Skip to content
All insights
Guide · Strategy

How to Choose Who Builds Your First AI Agent

The real choice is not consultant against fractional executive. It is whether the person deciding what to build sits in the same room as the people building it.

· 5 min read

Patrick Laughlin, Founder, Chief AI Officer

Naval Academy, class of 2011. Decorated naval aviator, 10+ years. President, 2023 Inc. 5000 company.

A retirement planning firm had a quote sitting on the chief technology officer's desk. Forty-two thousand dollars and eight weeks, for one agent.

The agent was not exotic. A prospective client fills in a family questionnaire before their first meeting. The agent reads it, summarizes what matters, and puts the summary where the advisor will see it. At the time, a copy was emailed to an assistant, printed and left on a desk, and the advisor read it in the four minutes before the meeting.

His objection was not the price. It was that his advisors needed it now, and that after eight weeks he would have exactly one agent and no better idea of what to build next.

That is the choice most executives are actually making. It is not the choice the market describes.

Advice, or hands

The market sells two things.

Advice with no hands is an assessment, a roadmap, a ranked list of opportunities. It is real work and we do it, because you cannot sensibly build without it. But if nobody builds anything for a quarter, the ranking is stale before it is used, and the document is filed in the place where documents go.

Hands with no advice is a shop that builds precisely what you specified. You are the architect now. Whatever you did not know to ask for does not exist. This is the quote above: one agent, exactly as described, in eight weeks, and the rest is your problem.

Most companies buying AI help for the first time do not know which of the two they are buying, because both are sold with the same words.

What we did instead

We built that same intake agent in the first week. Nothing came before it. The agent that had been quoted, in production and in use.

Since then we have built ten agents there, running across eight systems. Pre-meeting briefs, written each morning from the next forty-eight hours of meetings in the CRM and posted to the meeting record and the advisor's inbox. The intake agent. A sales scorecard on call transcripts, against their own sales method. An advisor scorecard. A sales tracker portal on an hourly sync. A seminar check-in and booking tool, which replaced a person working through a calendar one entry at a time with a legal pad. A marketing portal. A single sign-on landing page with roles.

None of that was in a scope document written at the start, because most of it came from people using the thing that shipped in week one and then asking for the next thing. That is the argument for putting the person who decides next to the people who build. The roadmap you write before anything exists is a guess. The roadmap you write in month three is informed.

The work keeps growing. The firm now keeps a wish list of roughly a hundred agents, and running a list that long is its own discipline.

What I cannot show you yet

I do not have a closing ratio lift for that firm, and I will not estimate one.

Their own method computes the closing ratio once a year, in January, with a grace period for meetings still in flight. Sales results also move with coaching, not only with software, and separating the two takes care.

So there is no such number yet. There may be one next January if we isolate the advisors who used the scorecard agent from the ones who did not, and if there is, I will publish it with the method attached.

Ask any vendor who shows you a percentage how it was computed and on what date. The good ones will tell you. The rest will tell you it came from a client, which is not the same thing.

Four questions worth more than a capability deck

Who decides what gets built, and do they attend the build meetings? If the strategist and the builders are different companies, you are the integration layer.

What ships in the first two weeks? Not what is designed. What is running.

Tell me about something you told a client not to build. A firm that has never talked a client out of anything is not advising, it is quoting.

When we stop paying you, what do we still own, and where does it run?

The answer to the last one should be everything, on your infrastructure, in accounts you control. At any point you should be able to walk away with the code, the data and the systems still running. If that is not true, you did not buy a system. You rented one, and the price of the rent is set by somebody else.

Tell us what you want to get done.

A first conversation with a Chief AI Officer about your goals and what you need, whether that is a build, training, change management or expert advice. We shape the engagement around your answer.

A developer at work on a build