5 Things We've Learned About AI Data Readiness

Opening — The Million People

Trisha Merriam has a way of setting up the whole show in one sentence. On May 6, with Erin Wiggers joining and Chris Carolan opening the recap, she pointed at the obvious thing nobody else was naming.

For the million people who haven't seen any of this — yeah, is it time for a recap?

Six months. Fourteen recorded episodes. Three voices showing up most weeks, plus the people who walked through the door on the way. The show began in November 2025 with Bill Barlas on data readiness through empowerment, and it has not stopped since. What started as a conversation about what to clean up before AI arrives has turned into a working theory of how an organization actually gets ready — and how it does not.

This article is the start-here for anyone who hasn't watched a single episode. It is also the synthesis for anyone who has watched all of them and wants to see the pattern from above. Five things have surfaced. They surface every time, in every conversation, across every shape of organization the team has worked with.

A note on the count, since Chris already corrected it on air: fourteen episodes recorded, seven with full notes. Not thirty. Claude got excited.

Six months in, here is what has shown up — every time, across every conversation.

1. The Unified Customer View Is a Hack, Not a Destination

Most "data integration" conversations spiral into vendor selection, integration cost, and a six-month wait. The Unified Customer View — UCV, as the show calls it — does something different. It reframes the question. Instead of asking "which systems do we connect and in what order?", it asks "what does the team need to see when they look at this customer?"

The first question is unanswerable in a session. The second one is answerable in an afternoon.

Chris named it on the recap with a phrase that lands harder the more you think about it: "The unified customer view is kind of a hack. It's a way to translate. If you want unified context or a data source of truth, there's so much other stuff that gets in the way."

The hack is the point. UCV is not the destination. It is the shortcut to the destination that organizations have been trying to reach for decades and never could — because they kept getting stuck on the integration layer. Once a team aligns on what the unified view needs to show, the architecture choices become legible. The integration questions get smaller. The vendor conversations get sharper.

Trisha pushed Chris on the second part of this — the part that scares people who have already heard the words "data integration" too many times.

You don't have to connect everything. Some people are going to get overwhelmed. If you tell me I have to connect all the external systems to make this actually work, I'm just going to throw my hands up because it feels like 2030.

She is right. And the answer to that overwhelm is the line Chris keeps coming back to, the line that defines the operating principle for everything that follows in this series:

Last Friday, Chris was on a call with a team that had invested heavily in HubSpot — but was still operating from a spreadsheet that held the actual truth of the work. The spreadsheet was a safety blanket. Every attempt to clean up HubSpot, integrate the data lake, or stand up a data hub crashed against the same wall: nobody trusted the system to give them what the spreadsheet gave them. So they kept the spreadsheet. So adoption stalled.

The fix was not bigger integration. The fix was smaller and more human. The team agreed to use the contact record as the surface of truth and added native CRM cards for the associated records — deals, line items, related companies — so the leader could edit any of it from one screen without context-switching. The spreadsheet's questions came home. The click happened. Adoption moved.

2. Your Data Is Incomplete and Bad

The original framing of this show was simpler: your data is incomplete. After six months, the framing got sharper. Trisha said it on the recap and Erin built on it for the next twenty minutes.

Number two was your data is incomplete. But I want to include in that — your data is bad. So you've got big buckets of data that are either wrong because they're missing stuff, or they're wrong because there's just problematic stuff in the data.

Both problems are real. Both have the same root cause. And — this is the part that matters — both have the same answer.

Erin named the principle. She has been saying it for weeks across different client work, and on May 6 she put it the cleanest way she has put it yet:

Make it hard to assume things. The only way to really do that is to make it dead simple to find the facts. Like if the facts are in black and white, hitting you in the face… a grumpy boss has less excuses to only see one slice.

She added: "It shouldn't feel like a scavenger hunt. People will grow to resent that."

Trisha then told a story that anybody who has ever worked for a difficult leader will recognize instantly. A boss — owner of the company, signer of paychecks, controller of mood — would walk in, see one thing out of context, and "make 17 assumptions about why you did or didn't do a thing." He was always wrong. Always. Not because he was incompetent. Because he was looking at one slice of a system that did not surface the rest of the picture.

The fix is not management training. The fix is not a memo about psychological safety. The fix is architectural: surface the full context where the decisions get made. If the decision happens at the contact record, the context lives at the contact record. If the decision happens at the dashboard on the wall, the dashboard tells the truth.

There is a deeper claim here about where data actually lives in most organizations. It is rarely where the org chart says it lives. Inboxes are the actual CRM for many sales teams. Spreadsheets are the safety blanket. Slack threads hold the rationale for half the decisions on a given week. The shadow data is not shadow because the team is hiding it. It is shadow because no system gave it a home where it belonged.

This is where AI has actually shifted the economics. What used to be expensive to capture — the conversation, the context, the rationale — is now cheap. The HubSpot AI note-taker has gotten good. Stop complaining that it is not perfect. Bring the call transcripts in. Let the AI summarize. Let the team check the summary. Now the marketing team and the sales team are reading from the same record, and the place where misalignment used to live has a context-rich answer instead of a one-line note that someone forgot to update.

3. Readiness Requires an Owner — and That Owner Is RevOps

Earlier in the series, the conversation was: someone needs to own the system. By May 6, the conversation had sharpened to a specific role: someone needs to own how the systems interact with the humans, and that role has a name. It is RevOps.

This is a meaningful shift, and it is worth slowing down on. "You need ownership" is the kind of advice that does not produce an org chart change. "You need a RevOps role whose job is to understand how the ten lanes of the highway intersect" produces a job posting.

Erin walked the framing.

What I'm seeing play out is more of an emphasis on a RevOps role. Everybody is going to work to stay in their lane and do their job because that's their job. They're not — if it's not specifically assigned to them, it's just noise.

Then she said the part that is hard to push back on once you hear it.

I think it's unfair to expect someone to take on systems outside of their kind of vertical at work. I trust the sales team to do your job well and just clue me in on what I need to know.

She extended the metaphor. "It's a lot of cognitive load on the average person trying to work in their lane to be aware of the entire ten lane highway."

Chris finished the thought: "It's a trap, too. The nine other lanes are not yours."

This is where the methodology stops being theoretical. The Three-Org Model gives you Customer, Operations, and Finance. Each org has people who own a function. RevOps is a different shape of role. RevOps does not own a function — RevOps owns how the functions interact. They sit in the seam. They are the person whose job is to know what the sales team needs to see, what the services team needs to see, what the finance team needs to see, and how to make those views agree without forcing every individual to track every system.

There was a moment on the recap that made the role definition crisp. Trisha had asked Claude — with no human context — for an AI orchestrator role description. What came back was ten technical acronyms, a passing reference to neural networks, and a line that read something like "you'll have to interact with other teams at some point." When she asked again with the human dimension specified, the template was completely different.

Chris's framing of the lesson:

It doesn't all have to be the same person, but if you don't, somebody needs to be in charge of asking the question — how do our humans work with AI? Not how does AI work? Right? And everybody's focused on that second question.

The role is not "AI specialist." The role is "person whose job is to ask how our humans work with AI." Sometimes the answer is a chief people officer paired with an AI leader. Sometimes the answer is a head of RevOps with sufficient authority to call cross-team meetings without permission. The shape varies. The function does not.

If you take only one thing from this article, take Trisha's framing of the priority order:

4. Foundation Before Agents

The instinct to "give every team an AI" runs hard into the same wall every time. The team does not have the unified context. The team does not have the data hygiene. The team does not have the ownership pattern. So the AI either confirms existing bias, hallucinates from incomplete context, or — worst of the three — makes the existing dysfunction faster and more confident.

The deepest treatment of this in the series is the late-March episode where Chris walks the 12 Complexity Traps in the order he has actually resolved them across client engagements. The order is not arbitrary. It is dependency-shaped. Earlier traps gate the later ones. The B2B Trap and the SaaS Trap are foundational — they have to clear before anything else moves. Skipping them to deploy agents creates a faster broken system, not a working one.

On the recap, Chris named the cascade pattern in his own voice:

All these things started to cascade. It's been an interesting journey to figure this out — and we're still trying. That's why we're here every week.

The reason this section is shorter than the others is honest: foundation is the precondition for everything in the other four sections. Theme 1 is foundation in data architecture. Theme 2 is foundation in data quality. Theme 3 is foundation in human ownership. Theme 5 is foundation in organizational learning. The "agents on top" question only becomes interesting once those four are real.

If a team is being told "you need agents now, the competition is shipping agents, you are behind" — and the team does not have a unified view, does not trust its data, has no ownership role for system-of-systems decisions, and has no cohort of people learning AI together — the agents are going to amplify the gap, not close it.

5. Readiness Is Organizational, Not Technical

This is the strongest moment of the May 6 show. It is also the part of the conversation that resists being reduced to a takeaway, because it is structurally a "position" rather than a tip.

Trisha set it up:

AI has never happened to us before. And it's happening to us universally. Every person in every country on the planet is, in the big scope of things, as new to AI as everybody else is. We're all going through this for the very first time together at the same time. We don't have infrastructure to lean on because of the universality of it.

That sentence is doing a lot of work. The universality is the part most readiness conversations miss. There is no precedent organization to copy from. There is no senior leader who already figured this out at a previous company, because there are no previous companies that did it at scale. Every team starts at the same place: not knowing. The leaders who acknowledge that out loud get further faster than the leaders who pretend otherwise.

Chris named the operational consequence:

Anybody can show up and learn AI in front of everybody else. That creates the safe space for everybody else to start learning in the same way. There's nothing else we can point to where literally the only way to learn is by building with it, like using it.

There is a specific anti-pattern this calls out: "Everybody go use AI to get your job done, and let me know how it goes." That sentence is the "opposite" of readiness. It outsources the learning curve to individual cognitive load. It produces inconsistent capability across the team. It produces silent frustration in everyone too senior to admit they do not know what is happening. It produces dashboards full of "AI tools deployed" with no underlying skill growth.

Cohort learning inverts that pattern. Visible learning. Shared friction. Compounding capability. Trisha closed with the prescription:

Learning goes fastest if you have a cohort that's in about the same stage as you are. If you can partner up with even just one person and both commit to learning together, you'll be learning at the same pace and you'll learn much faster by having somebody else you can talk to.

Chris named his own cohort by name — the inner loop with Mico and George — and was unambiguous about what it produced:

That's 100% the reason I am where I'm at. Cohort learning was already coming back. If you go back before digital — that's how people liked to learn. Apprenticeships and all these programs where it's people doing stuff together. Still the best way to learn for human beings.

What's Next

Next week's episode walks the other side of the readiness picture. The five themes above describe the "state of the inside" of the organization — the data, the people, the ownership, the foundation, the learning culture. Next week's conversation is the "state of the outside" — how to evaluate vendors, partners, and external help when you decide it is time to bring some in.

Together they form the complete picture. This article tells you whether "you" are ready. Next week's episode tells you how to recognize whether your "partners" are ready to meet you where you are.

If you are new to the series, the entry points are simple. Watch the May 6 recap embedded above. Read the URV Part 5 article series for the deep architectural cut. Subscribe to "Value-First Platform: AI Data Readiness" before next week's partner-readiness episode. The full series listing lives on the AI Data Readiness pillar page.

The thing the show keeps proving is the thing Chris said at the end of the recap, almost in passing: "let's figure it out together." That is the readiness posture. Everything else follows from it.