The One-Page AI Brief That Gets Executive Buy-In

The best AI project proposal I have ever seen fit on one page. The executive sponsor read it in 90 seconds, asked two questions, and approved it on the spot. The worst proposal I have seen was 47 pages. It went through four rounds of committee review over three months, got sent back for revisions twice, and the project window closed before it was ever approved.

The difference was not the quality of the AI idea. Both were solid. The difference was how the idea was packaged. One respected the executive's time and decision-making process. The other did not.

Why Most AI Proposals Fail Before They Are Read

Executives do not reject AI projects because they do not believe in AI. They reject them because the proposal does not answer their three questions fast enough: What is the problem? What will it cost? What do I need to decide?

Most AI proposals fail for one of three reasons. First, they lead with technology instead of the business problem. Your CTO might care that you are using transformer-based NLP with fine-tuned embeddings. Your CFO does not. Your CFO cares that invoice processing takes 4 hours a week and costs $180K a year in labor. Lead with the pain, not the solution.

Second, they are too long. A 20-page proposal signals that you have not done the hard work of simplifying. If you cannot explain the project in one page, either the scope is too big or you do not understand it well enough yet. Both are reasons to say no.

Third, they lack a clear ask. 'We should explore AI for customer service' is not an ask. 'I need $340K and two dedicated team members for 90 days to pilot an AI routing system that will cut average handle time by 35%' is an ask. One gets a head nod and no follow-up. The other gets a decision.

The 6-Section One-Page Format

After 20 years of watching AI proposals succeed and fail in banking, I have narrowed the one-page brief down to six sections. Every section earns its space by answering a specific question the decision-maker is thinking but may not ask aloud.

Section 1: Problem Statement (2 sentences). Name the specific task, the person who does it, and the pain it causes. Not 'improve operational efficiency.' Instead: 'Our 12-person accounts payable team manually processes 3,200 invoices per month. Each invoice takes 7 minutes of human review, costing $14K/month in labor with a 4.2% error rate.' The executive now knows exactly what you are talking about.

Section 2: Current State (the numbers table). Six rows: process owner, people involved, hours per week, current cost, error rate, and pain points. This section exists so the executive can verify the problem is real. If you cannot fill in these numbers, you are not ready to propose a solution. Go back and measure first.

Section 3: Proposed Solution (3 sentences). One sentence on what the AI will do. One sentence on the approach — buy, build, or partner. One sentence on what 'good enough' looks like for V1. Resist the urge to describe the architecture. The technical details belong in an appendix, not on the one-pager.

Section 4: Success Metrics (3 numbers). The ONE primary metric you will use to judge success at 90 days, plus two supporting metrics. Make them specific and measurable. 'Invoice processing time drops from 4 hours/week to 30 minutes/week' is a success metric. 'Increased efficiency' is not.

Section 5: Timeline (4 phases, 90 days). Foundation (Weeks 1-3), Build & Configure (Weeks 4-8), Pilot (Weeks 9-11), Full Deployment (Weeks 12-13). Each phase gets a one-line milestone. The 90-day frame is deliberate — it is short enough that the executive can commit without feeling locked in, and long enough to deliver real results.

Section 6: The Ask (budget + headcount). Total Year 1 cost broken into four categories: software, implementation, infrastructure, and people. Then the team commitment: who you need, at what time allocation. The executive needs to know two things — how much money and how many people. Give them both in one table.

The 90-Second Read Test

Here is how I test every brief before it goes to an executive: hand it to someone who knows nothing about the project and time them. If they cannot tell you the problem, the solution, the cost, and the timeline in 90 seconds, revise it.

The best one-pager I saw passed this test with 40 seconds to spare. It was for an AI-powered document classification system at a regional bank. Problem: compliance team spending 60 hours/week manually categorizing loan documents. Solution: AI classifier trained on 18 months of historical data, deployed as a pre-processing step in the existing document management system. Cost: $280K Year 1. Timeline: 90 days to pilot, full deployment by Day 100. Ask: $280K budget and a 50% time commitment from one senior compliance analyst.

The executive read it, asked 'What happens if the accuracy is below 90%?' and 'Who is the project owner?' Both answers were on the page. Approved in the same meeting.

The Three Additions That Close the Deal

The six sections get you to 'I understand the project.' Three more elements get you to 'yes.'

Top 3 Risks with mitigations. Executives do not expect risk-free projects. They expect you to have thought about what could go wrong. Three risks is the right number — enough to show diligence, not so many that you signal the project is dangerous. For each risk, include the mitigation. 'Data quality may be lower than expected — we will run a data audit in Week 1 before committing to the full build.'

Go/No-Go Criteria at each phase gate. This is the single most powerful addition. It tells the executive: 'We have built in exit ramps. If the data audit fails in Week 3, we stop. If pilot accuracy is below threshold in Week 11, we do not deploy.' This removes the executive's biggest fear — that they are approving an irreversible commitment. Phase gates make the project stoppable. Stoppable projects are easier to approve.

Approval signatures. A line for the executive sponsor, the technical lead, and finance. This is not bureaucracy. It is commitment. When people sign their name next to a decision, the follow-through rate goes up. It also protects you — six months later, when someone asks why this project was approved, the signatures are on record.

Common Mistakes to Avoid

Do not attach a slide deck. The one-pager IS the document. If you attach a 15-slide deck 'for reference,' the executive will flip through the deck and skip the one-pager. You lose control of the narrative.

Do not use the word 'explore.' 'We want to explore using AI for...' is a research request, not a project brief. Executives fund projects, not explorations. If you are still exploring, you are not ready for the one-pager.

Do not leave the success metric vague. 'Improved customer experience' is not measurable. 'NPS score increases from 42 to 55 within 6 months of deployment' is measurable. If you cannot put a number on success, the project is not defined well enough.

Do not skip the current state table. This is where you earn credibility. An executive who sees that you have measured the current process — hours, cost, error rate — trusts that you understand the problem. An executive who sees vague statements about 'inefficiencies' assumes you are guessing.

Actionable Takeaway