5/28/26: Your Top Rep Just Quit. Their $30M Brain Walks Out. AI saves the day!
Welcome again we got one hot off the presses regarding some amazing GTM AI-ness.
As a reminder, we have a TON of free options for you to subscribe to:
GTM AI podcast and newsletter with associated assets, apps, AI tools, etc
We also have Signal vs Noise which is more of a short and to the point update that is important for you to know.
On the paid side, we go DEEP on new AI tools, deep dives on how to use tech, and give you up to date education for GTM pros on how to use AI in your role. We would love for you to subscribe on any level that best for you.
Free or Paid, regardless of what you want to do, we will always shower you with as much valuable content as possible. Now lets get into it
You can go to Youtube, Apple, Spotify as well as a whole other host of locations to hear the podcast or see the video interview.
One of Fluint’s customers recently watched their top sales rep walk out the door. The same rep who personally closed a $30 million land contract. The single largest deal ever closed using their product. Two weeks notice. Gone.
Every signal that rep read before anyone else. Every implicit cue from a procurement email. Every gut-feel decision on when to hold ACV versus when to discount. All of it walked out with him.
This is the most expensive problem in enterprise sales right now, and I do not see it discussed honestly enough. So this week, I sat down with Nate and John, co-founders of Fluint, on the GTM AI Podcast. Nate is a repeat enterprise sales leader. John is the technical co-founder who built the architecture that captures what Nate could never document.
What came out of the conversation is, I think, the cleanest take I have heard on how to actually solve the “clone your top rep” problem after 15 years of GTM tech promising and never delivering. Four insights worth your attention.
1- The implicit signal example that explains everything
Nate gave this example, and I have not stopped thinking about it.
A rep runs two POC readouts on the same Friday. Two different prospects. Two procurement teams. Team A’s procurement emails the rep back at 9pm Friday night. Team B’s procurement says “we will keep this moving next Wednesday during business hours.”
Same outcome on paper. Both are “moving forward.” Both get coded the same in the CRM. But a top rep is going to negotiate completely differently in those two scenarios. Team A is screaming urgency. The exec could not wait. Friday 9pm follow-up means do not discount, they need this now. Team B is a normal business cadence. You can afford to play the commercial game.
Here is the issue: nobody trained the rep to read that signal. They picked it up over a decade of pattern matching. They cannot tell you they’re doing it. And every average rep on the team would miss it entirely, walk into Team A’s negotiation, and start discounting on instinct.
“That rep who reads that, they’re gonna hold onto ACV at the finish line versus somebody else is gonna miss that signal. So the question is, how do you capture that with AI?” Nate, Fluint
The implicit-signal example is the thesis of the whole episode. Tacit knowledge is the unfair advantage of your top reps. And until now, nobody has had a way to capture and redistribute it.
2- The photo vs. video architecture (the shrimp tacos moment)
This is where John, the technical co-founder, walked through the architectural decision that makes Fluint different from every other AI sales tool I have evaluated.
The conventional approach to AI agents is to give an LLM a giant context window: the CRM record, the call transcript, the email thread, all of it stuffed into one prompt. The agent then produces a summary or recommendation.
Nate’s framing for why this fails:
“Anytime you’re giving an LLM context, it’s like a snapshot. It’s like a photo, and it can write something. Think of it like a caption. A video file is gigabytes worth of data, and the whole meaning of a video comes from the motion in between those frames.” Nate, Fluint
Top reps do not pattern-match on photos. They pattern-match on motion over time. What was the deal doing 30 days ago vs. today? When the prospect went quiet for 4 days and then sent a flurry of emails about pricing, what did that combination mean? The signal lives in the change, not in any single moment.
Fluint’s architecture is event-driven and time-series-based. Every interaction gets tagged, timestamped, and stored so the system can reconstruct the motion of a deal, not just its current state. That is what lets ML models pick up patterns that no LLM-with-big-context-window can.
If you are evaluating AI sales tools right now, this is the single question to ask vendors: “Are you snapshot-based or time-series-based?” Most will not understand the question. Move on from those vendors.
3- AI is not LLMs. ML + LLMs is the actual stack.
John dropped a line that I want every GTM leader to internalize:
“LLMs are like a subcategory of AI which is a subcategory of machine learning. So it’s just this small little sliver of everything that we’re doing.” John, Fluint
The entire industry right now talks about “AI” when they mean “LLMs.” Those are not the same thing. ML has been solving pattern-recognition problems for decades. LLMs are good at human-to-human interaction and language generation. They are bad at retaining state across long conversations and bad at picking up subtle pattern shifts in large datasets.
John’s analogy for why LLMs alone cannot do this work:
“Have you ever seen the movie Fifty First Dates? There’s 10-Second Tom. ‘Hi, I’m Tom. Hi, I’m Tom.’ That’s what an LLM is good at. You have a very short window of time to talk to me before you lose everything that’s there.” John, Fluint
Fluint uses ML to do the pattern recognition (the judgment layer) and then passes the structured output to an LLM for the human-facing communication. Using the right tool for the right job.
This is the architectural decision that separates real enterprise AI from prompt-wrapped novelty. If your AI strategy is “we use ChatGPT and call it AI,” you have not built anything. If your AI strategy is “we use ML to capture pattern signals, then LLMs to communicate them to humans,” you have an actual system.
4- The “no perfect data” answer that ends the AI readiness debate
Nate asked John the question every GTM leader is asking themselves right now: “What if our data is too messy to start? Should we wait?”
John’s response was the cleanest answer I have heard:
“Data is a product of people. Humans are imperfect by nature. So it’s a logical fallacy that you could have perfect data because its originating source is imperfect. If you know what’s wrong with your data, that’s how you use it. That’s what turns it into knowledge.” John, Fluint
Then he walked through the racehorse model, which is genuinely smart. Fluint runs two models in parallel for every customer: a global baseline model trained on aggregated industry patterns, and a customer-specific model trained on that customer’s data. The two race nightly. Whichever produces better predictions on nearest-neighbor analysis gets promoted that day. The custom model wins eventually as it accumulates training data, but the customer gets value from day one because the global model is already trained.
The implication for your team: stop using “our data is not clean enough” as an excuse to not deploy AI. The data is never clean. The question is whether your tooling knows how to work with imperfect input. If the answer is yes, ship. If the answer is no, you need different tooling.
Real outcomes
Three numbers from the episode worth memorizing:
+$28K added to ACV per team per year (average, enterprise context)
32 days trimmed off median sales cycle time
Maturity curve: Q1 = win existing deals with less discount, Q2 = compress cycle times, Q3+ = win deals you would have lost
That progression matters. Most teams want AI to deliver Q3 results in week 2. Fluint’s curve says the highest-leverage wins happen second half of year one, after the model has had time to learn your patterns. Set expectations accordingly.
Why this matters more than usual right now
Every enterprise sales team is sitting on a hidden time bomb. Your top reps’ tacit knowledge represents the difference between 17% win rates and 50% win rates. None of it is documented. None of it is in the playbook. All of it walks when they take a recruiter call.
The teams that capture it first will compound their advantage every quarter. The teams that don’t will keep paying $5K to $50K per opportunity to generate pipeline that average reps then squander at half the win rate of their top performers.
This is the actual GTM AI play of 2026. Not “automate the BDR sequence.” Not “summarize the call.” Capture the tacit knowledge before it walks out the door.
My challenge to you this week
1- Name your top rep. Out loud, on paper, by name. 2- List 3 things that rep does that no one else on the team does. 3- If you cannot list the 3 things, that is the audit failing in real time. Start a 30-minute conversation with that rep this week. Ask them to walk you through their last 3 deals. Capture what they say.
That’s the smallest possible first step toward solving this. The full 5-step protocol is in this week’s free field guide: The Tacit Knowledge Audit below:
I hope this lands. More to come next week.
The Tacit Knowledge Audit
A 5-step protocol to capture what your top sales reps do unconsciously, before they leave.
A GTM AI Academy field guide. Inspired by my podcast conversation with Nate and John, co-founders of Fluint.
A few weeks ago, one of Fluint’s customers lost their top sales rep. He personally closed a $30 million land contract for them. The largest deal ever closed using the product. He gave two weeks notice and walked.
Every signal he caught before anyone else. Every implicit cue from a procurement email. Every gut-feel decision on when to hold ACV and when to discount. All of it walked out with him.
If you have led a sales team for more than two years, you have lived this moment. The recruiter call you saw coming. The exit interview where the rep tried to put it into words. The 90 days after, where the team’s win rate quietly dropped and nobody could quite say why.
This is the most expensive problem in enterprise sales right now. And almost nobody is solving it honestly. Most playbooks, enablement programs, and “AI for sales” tools attack the wrong layer. They try to document the explicit. They cannot get at the tacit.
I sat down with Nate and John of Fluint on the GTM AI Podcast this week. Nate is a repeat enterprise sales leader. John is the technical co-founder who built the architecture that actually captures what Nate could never document as a leader. What I learned from them, combined with my own pattern-matching across 13 years in enablement, became this 5-step audit.
This is the test I now run on every enterprise sales team I work with. Whether you are a CRO, a VP of Sales, a head of enablement, or a founder selling your own product, the 5 steps below give you a defensible read on how much of your top performers’ knowledge is captured vs. how much is one resignation letter away from gone.
Let’s go.
Why this matters more than it sounds
A small pocket of reps wins 50 to 60 percent of their pipeline. Everyone else sits at 17 to 19 percent. Same product. Same comp plan. Same playbook. The math is not subtle.
The difference is not training. Not coaching. Not effort. Not even talent in the abstract sense. The difference is tacit knowledge: the unconscious patterns top performers cannot even articulate. They just do things. The signal is invisible to them because it has been automated by their brain over a decade of reps.
Nate gave this example on the podcast, and it is the cleanest illustration of tacit knowledge in sales I have ever heard:
A rep runs two POC readouts on the same Friday. Two different prospects. Two procurement teams. Team A’s procurement emails the rep back at 9pm Friday night. Team B’s procurement says “we will keep this moving next Wednesday during business hours.”
Same outcome on paper. Both are “moving forward.” Both get coded the same in the CRM. But a top rep is going to negotiate completely differently in those two scenarios. Team A is screaming urgency. The exec could not wait. Friday 9pm follow-up means do not discount, they need this now. Team B is a normal business cadence. You can afford to play the commercial game.
That signal lives in a 4-hour timestamp difference on a procurement email. No CRM field captures it. No playbook explains it. No call transcript will surface it. A top rep notices it without thinking. An average rep walks into Team A’s negotiation and starts discounting because “the deal moved forward.”
The cost of that single missed signal across a year is six figures of leaked ACV per rep. Multiplied across your sales org, it is the difference between hitting plan and missing it.
Tacit knowledge is the unfair advantage of your top reps. It is the moat under your revenue. And until very recently, nobody had a way to capture and redistribute it.
That is changing. But the change requires you to do the audit first.
The 5-step Tacit Knowledge Audit
Each step has three parts: the question, the diagnostic test, and the action. Work through them in order. Do not skip ahead. Each step assumes the previous one is complete.
Step 1: Name the top rep, list the three things
The question: Can you name your top performer and list three specific things they do that no one else on the team does?
The diagnostic test: Right now, out loud, name the rep. Then write down three behaviors. Not “they’re great at discovery” (too vague). Closer to: “They always email the buying committee individually after a group call with one specific personalized observation about their function.”
If you cannot list three specific behaviors within 60 seconds, you have not actually been studying your top rep. You have been admiring them.
The action: Schedule a 30-minute call with the rep this week. Frame it not as a review but as a curiosity conversation. “I want to understand how you think about deals.” Take notes. Look specifically for the things they do that they think are “obvious.”
The 30-minute curiosity conversation script:
Most leaders skip this conversation because they do not know what to ask. Use this script verbatim. Adjust as needed.
Opening (2 minutes):
“I want to spend 30 minutes understanding how you think, not how you execute. The whole team learns from you whether you realize it or not, and I want to make sure I am capturing the right things to teach the rest of the team. There are no wrong answers. I am taking notes.”
Question 1 (5 minutes):
“Walk me through the last deal you closed. Not the highlights. The texture. When did you first sense it was going to close? Was there a specific moment? What did you see?”
Question 2 (5 minutes):
“Walk me through the last deal you lost or stalled. Same question. When did you first sense it was going sideways? What was the signal you wish you had acted on sooner?”
Question 3 (5 minutes):
“When a buying committee goes quiet for a week, what do you usually do? Walk me through your decision tree out loud.”
Question 4 (5 minutes):
“Tell me about a deal where your gut told you one thing and the playbook told you another. Which one did you follow? What happened?”
Question 5 (5 minutes):
“If I gave you one new rep to onboard, and you only had 60 minutes to teach them one thing that would make them better, what would it be? Why?”
Closing (3 minutes):
“Last question. If you left tomorrow, what is the one thing I would lose access to that I would not even know I had lost? That is the thing I most need you to help me capture.”
Take notes by hand or recording (with their permission). Do not edit during the conversation. The goal is capture, not synthesis. Synthesize after.
Why this is step 1: Everything that follows depends on you having a specific, observed top rep to anchor on. Without that, you are doing enablement in theory. With it, you are doing enablement with a target.
Step 2: Find the implicit signals
The question: What signals does your top rep notice that the rest of the team misses?
The diagnostic test: Pull the call recordings, email threads, and CRM updates from your top rep’s last three closed deals. Then pull the same artifacts from your average rep’s last three deals (won or lost). Read them side by side.
You are looking for moments where the top rep’s behavior changed because of a signal that is not in any structured field. Examples:
The procurement timing example above
A buying committee member who went silent for a week, then suddenly engaged with a specific question
A champion who started cc’ing a different stakeholder on follow-ups
A delay in a follow-up that was framed as “scheduling” but actually meant internal misalignment
The top rep responds to these. The average rep does not see them.
The action: Document at least 5 implicit signals your top rep responds to. For each one, write down: what the signal looks like, what the top rep does in response, what the average rep does instead, and what the financial consequence is of missing the signal.
Why this is step 2: This is where the explicit playbook breaks down. The signals top reps catch are not in any documented process because nobody could see them well enough to document them. Once you list them out, you have created your first “tacit-to-explicit” translation. That is real enablement IP.
The 5 most common implicit signal patterns (a starter library):
These are the patterns I see top reps catch over and over again, across industries, across deal sizes. Use this list as a starter. Add your own as you find them.
1- The timing tell. Procurement responds at 9pm Friday vs. next Wednesday. A champion who normally replies within 4 hours suddenly takes 36. A buyer who emails on weekends. Time-of-response carries information the content does not.
2- The CC pattern shift. Your champion was cc’ing two people. Now they are cc’ing four. Or: a new name appears on the thread without explanation. Or: a previously-cc’d executive drops off. Each shift means something. Top reps notice the change, not just the current state.
3- The vocabulary drift. A buyer who used “we are evaluating” suddenly says “when we go live.” Or the reverse: “this looks great” becomes “we have some concerns.” Top reps track linguistic shifts. The words change before the deal status does.
4- The post-call silence gap. Top reps know what “normal silence” looks like after each meeting type. 24 hours of silence after a demo is fine. 48 hours of silence after a pricing conversation is bad. The expected gap is contextual. The deviation from the expected gap is the signal.
5- The reverse pull. When a buyer starts asking questions YOU should be asking (”How quickly can your team implement?” “What does onboarding look like?” “Who is our day-one contact?”), the deal has shifted. Average reps answer the questions. Top reps recognize the shift and accelerate.
Add these to your own pattern library. Each one represents tacit knowledge that has been made explicit. The act of writing them down is the audit producing IP.
Step 3: Capture the motion, not the moment
The question: Are you capturing how your deals change over time, or just where they are right now?
The diagnostic test: Open your CRM. Look at a current deal. Can you reconstruct what happened in week 1 vs. week 4 vs. today? Can you see the motion of the deal, or only its current state?
For most teams, the answer is “current state only.” Stage changes are tracked, but the texture of how a deal moved through those stages is gone. CRM snapshots overwrite each other.
This is the same problem Fluint’s John talked about on the podcast. As he put it:
“Anytime you’re giving an LLM context, it’s like a snapshot. It’s like a photo. A video file is gigabytes worth of data, and the whole meaning of a video comes from the motion in between those frames.”
Top reps pattern-match on motion. They notice that a deal that took 3 weeks to get to demo and then closed in another 2 weeks behaves completely differently from a deal that took 6 weeks to get to demo and then sat for a month. Average reps look at the current stage and assume.
The action: Pick one deal type (e.g., enterprise renewals, or new-logo mid-market). For that deal type, define 5 to 7 events that should always be captured with a timestamp: first meeting booked, demo run, ROI doc shared, mutual action plan signed, procurement engaged, security review opened, contract sent, signed. Then capture them. Even in a spreadsheet at first.
Once you have the motion data, you can start asking “what did the average winning deal look like at week 4?” and use that as a guide.
Why this is step 3: You cannot capture tacit knowledge from snapshot data. The signal lives in the change between frames. If your data infrastructure is photo-based, you have already lost the most valuable inputs.
Step 4: Build the feedback loop into the daily workflow
The question: When your top rep does something brilliant on a deal, how does the rest of the team learn from it within a week?
The diagnostic test: Think about the last time your top rep did something unusual on a deal that worked. (Not a documented best practice. A nuanced judgment call.) How did the rest of the team find out about it?
For most teams, the answer is: they didn’t. Or they heard about it in a deal review weeks later. Or they noticed it once in a shadow session. The institutional knowledge transfer rate is approximately zero in real time.
This is where Fluint’s “Ollie deal room” approach is worth borrowing even if you do not use their tool. They run a public Slack channel where every morning their AI surfaces deal changes across the team: who used what move on what deal, what worked, what is stuck. It is the public ledger of tacit knowledge being captured and redistributed in real time.
The action: Pick one of these three minimum-viable feedback loops and run it for 30 days:
1- The Friday “moves of the week” Slack post. Every Friday, the sales leader posts 3 specific nuanced moves a rep made that week that the rest of the team should learn from. Name names. Show emails. Show the outcome.
2- The weekly 15-minute “weird deal” review. Each rep brings the one deal that had a weird inflection point this week. The group spends 15 minutes pattern-matching across the cases.
3- The “what would top rep do” spot check. When a deal hits a known inflection point (e.g., procurement involved), the rep posts the situation in a Slack channel and asks: “What would Cora do here?” Top rep responds in writing. Knowledge captured.
Any of these beats zero. All of these compound.
Why this is step 4: The capture is meaningless without redistribution. Tacit knowledge does not transfer through documentation alone. It transfers through proximity to people doing it, made visible.
Step 5: Stop waiting for clean data
The question: What is the excuse you have used most often for not building the system above?
The diagnostic test: Be honest with yourself. Did “our CRM data isn’t clean enough” come up in your head while reading steps 3 and 4? Did “we need to wait until RevOps finishes the data project” come up?
If yes, John’s answer from the podcast is your answer:
“Data is a product of people. Humans are imperfect by nature. So it’s a logical fallacy that you could have perfect data because its originating source is imperfect.”
Perfect data does not exist and will not exist. The question is not whether your data is clean. The question is whether you are doing something useful with what you have.
The best AI sales tools in 2026 are built on top of imperfect, human-generated data. They use ML to find signal in noise. They get smarter with use. Fluint’s racehorse model (a global baseline and a customer-specific model racing nightly, the winner promoted) is one example of how modern systems handle imperfection by design.
The action: Pick the single highest-leverage tacit knowledge capture (from steps 1 through 4) you have been delaying. Ship a v1 of it this week. Imperfect, ugly, partial. Ship it.
The reason: every week you wait for clean data is a week your top rep is closer to leaving. The cost of “let’s wait” is the brain drain you cannot recover from.
Why this is step 5: The audit is only useful if it ends in action. Most audits end in a slide deck. This one ends in a Friday Slack post, a 30-minute conversation, or a spreadsheet you started today. Choose action.
Scoring the audit
Score yourself 1 (not at all) to 5 (fully done) on each step:
Step Description Score (1-5) 1 Named top rep + 3 specific behaviors 2 Documented 5+ implicit signals 3 Capturing motion (event-driven), not just snapshots 4 A weekly tacit-knowledge feedback loop running 5 Shipped a v1 of capture without waiting for clean data TOTAL /25
20-25: You are ahead of 95% of GTM teams. Keep going. Layer on AI tooling that thinks in time-series (like Fluint or comparable architectures).
14-19: You have the right instincts but the system is incomplete. The most common failure mode in this range is good capture, weak redistribution. Fix step 4.
8-13: You are doing pieces but not the system. Most likely you have a top rep you admire but have not formalized any of the capture. Run step 1 this week.
Under 8: Your top rep is one resignation away from a quarter of pain. Start with step 1 today.
5 anti-patterns that kill tacit knowledge capture
I have seen these failure modes more times than I can count. Each one looks reasonable in isolation. Each one quietly sabotages the work.
1- Treating it as a one-time project, not a system. “We are going to capture tacit knowledge in Q2” never works. The capture has to be embedded in the weekly cadence (deal reviews, Slack posts, retrospectives) or it dies the moment the project sponsor moves on.
2- Capturing without redistributing. Some teams do excellent capture (transcripts, recordings, deal review notes) and then no one ever reads any of it. Capture without redistribution is a graveyard. The Friday Slack post or the 15-minute weekly review is what makes capture into transfer.
3- Asking your top rep to “write it up.” Top reps cannot articulate their tacit knowledge in writing. That is the entire definition of tacit. If you ask them for a “best practices doc,” you will get generic advice. You need conversation, observation, side-by-side review of actual artifacts. Then YOU write it up.
4- Letting the top rep become the bottleneck. Once you start using the top rep as the source of truth for everything, you have made the brain-drain problem worse, not better. The whole point is to distribute what they know across the team faster than they could ever do it themselves.
5- Mistaking explicit playbooks for tacit knowledge transfer. A 60-page sales playbook with everyone’s favorite frameworks is not what we are talking about. Tacit knowledge transfer is “Cora pinged the CFO on Wednesday after her champion went silent for 4 days, and the deal moved.” Specific, named, in-context, recent. The 60-page playbook is what teams produce when they are afraid to do the actual work.
If you find your audit work drifting toward any of these patterns, stop and re-center. The point is transfer, not artifact creation.
The 90-day implementation roadmap
You do not need a quarterly project plan to do this work. You need a 12-week cadence. Here is the calendar-ready version.
Weeks 1-2: Foundation
Week 1: Run step 1. Name top rep + 3 behaviors. Schedule the 30-minute curiosity conversation. Do step 5: pick the one thing you have been delaying and ship a v1 this week (could be a Slack post template, a deal review structure, a shared notes doc).
Week 2: Run the curiosity conversation. Take notes by hand. Within 24 hours of the conversation, write down 5 things you heard that you did not know before.
Weeks 3-4: Signal capture
Week 3: Run step 2. Pull artifacts from your top rep’s last 3 wins and your average rep’s last 3 deals. Read side by side. Document 5 implicit signals.
Week 4: Share the 5 implicit signals with the whole sales team in a Friday Slack post. Frame them as “things Cora notices that you might be missing.” Watch which signals generate the most discussion. Those are the ones to double-click on next.
Weeks 5-6: Motion infrastructure
Week 5: Run step 3. Pick one deal type. Define 5-7 events that should always be captured with timestamps. Even if you have to do it in a Google Sheet, do it.
Week 6: Start backfilling for the last 30 days of deals in that type. You are building a time-series record by hand. Painful but irreplaceable.
Weeks 7-8: Redistribution rhythm
Week 7: Run step 4. Launch one of the three feedback loops (Friday “moves of the week” Slack post, weekly 15-minute “weird deal” review, or “what would top rep do” Slack channel).
Week 8: Run the loop a second week. Pay attention to who engages and who does not. Engagement is your leading indicator.
Weeks 9-10: Pattern library
Week 9: Synthesize what you have captured. Build a “pattern library” of 10-15 tacit knowledge moves with concrete examples. This is internal IP. Treat it like product.
Week 10: Run a 30-minute team session walking through the pattern library. Have your top rep validate each pattern in their own words.
Weeks 11-12: Tooling decision and re-score
Week 11: With 10 weeks of pattern data in hand, evaluate AI tools that fit the architecture you have built (time-series, event-driven, ML + LLM stack). Avoid snapshot-only tools regardless of how good the demo looks. Reference the podcast for the architectural questions to ask vendors.
Week 12: Re-run the audit (score yourself out of 25). Compare to your starting score. Decide what to focus on for the next 90 days.
This roadmap is opinionated. Adapt it to your team’s reality. The key principle: do not skip step 1, do not delay step 5, and do not let any week pass without something concrete shipping.
Three things to do today
If this audit lands and you want to move on it without overcomplicating, do these three things in this order:
1- Today: Name your top rep. Write down 3 specific behaviors they do that no one else does. Time-box this to 15 minutes.
2- This week: Schedule the 30-minute curiosity conversation with that rep. Use the script above. Take notes specifically on what they describe as “obvious.”
3- By month end: Pick one of the three feedback loops from step 4 and run it for 30 days.
That is it. You do not need new tooling. You do not need a budget approval. You do not need a board sign-off. You need 90 minutes and the willingness to ask better questions.
The bigger picture
The next 18 months of enterprise sales will be won by teams who solve two things at once: capturing the tacit knowledge of their top reps before it walks, and building feedback loops fast enough that the rest of the team gets better in days, not quarters.
Most teams will not do either. They will keep paying $5K to $50K per opportunity to generate pipeline that average reps then squander at half the win rate of their top performers. They will keep losing $30M-deal-closers to recruiters. They will keep wondering why “AI for sales” has not delivered the productivity gains the keynotes promised.
The teams that move first on this will compound their advantage every quarter. By 2027 it will look like luck. It will not be luck.
My challenge to you this week
I want you to do one thing.
Name your top rep, out loud, by name. Then list three things they do that no one else on your team does. Be specific. Not “great at discovery.” Closer to: “They always send a personalized note to each buying committee member individually within 4 hours of a group call, referencing a specific question that person asked.”
If you cannot list three specific things in under 5 minutes, that is the audit failing in real time. That is your top rep’s brain being closer to walking out the door than you realized.
The good news: now you know. You can start the conversation today.
I hope this lands. If you do the exercise and want a second opinion on what you wrote down, reply to this post and I will read it. The point of the audit is what comes next, not the score itself.
If this field guide helped, the highest-leverage thing you can do is share it with the one sales leader on your team who is making AI roadmap decisions this quarter. Forward the email. Drop the link in your Slack. The point of an audit is that more people run it.
Subscribe to get the next field guide in your inbox. Listen to the full Nate and John episode on the GTM AI Podcast wherever you get your podcasts. And if you want to go deeper on the architecture (event-driven, ML + LLM stack, racehorse model evaluation), John’s recent blog post at fluint.io/blog is the technical companion to this strategic piece.
Appendix: Printable one-page scorecard
Print this section, take it into your next leadership team meeting, fill it out together.
THE TACIT KNOWLEDGE AUDIT · SCORECARD
═══════════════════════════════════════════════════════════
Date: _________________________ Team / Org: _________________________
Scored by: _________________________
TOP REP NAME: _________________________
3 THINGS THIS REP DOES THAT NO ONE ELSE DOES:
1) _____________________________________________________________
2) _____________________________________________________________
3) _____________________________________________________________
───────────────────────────────────────────────────────────
SCORE EACH STEP 1 (NOT AT ALL) TO 5 (FULLY DONE)
───────────────────────────────────────────────────────────
STEP 1 Named top rep + 3 specific behaviors [ /5]
STEP 2 Documented 5+ implicit signals [ /5]
STEP 3 Capturing motion (event-driven), not snapshots [ /5]
STEP 4 Weekly tacit-knowledge feedback loop running [ /5]
STEP 5 Shipped v1 without waiting for clean data [ /5]
─────────────────
TOTAL [ /25]
───────────────────────────────────────────────────────────
VERDICT
───────────────────────────────────────────────────────────
[ ] 20-25 Ahead of 95% of GTM teams. Layer on time-series AI.
[ ] 14-19 Right instincts, incomplete system. Fix step 4.
[ ] 8-13 Doing pieces, not the system. Run step 1 this week.
[ ] <8 Top rep is one resignation away from brain drain.
Start step 1 TODAY.
───────────────────────────────────────────────────────────
3 ACTIONS, NEXT 4 WEEKS
───────────────────────────────────────────────────────────
THIS WEEK: _____________________________________________
NEXT 2 WEEKS: _____________________________________________
BY MONTH END: _____________________________________________
───────────────────────────────────────────────────────────
RE-SCORE DATE (90 days out): _________________________
───────────────────────────────────────────────────────────
The Tacit Knowledge Audit · GTM AI Academy · Coach K
Save the filled scorecard. In 90 days, re-score and compare. The delta is your audit’s ROI.