The Seven Conversations

Every enterprise AI strategy I have seen makes the same mistake. It starts with IT. It lives in IT. And it dies in IT.

The CTO picks a platform. The data team builds a model. A pilot gets launched. The board gets a presentation with the words "AI-powered" on every slide. And somewhere in the building, the General Counsel's office is still reviewing contracts by hand, the CFO's team is running forecasts on spreadsheets from 2019, and the supply chain team is reacting to disruptions that AI could have predicted two weeks ago.

The mistake is not technical. It is structural.

The seven languages problem

AI transformation is not one conversation. It is seven different conversations happening in seven different languages.

The CLO talks about risk, liability, and compliance. The CFO talks about ROI, scenarios, and cost reduction. The CHRO talks about workforce impact, reskilling, and change management. The CTO talks about platforms, architecture, and infrastructure. The COO talks about process redesign and operational efficiency. The CMO talks about personalization and customer experience. And the CEO is trying to synthesize all of it into something the board can understand.

In most enterprises, these seven conversations happen in different rooms, on different timelines, using different vocabularies. The CLO does not know what the CTO just deployed. The CFO does not know the CHRO is piloting AI for workforce planning. The supply chain team has no idea that the marketing team's demand signals could improve their forecast by 30%.

Nobody is translating.

Why this gap exists

It exists because of how enterprises organized themselves for the last fifty years. Functions were designed as vertical silos — each with its own budget, its own metrics, its own leadership. That structure worked for efficiency. It fails for transformation.

The consulting firms know this. McKinsey, BCG, and Accenture — who collectively do more than $7 billion annually in AI consulting — all say the same thing: 70% of AI transformation success is people, process, and culture. 30% is technology. But most enterprises still treat it as a technology project with some change management sprinkled on top.

The seven conversations

Each function brings a distinct conversation to the AI table. Naming them is the first move toward translating between them.

1. Risk (CLO). What are we liable for, and how do we govern it before regulators or courts decide for us?
2. Numbers (CFO). What is the real ROI, how do we measure it, and which business case is real versus a vendor pitch?
3. Talent (CHRO). Whose job changes, who gets reskilled, who leaves — and how do we lead the workforce through that without losing the people who matter most?
4. Architecture (CTO). What platforms, data foundations and security perimeters do we need, and how do we build for the next ten years rather than the next demo?
5. Operations (COO). Which processes get reimagined versus paved-cowpath automated, and how do we sequence the redesign without breaking the business that pays for it?
6. Customer (CMO). What does AI change in the relationship with the customer — personalization, service, trust — and where is the line between useful and creepy?
7. Synthesis (CEO). How do these six conversations add up to a single coherent story the board, the market, and the workforce can hear?

What the winners do differently

I have spent the last decade watching how large enterprises absorb new technology. The pattern is remarkably consistent.

The companies that fail deploy AI into their existing structure. They pick a function, build a use case, celebrate the pilot, and wonder why it does not scale.

The companies that succeed deploy AI acrosstheir structure. They do not ask "which function gets AI first?" They ask "which business outcome requires multiple functions working together, and how does AI make that coordination faster?"

That reframe — from function-first to outcome-first — is the difference between an AI project and an AI transformation.

The legal conversation: a worked example

I spoke with a General Counsel whose team was reviewing 6,000 contracts a year. Her best associate could do twelve a day. The math was brutal — they were permanently 18 months behind on contract reviews. Not because of incompetence; because the volume of business had outgrown the capacity of the team.

She did not start with a vendor demo. She started with a question: What if we could review every contract the day it arrives?

That question reframed the entire project. They built an AI intake layer that flagged deviations from standard positions, an exception-routing layer so associates only read the unusual contracts, and a pattern layer that learned which counterparties consistently pushed on which clauses. Turnaround went from three weeks to three days. Same team. Fundamentally different work.

The hard part was not the AI. It was the conversation with senior associates who had built their entire career on slow, careful, thorough reading. Telling someone their craft is now a commodity is a leadership conversation, not a technology conversation. The GC reframed it: "AI is freeing you to do the work you went to law school for." The associates who adapted thrived. A few did not.

That is the texture of every one of the seven conversations. They are technology conversations on the surface and identity conversations underneath. The translator role exists to hold both at once.

Why Legal might be the most important conversation

A counterintuitive argument: the legal function might be the most strategically important function in your entire AI transformation. Not because legal work has the highest ROI for AI — supply chain does. Because Legal sits at the intersection of risk, governance, and compliance: the three things that slow every other function's AI adoption.

If Legal is blocking AI deployment, every function slows down. If Legal is enabling AI deployment by providing clear governance frameworks, every function speeds up. The companies where the CLO and the CTO have a strong working relationship are deploying AI two-to-three times faster than companies where Legal and IT barely talk.

That relationship — the Risk Conversation meeting the Architecture Conversation — is the canonical example of two of the seven conversations meeting in the right room.

What to do Monday morning

1. Map the seven conversations. Literally list who in your organization is making AI decisions for each function. If any of those roles are empty or defaulting to IT, you have found your blind spot.
2. Find the orphaned signals. What data does one function generate that another function could use? Marketing's demand signals for supply chain. HR's attrition predictions for finance. Legal's contract patterns for sales. These cross-functional flows are where AI creates the most value — and where most enterprises have zero infrastructure.
3. Ask who translates. If your AI strategy does not have someone who can sit in a room with the General Counsel AND the CTO AND the CFO and be credible with all three, you have a strategy without a translator. That role — the AI Translator — is the scarcest talent in enterprise AI.

The series ahead

The Seven Conversations is being developed as a fifteen-essay arc and a sixty-thousand-word book. The arc:

1. Act 1 — The Problem. The blind spot. Why pilots succeed and transformations fail.
2. Act 2 — Function Deep Dives. One essay each on Legal, Finance, IT, HR, Supply Chain, Operations.
3. Act 3 — Cross-Cutting Patterns. The translator role. Wardley mapping for AI. Governance as accelerant.
4. Act 4 — Synthesis. What Google, Uber, and Amazon actually teach. The talent gap nobody is measuring. AI transformation is not an IT project.

This is the master frame. The function essays — the Legal one, the Finance one — go deep on a single conversation at a time. They will surface here as they ship.