---
title: "Using ChatGPT as a Salesforce Admin"
episode: 5
podcast: "The LeanScale Podcast"
publisher: "LeanScale"
guest: "Christopher Martyen"
guest_title: "Systems Architect, LeanScale"
date_published: 2023-05-09
date_modified: 2026-07-22
duration: 00:15:12
word_count: 2711
topics: ["ai-in-gtm", "revenue-operations", "forecasting"]
canonical_url: https://leanscale-knowledge-hub.netlify.app/podcast/christopher-martyen-chatgpt-salesforce-admin/
source: "LeanScale Podcast Knowledge Hub — https://leanscale-knowledge-hub.netlify.app"
license: "Free to quote and cite with attribution to The LeanScale Podcast."
---

# Using ChatGPT as a Salesforce Admin

_LeanScale systems architect Christopher Martyen on debugging, generating, and translating Salesforce config with ChatGPT_

**Episode 5 · The LeanScale Podcast**  
Christopher Martyen, Systems Architect, LeanScale · Hosted by Anthony Enrico  
Published May 9, 2023 · Updated July 22, 2026 · 00:15:12  
Canonical: https://leanscale-knowledge-hub.netlify.app/podcast/christopher-martyen-chatgpt-salesforce-admin/

**Topics:** AI in GTM · Revenue Operations · Forecasting


## Executive summary

Recorded in May 2023, in the first wave of ChatGPT hype, this is one of the earliest LeanScale Podcast episodes — a candid time capsule of how a hands-on RevOps practitioner first reached for a general-purpose LLM inside a CRM. Anthony Enrico sits down with LeanScale systems architect Christopher Martyen for a live, screen-shared walkthrough of how he uses ChatGPT in his day-to-day Salesforce admin work. The mood is half-excited, half-terrified, and Christopher leans into the fear: he got into AI partly because it scared him, the same way people in the 1990s were scared that computers and email would kill letter-writing and the paper industry.

The substance is three concrete use cases. First, debugging: when the peer you'd normally ask is stuck in a meeting, you can paste a broken formula into ChatGPT and get an instant diagnosis of what's wrong. Second, generation: describe a business rule in plain English — 'people shouldn't be able to move an opportunity from Nurture to SQL unless the Type field has a value' — and ChatGPT writes the Salesforce validation rule, functions and punctuation and all, ready to copy in. Third, translation: paste an inherited 15–20-line formula and ask 'what does this formula do?' to get a plain-English explanation. The through-line is collaboration — AI lowers the technical barrier so non-admins can participate, and it shrinks the research tax of understanding a Salesforce org that is, in Anthony's words, a 'labyrinth of workflows and validation rules.'

Christopher is emphatic that a human stays in the loop. ChatGPT declared 'no errors found' on a rule that would have failed in production because it used ISBLANK on a picklist field — something Christopher caught by eye and had already had to correct earlier that morning. His closing worldview is squarely augmentation, not replacement: he cites a Goldman Sachs analyst estimate that 20–25% of jobs could be replaced by AI, then counters that email displaced some work while creating far more — history rhymes. For RevOps operators and Salesforce admins, the episode doubles as a practical starter kit and an early, prescient framing of the augmentation thesis LeanScale returns to for years afterward.


## Key takeaways

1. **ChatGPT is an always-available technical pair partner** — When the peer you'd normally ask is in a meeting, ChatGPT fills the gap. Christopher had a big formula that wasn't working, pasted it in, and it told him exactly what was wrong so he could test the fix — no waiting on a human to free up.
   _Why it matters:_ Salesforce admins and RevOps engineers can compress the debugging loop from 'wait for a colleague' to 'get a second opinion on demand,' which matters most on small teams where technical reviewers are scarce.
   _For:_ RevOps Leaders, Revenue Executives

2. **Plain-English in, working Salesforce config out** — Christopher pastes a plain-language request — 'people should not be able to move an opportunity to SQL from Nurture unless the Type field has a value' — and ChatGPT returns a validation rule using ISCHANGED, PRIORVALUE, ISPICKVAL, ISBLANK and TEXT, which he copies straight into Salesforce.
   _Why it matters:_ The syntax barrier that once gated admin work drops away. Stakeholders can describe rules the way they think about the business, and the admin's job shifts toward reviewing and refining rather than hand-typing every formula.
   _For:_ RevOps Leaders, Revenue Executives

3. **Keep a human in the loop — 'no errors found' is not 'correct'** — ChatGPT reported no errors on a validation rule that used ISBLANK on a picklist field, which doesn't behave as expected (you need to wrap it in TEXT). Christopher caught it by eye because he already understood the Salesforce gotcha.
   _Why it matters:_ LLM output still needs an operator who knows the platform's edge cases and the business context. Trusting generated config blindly ships silent bugs; a human review gate is non-negotiable.
   _For:_ RevOps Leaders, Revenue Executives

4. **Re-prompting refines the output like a peer-review loop** — Christopher told ChatGPT its ISBLANK version had an error and it corrected itself; he jokes 'maybe it learned from this morning.' The 20–30-minute client formula he shows took a few prompts and re-prompts to get exactly right.
   _Why it matters:_ Treat the tool as an iterative collaborator, not a one-shot oracle — the same back-and-forth you'd have across a couple of meetings with a teammate, compressed into minutes.
   _For:_ RevOps Leaders

5. **'What does this formula do?' turns AI into a translation layer** — Pasting an inherited, jargon-dense formula and asking ChatGPT to explain it returns a plain-English breakdown — e.g., 'if the term end date is blank, it calculates the day by adding one year to the closed date' — including how it handles a leap year via MOD(YEAR, 4).
   _Why it matters:_ Admins can understand and document configuration they didn't write, and explain it to others in seconds instead of spending 20 minutes writing a page of explanation.
   _For:_ RevOps Leaders, Revenue Executives

6. **AI collapses the research tax on inherited Salesforce orgs** — As Anthony notes, every new Salesforce instance is 'a complete labyrinth of workflows and validation rules' that a team has to reverse-engineer before doing any real work. Translation-on-demand shortens that discovery dramatically.
   _Why it matters:_ Onboarding to an unfamiliar org — a constant in fractional and agency RevOps — gets faster and less error-prone when you can ask the system to explain itself.
   _For:_ RevOps Leaders, Revenue Executives

7. **Two value modes: do it faster, or solve something more complex** — Anthony's framing: leveraging AI either helps you do a known task faster, or helps you solve something harder by giving you inspiration for potential solutions you couldn't reach alone.
   _Why it matters:_ Evaluate AI use cases against both axes. Speed wins are obvious; the underrated win is unlocking work that was previously beyond an individual's skill ceiling.
   _For:_ RevOps Leaders, Revenue Executives

8. **Augmentation, not replacement — the email-and-paper precedent** — Christopher counters the fear with history: in the '90s people thought email would kill paper, but it grew the industry (paper companies pivoted to the cardboard boxes now shipping from Amazon). He cites a Goldman Sachs estimate that 20–25% of jobs face automation, then asks how many jobs email erased versus created.
   _Why it matters:_ The default posture toward AI in RevOps should be adoption and augmentation, not defensive avoidance. History rhymes: new tooling tends to enhance technical roles and expand the work, not eliminate it.
   _For:_ Revenue Executives, RevOps Leaders


## Frameworks

### ChatGPT as an Always-Available Pair Partner (03:14)

**Definition:** Use ChatGPT as the technical collaborator you turn to when a human peer is unavailable — paste a broken formula or a stuck problem and get an immediate diagnosis and a testable fix.

Christopher frames the tool as filling the role a trusted colleague would: reviewing your work and catching what's wrong, on demand, so you're not blocked waiting for someone to come out of a meeting.

### Plain-English In, Working Config Out (05:40)

**Definition:** Describe a Salesforce business rule in ordinary human language and let ChatGPT translate it into the validation rule or formula, then review the output for business context before saving.

The live demo turns 'people shouldn't move an opp from Nurture to SQL unless Type is filled in' into a copy-ready validation rule, opening admin work to teammates who don't write formulas.

### Ask It What This Does (Translation Layer) (10:22)

**Definition:** Paste an existing, complex formula and ask ChatGPT 'what does this formula do?' to get a plain-English explanation you can understand and pass on to others.

This democratizes review of inherited configuration and lets an admin explain a dense formula to a stakeholder in seconds instead of writing a page of documentation.

### History Rhymes: The Email-vs-Paper Precedent (01:30)

**Definition:** A mental model for reacting to disruptive technology: history doesn't repeat but it rhymes, and past waves (like email) augmented and grew work rather than eliminating it.

Christopher uses the 1990s fear that email would kill paper — which instead grew the paper-and-cardboard industry — to argue AI will enhance technical roles, tempering the Goldman Sachs job-automation figure.


## Quotes

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

> "And just like that, ChatGPT instantly created this validation rule for me. Salesforce, in general, is already low to no code, but this is something that makes it much more inviting to almost anyone to come collaborate."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (00:00)

> "I still have mixed feelings about ChatGPT, the power of AI, but at the same time it's still absolutely fascinating — and I really love the applications it has for revenue operations."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 5 (00:00)

> "The reason I got into it was because I was a little bit terrified of it, like many people in the '90s were terrified of computers and emailing."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (01:30)

> "It can be terrifying, but history tends to not necessarily repeat itself, but tends to rhyme. I've seen this story a few times."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (02:04)

> "Instead of relying on someone else, waiting for them to be available and then collaborating on that problem, ChatGPT was that partner — and I think ChatGPT and other AI tools can be that partner for a lot of people in technical roles."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (03:50)

> "The more people that we can get to collaborate — opening up things like code and formulas to people who might not understand it — the better the work becomes."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (04:57)

> "Before this demo I tried this and it had ISBLANK on the type, and I told it, 'no, that has an error, please correct it,' and it corrected it. So maybe it learned from this morning."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (07:28)

> "For most people, leveraging AI is either going to help you do something faster, or it's going to help you solve something more complex by giving you inspiration of potential solutions."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 5 (08:25)

> "When we got translation tools, we were able to collaborate across borders like no other time in history. Now, with ChatGPT and other AI tools, we can translate a lot of these technical details so fast."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (12:10)

> "Every time we run into an instance of Salesforce, there's usually a complete labyrinth of workflows and validation rules that we need to do our own research on just to understand what was implemented before we begin our work."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 5 (12:43)

> "Some analysts from Goldman Sachs predicted that about 20 to 25% of jobs would be replaced by AI. Well, how many jobs were replaced by email? We gained a lot more jobs with technology as we grow."
>
> — Christopher Martyen, The LeanScale Podcast Ep. 5 (14:01)

> "I have a slight feeling of terror when I look at it, just because it seems so powerful — but I agree, the future is bright."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 5 (14:42)


## Practical advice by role

### RevOps Leaders

- When a technical reviewer is unavailable, paste your broken formula into ChatGPT for an instant diagnosis — but treat it as a second opinion, not a verdict.
- Let stakeholders describe rules in plain English and use ChatGPT to draft the validation rule or formula; keep the admin as the reviewer who owns business context.
- Always review generated config for platform edge cases before saving — 'no errors found' missed an ISBLANK-on-picklist bug that needed a TEXT() wrap.
- Use 'what does this formula do?' to translate inherited configuration when onboarding to an unfamiliar Salesforce org, and to explain your own formulas to non-technical teammates in seconds.
- Re-prompt like a peer-review loop; the best output comes from a few iterations, not one shot.

### Revenue Executives

- Frame AI adoption as augmentation, not replacement — the email-and-paper precedent suggests new tooling grows technical work rather than erasing it.
- Encourage the team to lean into the tool despite the fear; the fastest adopters compress both simple tasks (do it faster) and hard ones (solve something more complex).
- Invest in the human judgment layer — the value is an operator who can catch what the model misses and translate between the business and the system.


## AI takeaways

**Thesis:** ChatGPT turns Salesforce admin work into a plain-English, collaborative exercise — an always-on pair partner that debugs, generates, and translates configuration — and, like email before it, augments technical roles rather than replacing them, provided a human keeps ownership of business context and platform edge cases.

- **A pair partner on demand** — Paste a broken formula and get an instant diagnosis when your go-to human peer is in a meeting — the debugging loop shrinks from 'wait for a colleague' to 'get a second opinion now.'
- **Plain English becomes config** — Describe a business rule the way you think about it and ChatGPT writes the validation rule; the admin's job shifts from typing syntax to reviewing and refining.
- **A translation layer for legacy orgs** — Ask 'what does this formula do?' to understand inherited, jargon-dense configuration in seconds — collapsing the research tax on every unfamiliar Salesforce instance.
- **Human-in-the-loop is non-negotiable** — ChatGPT said 'no errors found' on an ISBLANK-on-picklist rule that would have failed; only an operator who knows the platform caught it. Re-prompting fixed it.
- **Augmentation over replacement** — The email-and-paper precedent — and the Goldman Sachs job-automation figure it rebuts — frame AI as a force that grows technical work rather than erasing it.

**Agent & automation ideas**

- A Salesforce 'formula copilot' that turns plain-English business requirements into validation rules and formulas, and auto-flags platform gotchas (like ISBLANK on a picklist that needs a TEXT() wrap) before an admin saves.
- An org-onboarding agent that ingests an unfamiliar Salesforce instance and produces plain-English documentation of every validation rule, workflow, and formula to shorten discovery.
- A request-to-config intake that lets non-technical stakeholders describe a rule in plain English and drafts the config for admin review, keeping a human approval gate.
- A date/renewal-logic checker that computes term-end and renewal dates with leap-year-safe logic and validates them against close dates.


## Operations takeaways

### Revenue operations

- **Democratize admin work.** Plain-English requests become drafted config that an admin reviews — non-technical teammates can participate in building and reviewing rules.
- **Translate inherited orgs.** Paste formulas and rules and ask what they do; discovery on an unfamiliar Salesforce instance goes from hours of research to minutes.
- **Guard the edges.** LLMs miss platform-specific gotchas (ISBLANK on a picklist) — review before you save, because 'no errors found' isn't 'correct.'
- **Iterate, don't one-shot.** Treat generation as a peer-review loop; a few prompts and re-prompts get a complex formula exactly right.


## Metrics mentioned

| Value | Metric | Context |
| --- | --- | --- |
| 20–25% | Jobs at risk from AI (Goldman Sachs estimate) | An analyst estimate Christopher cites for jobs that could be replaced by AI, before countering with how many jobs email created versus destroyed. |
| ~20–30 min | Time to build a complex formula with ChatGPT | A 15–20-line client formula — including a leap-year edge case Christopher didn't know how to solve on his own — built via a few prompts and re-prompts. |


## Entities mentioned

- **LeanScale** (company) — Christopher's and Anthony's employer; Christopher is one of LeanScale's systems architects and demos how he uses ChatGPT in client Salesforce work. · https://leanscale-knowledge-hub.netlify.app/company/leanscale/
- **Amazon** (company) — Used to illustrate that disruptive tech grows industries rather than killing them — email didn't kill paper; paper companies pivoted to the cardboard boxes now shipping from Amazon 'from all over the world.' · https://leanscale-knowledge-hub.netlify.app/company/amazon/
- **Goldman Sachs** (company) — Cited for an analyst estimate that ~20–25% of jobs could be replaced by AI — the figure Christopher counters with the email precedent. · https://leanscale-knowledge-hub.netlify.app/company/goldman-sachs/
- **Christopher Martyen** (person, guest) — Systems architect at LeanScale; an early adopter of ChatGPT for Salesforce admin and RevOps work. · https://leanscale-knowledge-hub.netlify.app/guest/christopher-martyen/
- **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 subject of the whole episode; used to debug a broken formula, generate a Salesforce validation rule from a plain-English request, and translate/explain a complex inherited formula.
- **Salesforce** (tool, CRM) — The CRM Christopher administers; the demos cover validation rules, formulas, picklists, and opportunity stage progression (Nurture → SQL). Called 'already low to no code,' made even more accessible with ChatGPT.


## FAQ

**Q: How can you use ChatGPT as a Salesforce admin?**

A: In three main ways, per LeanScale systems architect Christopher Martyen: debug broken formulas by pasting them in for an instant diagnosis; generate Salesforce validation rules and formulas from plain-English business requirements; and translate or explain existing complex formulas by asking 'what does this formula do?' Each lowers the technical barrier and speeds up admin work, provided a human reviews the output.

**Q: Can ChatGPT write Salesforce validation rules?**

A: Yes. If you describe the rule in plain English — for example, 'people should not be able to move an opportunity to SQL from Nurture unless the Type field has a value' — ChatGPT returns a validation rule using functions like ISCHANGED, PRIORVALUE, ISPICKVAL, ISBLANK, and TEXT, which you can copy into Salesforce. You should still review it for business context and platform edge cases before saving.

**Q: Should you trust ChatGPT-generated Salesforce formulas without review?**

A: No. In the episode, ChatGPT reported 'no errors found' on a validation rule that used ISBLANK on a picklist field, which doesn't work as expected — you need to wrap the field in TEXT. The admin caught it because he knew the Salesforce gotcha. Generated config should always pass through a human who understands the platform's edge cases.

**Q: How does ChatGPT help you understand an inherited Salesforce org?**

A: You can paste an unfamiliar formula and ask 'what does this formula do?' and ChatGPT explains it in plain language — for instance, that a formula adds one year to the closed date to set a term end date and handles leap years. This collapses the research needed to understand configuration someone else built, which is especially useful when onboarding to a new Salesforce instance.

**Q: Does using ChatGPT make Salesforce admin work more collaborative?**

A: Yes. Because rules can be described in plain English and formulas can be translated back into plain English, teammates who don't write code can participate in building and reviewing Salesforce configuration. Christopher Martyen argues that opening technical work up to more people makes the work itself better, with the admin acting as the reviewer.

**Q: Will AI replace RevOps and Salesforce admin jobs?**

A: Christopher Martyen's view is no — the benefits currently outweigh the cons, and AI is more likely to augment technical roles. He draws on the 1990s fear that email would kill paper, which instead grew the paper-and-cardboard industry, to argue that AI, like past technology waves, will enhance jobs and create new work rather than eliminate it.

**Q: What did Goldman Sachs predict about AI and jobs?**

A: Christopher cites a Goldman Sachs analyst estimate that roughly 20–25% of jobs could be replaced by AI. He acknowledges the concern is valid but counters that history rhymes: technology waves like email displaced some work while creating far more, so he expects AI's benefits to outweigh the downsides over the long run.

**Q: How long does it take to build a complex Salesforce formula with ChatGPT?**

A: In one example from the episode, a 15–20-line client formula took about 20–30 minutes using a few prompts and re-prompts — including handling a leap-year edge case that the admin didn't initially know how to solve. The iterative back-and-forth mirrors collaborating with a teammate across a couple of meetings, compressed into minutes.


## Timeline

- **00:00** — Cold open: excited and terrified about ChatGPT
- **01:30** — Why Christopher got into AI — and the '90s email precedent
- **03:14** — ChatGPT as an always-available technical partner
- **04:22** — Use case: plain-English request to a Salesforce validation rule
- **07:28** — Keeping a human in the loop: the ISBLANK catch
- **09:04** — Use case: a leap-year-safe date formula for a client
- **10:22** — 'What does this formula do?' — AI as a translation layer
- **13:22** — Should you be worried? Jobs, Goldman Sachs, and the future


## Related episodes

- **Ep. 95: Why AI Means More RevOps Hires, Not Fewer** (Jimmy O'Halloran) — The mature, full-circle version of this episode's augmentation thesis — AI as a productivity multiplier that should grow the team, not shrink it. · https://leanscale-knowledge-hub.netlify.app/podcast/jimmy-ohalloran-new-relic-revops-consumption-revenue/
- **Ep. 85: Why AI + GTM Engineers Can't Replace RevOps** (Tessa Whittaker) — Extends the 'human stays in the loop' point — AI and engineers execute judgment but don't replace the operator. · https://leanscale-knowledge-hub.netlify.app/podcast/tessa-whittaker-ai-gtm-engineers-revops/
- **Ep. 88: Why AI Won't Close Your Biggest Deals** (Michael Kiernan) — A CRO's take on the limits of AI, pairing with Christopher's insistence on human review and business context. · https://leanscale-knowledge-hub.netlify.app/podcast/michael-kiernan-nextdoor-ai-wont-close-deals/
- **Ep. 6: Why Your Forecast Is Broken** (LeanScale) — Foundational forecasting episode that connects to the contract-date and renewal formulas Christopher automates. · https://leanscale-knowledge-hub.netlify.app/podcast/why-your-forecast-is-broken/
- **Ep. 15: Where Should RevOps Report?** (LeanScale) — Org-and-systems companion to a RevOps-and-Salesforce-focused conversation. · https://leanscale-knowledge-hub.netlify.app/podcast/cameron-legge-where-revops-report/


## Full transcript

_Machine-transcribed and not diarized; speaker attribution is inferred._  
_Transcript only, as a separate file: https://leanscale-knowledge-hub.netlify.app/podcast/christopher-martyen-chatgpt-salesforce-admin/transcript.md_

### 00:00 — Cold open: excited and terrified about ChatGPT

**[0:00]** And just like that, ChatGPT instantly created this validation reform. Salesforce, in general, is already low to no code, but this is something that makes it much more inviting to almost anyone to come collaborate. Welcome to The LeanScale Podcast where we talk about everything RevOps. Thank you for listening. I am really excited to be going through this topic. I don't know if excited is the right word, maybe I'm a little terrified to go through it. I still have mixed feelings about ChatGPT, the power of AI. I know a lot of technology leaders in the space have mixed

**[0:48]** feelings about it, but at the same time, it's still absolutely fascinating. And I really love the applications that it has for revenue operations and some of the work that we do. So today, I have one of LeanScale's systems architects, Christopher Martyen, who's going to walk through some ways. He uses ChatGPT in his daily work with Salesforce. Christopher, thank you so much for being here. Yeah, happy to be here. Happy and excited to talk about such a terrifying topic too. I think we can maybe ease people's minds a bit and shed some light on the future of AI and what ChatGPT does

### 01:30 — Why Christopher got into AI — and the '90s email precedent

**[1:30]** for me and what it can do for you and everyone. You mentioned a word terrifying. I think that's actually one of the reasons that I started to get into working with AI tools like ChatGPT and some other AI tools. The reason I got into it was because I was a little bit terrified of it. Like many people in the 90s were sort of terrified with computers and emailing. A lot of people sort of said that the computers and emails would take away all of the jobs for writing letters and paper. Paper would be dying. Paper would be dead. The computers and email actually enhanced paper so

**[2:04]** much. People began typing everything out and printing everything in math quantities. And paper companies even shifted to cardboard boxes. If you check your house tonight you'll probably get two or three cardboard boxes from Amazon shipped from all over the world. So it can be terrifying but history tends to not necessarily repeat itself but tends to rhyme. I think I've seen this story a few times. Certainly read about it at least. And it really helps augment people's lives and their jobs and enhance them. So hopefully people aren't too terrified. Especially after today

**[2:38]** you'll see how friendly chat GPT can be in fun. So yeah that's pretty much kind of why I started with it. Yeah and I know you are actually using friendly language with chat GPT just in the off chance there's an AI apocalypse coming our way. At least you were one of the friendly ones to the friendly chat GPT bot. Yeah exactly. We even develop a little bit of relationship you know friendly good morning every now and then you know no tasks required so why not you know. A little kindness goes a long way. Of course of course. So maybe walk us through what was the first time you thought hey

### 03:14 — ChatGPT as an always-available technical partner

**[3:14]** I have a problem and I think chat GPT might be able to help me with some of the work that I'm doing in Salesforce. Yeah absolutely. Well chat GPT is a very technical tool but it sounds like a human and in the technical job that we do at Salesforce we often partner with other peers and reviewing each other's work. Maybe if somebody has a problem or a question or wants to learn about something we collaborate. But when somebody that I normally reach out to is in a meeting but I wanted to do something pretty quickly I thought you know there's something just a little bit off about what I'm

**[3:50]** doing right now. I have this big formula but I don't know why it's not working. I threw it into chat GPT. Chat GPT analyzed it and gave me exactly what was wrong with it and then I was able to test it out. So instead of relying on someone else waiting for them to be available and then collaborating on that problem chat GPT was that partner and I think chat GPT and other AI tools can be that partner for a lot of people in technical roles and any job in general. That's great. That's great. Well I know you have a couple things you want to share with the listeners and viewers today.

### 04:22 — Use case: plain-English request to a Salesforce validation rule

**[4:22]** I'm really interested to go through these and I think it's going to prompt a lot of discussions. I hope people watching this also if you have your own ideas of how you've used chat GPT and this type of work it'd be interesting to add that as well. But Christopher what's one of the problems chat GPT has helped you solve? One of the problems it's helped me solve is working and collaborating with other people. You just mentioned that you wanted to get more information from people that might be watching or listening to this on how they use AI tools and that's exactly what I wanted to do.

**[4:57]** The more people that we can get into collaborating together at work in general but specifically in technical tools opening up a lot of these things like code and formula to people that might not be able to understand it or don't really have the interest in doing that the more people that we can get to collaborate the better the work becomes. So for example I'm going to use chat GPT here up on the screen and I'm going to just tell it some human language really and it's going to do exactly a little bit of what I want. So I'm going to paste a message in there that says hello hope you are having

**[5:40]** a nice day just for a little bit of kindness. I need help with the Salesforce validation rule. People should not be able to move an opportunity to the stage SQL from the stage nurture unless the field type has a value in it. So this is something that I get pretty often in the technical role. People will say hey I want to add this to Salesforce. I want to make this requirement and they'll use plain simple human language like this but they don't know how to do the technical back end. So I'm going to send that to chat GPT. Chat GPT is thinking because hello I'm happy to

**[6:15]** didn't ask about my day but that's okay I forgive him. This is a workhorse. Here's a validation rule that will prevent people from moving an opportunity stage SQL to nurture unless the field type has a value in it. And as you can see here the majority of people don't know how to write this type of formula. They don't know how to use this type of punctuation. They don't know these functions for example is changed is pick val prior value is blank text and all these parentheses. For most people it looks boring. Chat GPT also explains it but with this validation rule I can simply

**[6:46]** click on copy code and I have a test Salesforce validation rule right here. I'm going to throw it in here and as I'm looking at this as a human it of course typed this out faster than I would manually type this entire formula out. I do see that something might be a little off and it says no errors found. What I was thinking is that is blank text type type is actually a pick list value so is blank won't work. That's why it threw text in there. I will tell you that before this demo I did try this and it had is blank type and I told it no that has an error please correct it

### 07:28 — Keeping a human in the loop: the ISBLANK catch

**[7:28]** and it corrected it. So I think maybe it learned from this morning. Wow. So now that it threw that

**[7:36]** formula into there I can choose save and if I were to move an opportunity like this from nurture

**[7:48]** into SQL and Mark is the current stage it will tell me chat GPT says you cannot move forward unless you fill out type. Thank you chat GPT. So then I'll come in here and say great thanks love you and just like that chat GPT instantly created this validation for me. Salesforce in general is already low to no code but this is something that makes it much more inviting to almost anyone to come collaborate. So then now that I'm able to open this up to other people that might not have written this code manually like I was doing mostly in the past they're able to

**[8:25]** create this and I can quickly review it for them. We can collaborate together on how to make this code better or more efficient or add things and just like that it created the validation before me instead of me doing it manually. In reality I would save a lot of time doing that and able to collaborate with others that might not normally do this on their own. Yeah that's I think for most people leveraging AI is either going to help you do something faster or it's going to help you solve something more complex by giving you inspiration of potential solutions. I think this is a this is a great example. Clearly it's pretty quick to just ask a question

### 09:04 — Use case: a leap-year-safe date formula for a client

**[9:04]** copy paste it in put the human element of making sure it fits the business context of of your problem and then and then off you go. This is an excellent example. What's another way you've used chat GPT? I've used chat GPT in a number of ways. I have another example right here that I wanted to show you. This is from something that I've done a few days ago. Actually for one of our clients and this took about I'd say 15 20 about 20 or 30 minutes with chat GPT and a few prompts and re-prompts to get it to do exactly what I wanted. Just like you would collaborating

**[9:47]** with the team you might have one two or three meetings you might hear review back and forth etc and after I had that it created this formula for me. It's about I don't know 15 20 lines and it was something that I didn't know how to do on my own after I ran into a problem and chat GPT helped me learn about it and then also right now it's translating it. So I then recently just copied and pasted the big formula in here and I said what does this Salesforce formula do and I threw it in here and it explained to me that it says if the term end date is blank it calculates the day by

### 10:22 — 'What does this formula do?' — AI as a translation layer

**[10:22]** adding one year to the closed date field etc etc. Now the problem I was running into and some people might want to double check their calendars their forecasting maybe some of their own formulas is that next year is actually a leap year so if you're using if you're counting days weeks months and February is going to have one last day in it that could throw a little bit of a wrench into some of your plans so what I needed to do was calculate a specific date but I was running into a lot of trouble with the leap year and chat GPT has this for me here and it created if is blank

**[10:56]** populate the record term end date using based on the closed date it'll take whatever the month is and for example you see one comma 31 and three comma 31 that's January and March both have 31 days four being April has 30 and then number two it says two comma if mod year record closed date comma four this is all a bunch of technical jargon that most people don't even care about but chat GPT was able to throw it in and it can explain exactly what it does so that if the year is divisible by four it will know what if it's a leap year or not and then it will calculate

**[11:31]** the date perfectly instead of being off by one day so by asking it what does this formula do that's essentially what it does and it breaks it down here for me for February return 28 or 29 depending on whether the year of the closed date field is divisible by four so that in contrast to to comma if mod year dollar sign record dot closed date most people get snoozefest when they read that line but reading this you can easily see what it does so again kind of circling back to the collaborative effort this tool helps translate things history doesn't necessarily

**[12:10]** repeat but it does rhyme when we got translation tools we're able to collaborate across borders like and no other time in history and now with chat GPT and other AI tools we can translate a lot of this technical details so fast so quickly and people are able to collaborate on it pretty quickly so it helps me create that formula but it also helped me explain that to somebody else without me taking 20 minutes to type up this page explaining it yep they're able to throw it in in a few seconds yeah and see the human language behind it and that makes a lot of sense i mean

**[12:43]** every time we run into an instance of sales force there's usually a complete labyrinth of workflows and validation rules and things that we need to do our own research on to even begin understanding what was implemented before we begin to do our work so you know getting faster solving more complex problems and then doing research on something that already exists and learning excellent use cases for using this chris for this this was really awesome i think um i know you're going to have more uh examples to share in the future and it's only going to get better and better

### 13:22 — Should you be worried? Jobs, Goldman Sachs, and the future

**[13:22]** over time um anything you're worried about by using chat GPT to do work in sales force i'm really not worried about chat GPT doing work in sales force or other AI tools and technical programs i think the benefits just at this time outweigh the cons and i think for a while they will um again back to the analogy that paper in the 90s a lot of people were terrified that letter writing would be dead with email i think we have some colleagues down in brazil i much prefer to email them today than to write a letter waiting a couple weeks to get a response um just today on the radio

**[14:01]** i was listening on the drive over here and i heard that some analysts from Goldman Sachs predicted that they about 20 to 25 of jobs would be replaced by AI well how many jobs were replaced by email though we gained a lot more jobs um with technology as we grow so i think that if anyone's worried or concerned they have a right to be and that is good but i think the benefits will outweigh the cons in the long run i tend to agree i know i have a slight feeling of terror when i look at it just because it seems so powerful but um i agree i think the future is bright chrisford thank

**[14:42]** you so much i really appreciate it i think we're gonna have to do a refresher of this in the future to see what other tips and tricks you've come up with um and i'm looking forward to it thank you so much at this rate probably next week thank you for having me thank you chrisford thank you for listening to this episode if you like the discussion please like share and subscribe to wherever you listen to podcasts so you never miss a new episode


---

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