Building the Business Case for AI Adoption

Tier 1 — Action counts

The most effective AI business cases work on three levels simultaneously. Start with what is easy to measure, then build toward what is most powerful.

Tier 1: Action Counts (Easy — Start Here)

Measure adoption itself:

• How many people are using AI tools?
• How frequently? Daily, weekly, ad hoc?
• What tasks are they using it for?

Why this matters: It demonstrates demand. If 80% of your analysts are already using AI informally, there is a clear argument for providing sanctioned tools with proper controls.

Tier 2 — Workflow efficiency

Tier 2: Workflow Efficiency (Most Convincing)

Measure time savings on specific workflows:

Why this matters: Time savings translate directly to capacity. If your deal team can produce DD memos in 2 days instead of 5, they can work on more deals simultaneously — or submit faster and win more mandates.

Tier 3 — Business outcomes

Tier 3: Business Outcomes (Most Powerful)

Connect efficiency to revenue and competitive advantage:

• Faster deal execution — more competitive bids — higher win rate
• Higher throughput — more deals with the same headcount — better revenue per professional
• Better quality — fewer errors, more thorough analysis — better investment decisions
• Client satisfaction — faster turnaround — stronger relationships and retention

Why this matters: This is the language your leadership speaks. But these metrics are harder to isolate, so you need Tier 2 data as the foundation.

Calculating the cost

A credible business case includes costs as well as benefits.

Direct costs:

• Enterprise AI licences: Typically $20-60 per user per month for enterprise tiers
• Implementation support: Internal or external resources to set up workflows, train teams
• MCP integration: If connecting to enterprise systems, factor in development time
• Training: Time invested in getting the team proficient

Total cost example:

For context, one senior associate at a major financial services firm costs the firm $200,000-$400,000 per year fully loaded. If AI saves 20% of that person's time, the tool pays for itself on one person's productivity alone — and you are deploying it across 50.

The executive presentation

When you present the business case, structure it as follows:

Slide 1: The Opportunity — "Our competitors are using AI to execute deals faster. Firms with embedded AI workflows report 3x higher ROI on their AI investments. We are currently at Stage 2 (individual experimentation) — organised adoption would move us to Stage 3."

Slide 2: What We Tested — "We ran a pilot with [team name] on [specific workflow]. Here are the results." Before: [time/cost]. After: [time/cost]. Sample output: [show an actual example of AI-assisted work product].

Slide 3: The Proposal — Scope: Deploy enterprise AI to [N teams / N users]. Cost: [$X Year 1, $Y ongoing]. Expected return: [Z% time savings on named workflows]. Timeline: [pilot — expand — full deployment].

Slide 4: Risk and Governance — "We have designed a governance framework (see Module 4) that addresses data privacy, compliance, and model risk." Enterprise tier = no training on our data. Use case classification = appropriate oversight by risk level. Audit trails = full compliance visibility.

Slide 5: Ask — Specific budget, timeline, and next step.

Designing a pilot

Do not propose organisation-wide deployment from day one. A focused pilot is more credible and lower risk.

The ideal pilot:

• One team (5-10 people) that is motivated and has a clear, measurable workflow
• One workflow that is document-heavy, repetitive, and high-value (DD is often the best candidate)
• 4-6 weeks — long enough to see real results, short enough to maintain momentum
• Clear metrics defined before the pilot starts

Pilot selection criteria — choose a workflow that is:

• Document-heavy (AI's biggest strength)
• Time-consuming (large potential savings)
• Repeatable (can measure before and after)
• Not client-facing initially (lower risk for errors)

Avoid workflows that are:

• One-off projects (cannot measure improvement)
• Highly regulated initially (keep governance simple for the pilot)

After the pilot, present results to leadership with specific time savings and before/after data, examples of actual work product (redacted as needed), team feedback on quality and experience, and a proposed expansion plan.

Change management

The best AI tools fail without adoption. Here is what drives adoption and what kills it.

What drives adoption:

• Executive sponsorship: When leadership visibly uses and advocates for AI, teams follow
• Early wins: Small, tangible successes that people can see and reference
• Peer champions: Train 2-3 enthusiastic users intensively, then let them mentor others (the "flight instructor" model)
• Remove friction: Pre-configured tools, shared prompt libraries, clear permissions

What kills adoption:

• Mandates without support: Telling people to "use AI" without training or workflows
• Fear: People worry AI will replace them. Address this directly — AI replaces tasks, not roles
• Bad first experiences: If someone's first try with AI produces garbage, they will not try again. Start them with a use case that works well.
• Governance paralysis: Overly restrictive policies that make AI harder to use than not using it

Key takeaways

1. Start with Tier 2 metrics — time saved on specific workflows is the most convincing evidence.
2. Calculate costs honestly — licences, implementation, training. Then show ROI that makes the cost trivial.
3. Run a focused pilot — one team, one workflow, clear metrics, 4-6 weeks.
4. Present with data — before/after comparisons, actual examples, specific dollar figures.
5. Manage the change — executive sponsorship, peer champions, early wins.

You have completed the Practitioner tier. The Advanced tier begins with AI agents for deal lifecycle automation.