Before you hire AI help, read this.

The default sequence for hiring AI help follows a familiar pattern. The consultant arrives. They walk through ten or fifteen tools. The operator asks where to start. The consultant says, "Find your pain and solve it." The room nods. Six weeks later, the deliverable is a custom GPT, a couple of Zapier flows, and a Notion workspace. The work is technically clean. None of it changed how the business runs.

The advice is half-right at best. Pain points are symptoms, not structure. A tool plugged into one of those symptoms relieves the immediate complaint and creates a new dependency. Six tools plugged into six symptoms produce a stack that fragments instead of compounding. The team adds AI to their workflow without anything moving in the business that matters.

The data has caught up. Gartner reported on April 7, 2026 that only 28% of enterprise AI infrastructure projects deliver the promised return. One in five fails outright. Across the failures, 73% had no agreed definition of success before work started. The pattern is consistent. Pain-driven discovery, generic recommendation, fragmented build, no system-level outcome.

The fix is structural. Walk into the AI conversation with a one-page brief that maps how the business actually runs. The brief surfaces structure, not pain. It is the operator's preparation document, written before any vendor or consultant gets involved, and it changes every conversation downstream.

Why it matters now: Anthropic announced a $1.5 billion mid-market AI services venture three days ago. OpenAI announced a $4 billion version the same week. Forward-deployed AI engineering just became accessible at scales that historically couldn't afford McKinsey. The inbound is about to multiply. The pain-first script will be everywhere. The operator who walks in unprepared will be sold what the consultant brought.

What to do this week: Write the brief before the next AI conversation. The operator who walks in clear on how the business runs gets a different conversation than the one who walks in waiting to be told.

The fifteen minutes to write the brief is the most useful work an operator can do before any AI conversation. Not to test the consultant. To force clarity on what the business actually needs, in plain language, before someone else fills that vacuum with their framing.

The questions are about the business, not about AI. They require honesty, not expertise. Most can be answered by the leadership team in a single working session.

By the end of it, the operator has a written picture of how the business actually runs. What is slowing revenue, separate from the loudest complaints. What is already in place that any new work has to integrate with. What success would mean in numbers a board would care about.

That clarity is what makes the AI conversation productive. The consultant gets a clear picture of the business in five minutes instead of five meetings. The operator can evaluate every recommendation against the brief, not against a sales pitch. The conversation ends with something the organization actually needs, not something that sounded good in the room.

The cost of skipping the prep is invisible until it lands. Three meetings in, the conversation has drifted to a tool the consultant happened to bring. The proposal solves a problem the operator never named. Six weeks later, the build does not fit. The fifteen minutes upfront would have changed every conversation downstream.

The AI Brief Builder makes the prep take fifteen minutes instead of an afternoon. Use it before any AI hire, internal planning meeting, or vendor demo. The operator who walks in clear gets a different conversation.

Note: Everything you enter is processed in real time and never stored. The data is not visible to anyone, including this newsletter. Put in what is actually true and the output will be worth acting on.

Default belief: AI expertise is the qualification. Hire someone with AI credentials.

Flip: Systems thinking is the qualification. The AI part is learnable.

A mid-market engineering firm hired an AI consultant after a polished demo. The consultant arrived with a deck of fifteen tools and asked the team about their pain points. The team listed a few. Proposals took too long. Project tracking lived in spreadsheets. Knowledge transfer between engineers was uneven. Six weeks later, the deliverable was three custom GPTs trained on past proposals, a Notion workspace for project tracking, and a Slack bot for knowledge queries. Each tool worked. None of it changed how the firm actually delivered projects.

What the firm should have caught is what they would have caught instantly in their own discipline. An engineer does not accept another engineer who shows up with a design before understanding the building. They map the loads, the soil, the use case, the existing system, then recommend a structure. The AI consultant did the opposite. Showed up with the design already drawn. Asked the firm to point at where to install it.

Most "AI consultants" are tool-category specialists. They lead with what they know, which is tools. They are not trained to map systems before recommending solutions. The engineering firm CEO would have caught this in their own field in five minutes. In an unfamiliar field, the same instincts went quiet.

The fix is to apply the same standard. Treat AI hires the way any other technical hire gets treated. Watch what they produce in week one. Architects produce maps. Tool peddlers produce tool inventories.

AI is a tool category. Systems thinking is the discipline that picks the right tools.

Most AI consultants will start with your pain. The hire that actually moves the business starts with how the operation runs. The brief is the operator's preparation document. It is what gets the organization what it needs out of the conversation.

Could you map how your business actually runs on one page right now?

Write your brief. The AI Brief Builder walks through six questions in fifteen minutes. Open the AI Brief Builder →

Last week's hidden cost calculator is still live. Open the calculator →