How to Convince Leadership to Adopt AI

Good ideas don't get approved. Good explanations do.

The Loneliest Moment for Practitioners

You've defined the problem. Explored tools. Even ran a small experiment and saw promise. Now you need budget and headcount.

So you stand before leadership.

"With AI, we can do all of this. Productivity goes up, costs go down, competitiveness gets stronger."

Leadership nods along, then asks:

"So how much does it cost, and when do we see results?"

Most AI proposals stall right here. You can explain the technology's potential, but you haven't translated it into the language leadership wants to hear.

Leadership Doesn't Want to Hear About Technology

Three questions live in every executive's head.

"How much do we spend, and how much do we make?" Return on investment. The default frame for every decision.

"What happens if we don't do it?" Opportunity cost and risk. The cost of inaction is also a powerful persuasion tool.

"How much does it hurt if it fails?" The size of the risk. Executives are more sensitive to the magnitude of failure than the probability of success.

If you can answer these three, you can minimize the technology explanation. Leadership doesn't need to know how AI works. They only need to know what problem it solves and at what cost.

The Structure of Persuasion: Speaking in Four Blocks

When proposing AI adoption, present these four blocks in order.

Block 1: Show the Size of the Problem in Numbers

"There's inefficiency" moves no one. "This inefficiency costs us 240 million won annually in labor" opens ears.

The formula for turning a problem into numbers is simple.

[Number of people involved] × [Time spent] × [Hourly labor cost] × [Annual frequency]

For example: 8 sales reps × 6 hours per quote × 30,000 won per hour × 20 quotes per month × 12 months = 345.6 million won annually. If AI can cut 70% of that, the annual value is roughly 240 million won.

The numbers don't need to be perfect. Executives know they're estimates. What matters is the sense of scale.

Block 2: State the Solution in One Sentence

Technical architecture and vendor comparisons come later. Lead with one sentence.

"We'll build a system where you input a customer request email and it auto-generates a draft quote. The sales rep only needs to review it and decide on the discount rate."

What this one sentence needs: ①what goes in, ②what comes out, ③what the person does. If these three are clear, leadership can picture it.

Block 3: Propose a Scale Where Failure Is Acceptable

The biggest reason executives hesitate on AI projects is uncertain success. Acknowledge the uncertainty head-on. Then propose a structure that minimizes the cost of failure.

"We need 100 million won to build the full system" versus "We need 5 million won for a 4-week pilot. If results are positive, we expand. If not, we stop here." These are completely different proposals.

The latter gets approved. Reversible decisions are ones executives make comfortably.

Key elements of a pilot proposal:

• Scope: Which team's which process
• Duration: How many weeks to test
• Cost: How much (including labor)
• Success criteria: What confirms we proceed to the next stage
• Failure plan: What happens if it doesn't work

Fit these five on one page, and you have a complete proposal.

Block 4: Show What Happens If You Don't Act

If you only talk about the benefits of adoption, the response is "we can do this later." To create urgency, show the cost of inaction.

Competitive landscape: "Competitor A has already deployed similar automation and cut customer response time in half. This gap compounds every quarter."

Talent market: "Positions heavy on repetitive work have high turnover. Last year, turnover in this role was 25%, and recruitment plus training costs 30 million won per person annually."

Opportunity cost: "If our sales team could spend their data aggregation hours on client-facing work instead, a conservative estimate suggests a 5% quarterly revenue increase."

Show that doing nothing has a cost too.

Three Mistakes to Avoid

Explaining in Technical Jargon

"We'll fine-tune a RAG-based LLM on our internal documents..." — this is where you lose executive attention. Talk about technology only in terms of "what it solves," not "how it works." Save technical details for when they ask.

Presenting Only Rosy Projections

Numbers like "300% productivity improvement" or "80% cost reduction" actually erode trust. Present conservative estimates and say, "This is the minimum expected impact — the actual result could be larger." That's far more convincing.

Asking for Everything at Once

Requesting budget, headcount, system access, and organizational restructuring all at once makes the decision too heavy. Keep the first ask light. "4-week pilot, 5 million won budget, 30% of one person's time." Start with the minimum unit.

Persuasion Isn't a One-Shot Event

Convincing leadership doesn't end with one presentation. It's a process.

Round 1 — Share the problem: "We have this problem, and it's costing us this much." Don't mention AI yet. The goal is to build empathy for the problem's scale.

Round 2 — Propose a direction: "We think we can solve this problem in this way. We'd like to run a small test." The goal is pilot approval.

Round 3 — Share results: "Here's what we found in 4 weeks of testing. To expand, we need these resources." Data does the persuading.

Follow these steps, and by the third meeting, leadership will be the ones asking "Can we expand this further?"

In the End, It's a Translation Problem

The essence of an AI adoption proposal is translation. Translating technology's potential into business language. Translating practitioners' pain into management's numbers. Translating uncertainty into manageable risk.

A good translator stays faithful to the original while converting it into words the listener can understand. AI proposals work the same way. Don't distort the technology — just reshape it into a form executives can judge.

You, reading this, are probably that translator. Good luck with the translation.