You've noticed it. Half your team barely opens GitHub Copilot. The other half uses it for autocomplete and nothing else. You know with better training they'd get 3x the value.
Now you have to convince your CFO.
Here's the exact math that works.
Start With What You're Already Spending
Most companies don't do this calculation. Do it now.
Your current AI tool cost:
- GitHub Copilot Business: $19/seat/month
- Microsoft Copilot for M365: $30/seat/month
For a 50-person engineering team on Copilot: $950/month = $11,400/year
If your team's utilization is 25% (industry average without training), you're effectively paying $11,400 for the equivalent of 12.5 seats of real usage.
That's the cost of inaction. Put that number in your deck.
The ROI Math — Conservative Version
Training cost for a 50-person team: ~$6,500 (one-day workshop, flat fee for the whole team)
Conservative time savings assumption:
- 1 engineer saves 30 minutes/day from better AI tool use
- Average fully-loaded hourly rate: $80/hour
- Savings per engineer per day: $40
- At 50% utilization improvement: 25 engineers × $40/day = $1,000/day saved
Payback period: 6–7 working days.
Present this as the conservative case. Most teams see more.
The Questions Your CFO Will Ask
"How do we know the training actually works?"
You measure baseline utilization before training (your Copilot admin dashboard shows this) and re-measure 30 days later. Industry average post-training: 65–75% utilization. Pre-training: 20–35%. That's a measurable delta you can show leadership.
[Model your specific numbers with the free ROI calculator at askpatrick.co/roi-calculator.html]
"Couldn't developers just watch YouTube videos?"
They could. They haven't. Self-directed learning fails for productivity tools because it lacks job-specific context. Generic tutorials don't cover your codebase, your workflow, your PR process. Live, contextual training does.
"Can't we train one person and have them teach the rest?"
You can — but that person needs to be trained first. And they'll need a structured curriculum, not tribal knowledge. That's exactly what a professional playbook delivers.
The Slide That Closes It
One comparison that works:
| Current State | Post-Training | |
|---|---|---|
| Utilization | 25% | 70% |
| Annual tool cost | $11,400 | $11,400 + $6,500 training |
| Effective cost per active seat | $456 | $159 |
Effective cost drops by 65%. The tool becomes 3x more valuable. Training pays for itself in the first month.
Template: One-Page Business Case
Problem: We're paying $[X]/year for AI tools. Current utilization: ~[%]. We're leaving $[Y] in productivity value unused annually.
Proposed solution: One-day team training workshop. Flat cost for the whole team.
Expected outcome: Utilization increases to 65–75% within 30 days, based on industry data from trained teams.
Payback period: [X] working days, based on [Y] hours/engineer/week saved × [Z] average hourly rate.
Measurement plan: Copilot admin dashboard, before/after utilization report, 30-day follow-up check-in.
Next Step
If you want to model your specific numbers before taking this to leadership, the free ROI calculator at askpatrick.co/roi-calculator.html does the math in 2 minutes.
If you're ready to talk about a workshop for your team: askpatrick.co/contact.html
Ask Patrick Co-Work runs half-day AI tool training sessions for engineering teams. Tool-specific. Live. Built around your actual workflow.
Top comments (0)