How do I prepare an AI ROI report for the board?
Turn raw AI spend into a board-ready ROI narrative: cost per outcome, efficiency trend, and a one-page PDF — in under 30 minutes.
Goal: Walk into a board meeting with a one-page PDF showing AI spend, cost per business outcome, efficiency trend, and net savings — with an AI-written narrative so you're not assembling slides the night before.
Prerequisites: At least 30 days of AI calls through Cognocient. Outcome tracking set up for at least one key metric (see Step 1).
Step 1 — Define your outcome metric
Go to Executive → Outcomes → Define Outcome. Pick the metric that makes AI spend feel like an investment rather than a cost — name it, choose Correlation Method (By Session or By User), and set the correlation Window (minutes):
| Use case | Outcome name | Correlation method |
|---|---|---|
| Customer support | Support ticket resolved | By Session |
| Legal / contracts | Contract drafted | By Session |
| Content | Article published | By Session |
| Engineering | PR reviewed | By Session |
Once the outcome is defined, push events from your application whenever the outcome actually occurs:
The more outcome events you record, the more reliable your cost-per-outcome figure is. Set this up now — the data starts accumulating immediately. See Outcomes for the full setup guide.
Step 2 — Let the data accumulate
Once outcome tracking is in place, click into the outcome on Executive → Outcomes to see:
- Avg Cost per Outcome — e.g., $0.42 per ticket resolved
- Total Outcomes — e.g., 28,500 tickets
- Total AI Investment — total spend matched to those events
There's no "AI efficiency vs. alternative" benchmark field here — if you want the human-cost comparison, track that number yourself and bring it into the narrative as context. You don't need to do anything else during this period; the metrics update as events come in.
Step 3 — Open the Executive View before your meeting
Go to Executive View. This is the CFO-facing dashboard. Check:
| Section | What to note |
|---|---|
| Total AI Spend | This period's total — you'll say this number out loud |
| AI Efficiency Score | Cost per outcome in dollars, with a month-over-month trend — not a 0–100 score |
| VS Last Month | % change in spend — ideally flat or declining vs. rising outcomes |
| Savings Summary | What optimisations delivered — concrete dollar amounts |
If any numbers look off — a big spend spike, an efficiency trend reversing — investigate before the meeting, not during it.
Step 4 — Choose your narrative tone
Go to Reports and click into the Generate Report tab. Pick a narrative tone:
| Tone | Use when |
|---|---|
| Board / CFO | Presenting to non-technical stakeholders, quarterly business review |
| Technical Finance | Finance team that understands engineering trade-offs |
| Engineering | Internal team retrospective, sprint review with cost context |
Choose Board / CFO for board meetings. The AI generates a short narrative (150–200 words) summarising your AI investment, efficiency trend, and outlook — written in plain business language.
Step 5 — Download and review the PDF
Download the AI ROI Summary PDF for the selected month. It includes the AI-written narrative alongside your spend, cost-per-outcome, and waste-recovered figures.
Review the narrative for accuracy. The AI writes from your data but you know the context (a Q4 push, a new product launch) that explains any unusual numbers.
Step 6 — Set up monthly auto-delivery
After your first successful board report, automate it. Go to Reports → Scheduled Delivery and configure:
| Setting | Value |
|---|---|
| Report type | AI ROI Summary (1 page) |
| Frequency | Monthly (delivered on the 1st) — weekly and end-of-billing-cycle are also available |
| Recipients | CFO email, your email |
| Narrative tone | Board / CFO |
There's no custom day-of-month picker — Monthly delivery goes out on the 1st. Toggle Require approval before sending if you want a named approver to review it before it reaches recipients.
What a good ROI story looks like
A strong AI ROI narrative has three numbers:
- What you spent — "$47,000 on AI this quarter"
- What you got — "85,000 support tickets resolved at $0.55 each"
- What the alternative costs — "Our human agents cost $8.40 per ticket — AI is 15× more cost-efficient"
If you can answer "what would this have cost without AI?", you have a compelling investment case. If you can't answer that yet, defining a benchmark cost in Outcomes is the fastest path to having that number for next quarter's meeting.
Related articles
Tag Your First AI Call
Add 2 headers to your existing code and see per-feature spend in under 5 minutes.
Set a Monthly Spending Limit
Create a hard budget enforced at the proxy before charges reach your provider bill.
Cut Your AI Bill with One Click
Use AI Advisor recommendations to apply model downgrades and caching without code changes.