can be doing to put our best foot forward with these LMS and just give them the information they need.

### 24:48 — Getting started: measure, then experiment

**[24:48]** Yeah, I'm a big fan of instead of reading something or digesting some content or listening to this podcast to go out there and actually try it yourself like getting your hands on, especially in a new market like this. There's a new channel like the best way to learn it is to go out there and do it versus listen to people like me talk about it conceptually. That said, I do feel like there is kind of like a minimum thing you can do, which is just to start measuring the stuff, measuring how your brand's showing up and prompts measuring where your competitors are showing up if you're being cited or not.

**[25:27]** Like that baseline, at least give it's kind of like Google analytics in their list. You need a baseline to understand what's happening with analytics and then you can kind of like make inferences and insights and decisions based off of that. And so I would encourage people to look at look into a you can do this manually. Actually, you can like put these prompts into a spreadsheet. You can prompt the LN models yourself. You can probably build a custom GPT or something like do all this for you.

**[25:52]** I would say that's probably not a good use of time when there's like a, you know, credit card transaction type of product like scrunch out there where you can just, you know, set this up and we can do all that mess for you. But try one of these products. I mean, I hope everyone here will give scrunch or try to, but like, you know, there's plenty of other products out there to just measure this from a baseline. And then from there you'll I guarantee you'll start to get ideas like, oh, we're seeing this this corner of prompts aren't being aren't being addressed the right way.

**[26:25]** We can create content around that or we can like change some of our site content to address that and then you'll start building and learning over time. And so that's my advice is start measuring it, but then also start doing and start trying things because that's where you'll get the most learning from. Yeah, I think that's great advice. Just get the feedback we've going now start iterating moving forward. And I think to your point, yes, hands on best way to learn things. I may arguably the only way to learn things. Well, I'm I'm curious. I'd love for you to share just I think you've had such an interesting career in marketing.

### 27:01 — Kevin's career: curiosity and growth marketing

**[27:01]** And I'd love if you could just share how you ended up at scrunch and kind of the experiences that stacked along the way to get you to where you are today. And one thing I always ask if our guests can share if there's anything you attribute your success to or anything that you think someone could copy and emulate for themselves. Yeah, yeah. We'll see. That's a tough, tough, tough thing to answer. I feel like I'm like reflecting back introspecting on my my career. No, I think, you know, the thing that got me to where I am is curiosity.

**[27:45]** And so what I mean by that is when I first started in SaaS marketing, essentially I this was back in the day where I I joined the SaaS companies called Gigia. Actually, the the CEO of Gigia is now an investor in scrunch. And one of the reasons why I joined scrunch. Well, that's a different story. That's a shout out, Patrick Salier, if you're listening. But like when I joined this company, I was very junior, not, you know, not knowing how SaaS and PLG and all this kind of stuff works.

**[28:19]** They're just like, hey, we bought this thing called Marketo, like go figure it out. And so that essentially just got my curiosity building and learning kind of learning by doing mentality going or just like, I need to figure this thing out. I need to understand how to generate revenue, how to like get people interested. And from there, like, I kind of just that got me into more of like a growth mindset where it's like, oh, we built out this infrastructure for lifecycle marketing. How do we get people on the top of the funnel? How do we get them to engage with this stuff?

**[28:51]** And so, you know, it just by by setting something up, you kind of like get this curiosity of like, how do I go and like learn how to achieve this objective that we want to do? And then I feel like that leads a lot more to like a growth type of mindset, like a growth marketing type of mindset, which is, hey, we need it, we need to solve this thing. I we need to do both like the analytical and the creative part of it. So like the creative part is like, oh, what offer can we put together?

**[29:22]** What channels can we reach to get to this audience? But then also the the infrastructure and data driven side of things like when someone actually does this action, you know, how do we report on that? What insights can we learn from it? Like what data science can we apply to that? And like merging those two things like tend to be a really good combination for a growth marketer. And growth marketer is a really good stepping stone to leading all of marketing, I would say, because I mean, everyone needs growth and distribution.

**[29:53]** Otherwise, you don't have a business. So I would say that has been what's got me to where I am in my career to leading growth at segments and then also, you know, leading marketing at retool and then common room and then now scrunch. So, yeah, hopefully that somewhat answers the question. No, I think it makes a ton of sense. And I think taking that growth marketer path as the path to, you know, running the whole show ahead of marketing CMO, I think I think you're right. Closer you can get to the revenue performance and then use everything else as a tool brand and your channels and everything else to manage and techniques.

### 30:30 — Be careful what you wish for: the IC path

**[30:30]** But being can I add one other piece of career advice? Yeah, to this as well. So on paper, that career trajectory looks really great. But be careful what you wish for because I got myself to a point where I was leading a marketing team, I was really ambitious and what your day to day looks like from where you're starting to grow to where like you're leading a team of like 30 people is very different.

**[30:59]** And so I ended up at common room actually taking like a step back and taking more of a super icy role because my day to day was just filled with board updates, internal marketing, dealing with drama from the SDR team, like all this kind of stuff that was just like killing my joy.

**[31:16]** And so if that's there are people who love doing that stuff. And if you aspire to do that, like do it in great. But also, there's another path for being an IC. And I feel like especially nowadays with AI companies being small and being pretty flat, like that type of role is actually something that, you know, if I was starting early, I might go in more in that direction, knowing what I know now. Yeah, I think that's great. And you may have to go through that experience to even know because it looks so alluring, like, when you're early in your career and, you know, wanting to be in a position to make an impact and be a leader.

**[31:51]** But to your point, the day to day, it's pretty treacherous and mundane and kind of boring sometimes. So, and it pulls you away from some of the like newer, exciting hands on work that you could be doing in that temporal team. Yeah, and the pay is not all that much more for the stress. So, maybe equity wise it is. But yeah, anyway, different topic. No, I think that's really, that's sage advice. And I appreciate you sharing that because I think most people have in their heads, like, hey, how do I rocket ship to the C-suite? But are you sure that's actually what you want to do?

**[32:27]** And I do think there are more opportunities now because the organizations are relatively flat, where even if you have a C title, like, oftentimes, you still have to roll up your sleeves and do stuff. So, I think that's a good thing. I think most of us get attracted to the work, not necessarily just professional management, and especially board management and things like that. So, I think it's a great path, if you can carve it out. Yeah, I feel like I wouldn't be able to give advice if I wasn't hands on with things. And so, despite driving some of my team members crazy, and I'm a team of one right now at Scrunch, so I can drive myself crazy.

### 33:05 — Staying hands-on + wrap

**[33:05]** Despite that, I still feel like I must be hands on with something. And for the people who are in that, you know, path of destruction, who are working with me on that thing, I feel sorry for them. But I'll pull back and bother another team member at some point as I build my game. So, yeah. Amazing, amazing. Well, Kevin, this has been awesome. I think you really illuminated this stuff for me where, hey, a lot of the traffic already 10, 20% of the traffic is really guided by LLMs retrieving information, teeing it up to people who are prompting them to get the information they need.

**[33:43]** That's only going to increase over time. I love the practical recommendations and tactics that you gave to help put your best foot forward and just get started measuring so you know how much of an impact it has on your business. Scrunch AI, your technology, your platform looks incredible. So people can leverage that to start getting a baseline. And then who knows what the future holds. But I trust someone like you is going to figure it out and move with the waves as they come. And thank you so much for sharing everything you did today. Of course, yeah, thanks for having me. It was a blast. I had a good time. Awesome. Thanks, Kevin.


---

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