GEO KPIs to Track: The CMO's AI Visibility Board Scorecard
The CMO’s AI visibility board scorecard is a one-slide report with five numbers and one sentence: Share of Model against your top two competitors, citation rate on the money prompts, how accurately AI describes your brand, AI-referred and AI-influenced sessions, and the actions shipped. This page gives you the dashboard behind the slide and what to do when a number slips.
By Vijay Vasu, Founder of Indexable AI. SEO at Uber and the first SEO hire at Uber Eats in the US, former Director of SEO at Zendesk, and former Director of Technology, SEO & AI Innovation at Williams-Sonoma. Updated October 28, 2026.
- Why Don't Traditional SEO Metrics Work for AI Search?
- What KPIs Should a CMO Track for AI Search Visibility?
- The Board Slide: Five Numbers and One Sentence
- How Do You Build an AI Visibility Dashboard for Reporting?
- What Should You Do When an AI Visibility Metric Underperforms?
- What Tools Do You Need to Measure AI Visibility?
- What Will Your CFO Ask About AI Visibility?
- Frequently Asked Questions
Why Don't Traditional SEO Metrics Work for AI Search?
Traditional SEO dashboards measure impressions, clicks, and rankings, but AI search removes the click, so those metrics are blind to the fastest-growing discovery channel. Your marketing stack has zero instrumentation for AI search, and that channel is shaping buyer perception without your knowledge. The AI search metrics playbook defines and calculates the KPIs; this page is the board view, the five numbers that go on the slide.
AI search is eating your pipeline in ways your current tools cannot detect.
Key facts
- ChatGPT has 1.2 billion weekly users (OpenAI, 2026).
- According to Google, AI Mode passed one billion monthly users by May 2026 (Google, 2026).
- 68% of U.S. Google searches ended without a click in early 2026 (SparkToro, 2026).
- Google AI Mode and Gemini shared only 3.5% of cited domains across 42,971 citations (Shashko, 2026).
- According to our own Search Console, marketing leaders asked 229 questions about AI search in the 90 days to Oct 24, 2026; 79% of those questions were on page 1 and they sent 0 clicks (Indexable, 2026).
Sources: OpenAI DevDay 2026 recap · Google I/O 2026 · SparkToro 2026 zero-click study · Shashko 2026 citation study · Indexable Search Console read, 90 days to Oct 24, 2026
When a buyer asks ChatGPT "What's the best CRM for a 100-person sales team?" and gets an answer that mentions three competitors but not your brand, that is a lost impression. You will never see it in Google Search Console.
If the buyer picks a competitor based on that AI recommendation, your organization has lost a deal without knowing a touchpoint existed.
The dark funnel is real. The five numbers below make it visible to the board.
What KPIs Should a CMO Track for AI Search Visibility?
A CMO should track five numbers for AI search visibility: Share of Model against the top two competitors, citation rate on the money prompts, how accurately AI answers describe the brand, AI-referred and AI-influenced sessions, and the actions shipped that moved them. Measure each per engine on a fixed panel of buyer prompts; the formulas are in the AI search metrics playbook.
Share of Model vs the top two competitors
Your share of all brand mentions in AI answers to your buyer prompts, per engine, trended quarterly against the two competitors buyers compare you with. How to calculate it →
Citation rate on the money prompts
Of the answers that name you on your highest-intent prompts, the share that link one of your pages as a source. How to calculate it →
Accuracy of how AI describes you
The share of answers that get your pricing, positioning and product right: accurate descriptions ÷ answers that describe you × 100, with tone tracked beside it. Sentiment and the other GEO KPIs →
AI-referred and AI-influenced sessions
AI-referral sessions from GA4, reported beside the branded and direct demand that moves with your Share of Model, as separate lines. The pipeline bridge →
Actions shipped
The fixes your team shipped this period, and which of the four numbers above each one moved. How to close the gap →
The board slide: five numbers and one sentence
Executives don't need the dashboard; they need the trend and what you are doing about it. Put the five numbers on one slide, each with its current value, prior value and direction, and put one sentence on top: "Our brand appears in X% of the AI answers our buyers see, up from Y%, and here is what we shipped to move it." That phrasing survives CFO scrutiny because it claims presence and trajectory, not magic attribution.
The two-week integrity rule: no number goes on the slide, or anywhere outside the team, unless it rests on at least two weeks of data. Single AI answers vary from run to run, and a figure quoted from one run may not survive the next.
How Do You Build an AI Visibility Dashboard for Reporting?
Build the AI visibility dashboard around the five scorecard numbers: one row each, with current value, 90-day trend, target and status, and a drill-down by engine, segment and prompt. This is the executive view Indexable AI recommends; the values below are illustrative, not benchmarks, so set targets from your own baseline.
| Metric | Current | 90-Day Trend | Target | Status |
|---|---|---|---|---|
| Share of Model (vs top two competitors) | 12% | +3 pts | 20% | On Track |
| Citation rate (money prompts) | 25% | Flat | 35% | At Risk |
| Description accuracy | 88% | +4 pts | 95% | On Track |
| AI-referred sessions (GA4) | 1,240 / mo | +18% | 1,500 / mo | On Track |
| Actions shipped | 9 this quarter | +3 | 8 | Met |
For AI visibility reporting, show each number per engine first and any blended view second, put the date and the number of answers behind every figure, and keep per-prompt detail in an appendix. A number with no engine and no sample size beside it is a claim, not a measurement.
How should a CMO read AI visibility metrics next to traditional KPIs?
Keep them on one page, for the same period. Put AI share next to organic share, AI-referred sessions next to organic sessions, and never blend the two into one number. Read the AI column for direction first, because the base is small and moves week to week. Read the organic column for scale. If AI share rises while organic share holds, the program is working. The AI search board report template lays the page out this way.
Get Your Baseline in 15 Minutes
Our AI Visibility Diagnostic calculates your current SoM, CFR rank, and APR across your top 10 prompts. No tools required. Just data.
Ready to Deploy AI SEO Agents?
See how 10 autonomous agents can transform your enterprise SEO. Talk to an architect for a live demo with your actual domain.
Talk to an ArchitectWhat Should You Do When an AI Visibility Metric Underperforms?
Tracking KPIs is useless without action. Here's what to do when each metric underperforms its target.
- Audit AI crawler access in robots.txt and at the edge, including OAI-SearchBot (ChatGPT search answers) and PerplexityBot
- Check that key copy is in server-rendered HTML, since many AI crawlers fetch pages without running JavaScript
- Earn mentions on the third-party sources engines cite for your prompts: reviews, communities and editorial lists
- Add visible, answer-first FAQ text to your top traffic pages
- Rewrite key sections so each claim stands alone in one sentence
- Earn links from the pages engines already cite for your prompts
- Identify the specific sources behind wrong or negative descriptions
- Publish a current fact page (pricing, positioning, what you do) and correct third-party listings
- Address criticism directly, with evidence, on the pages engines cite
- Run a competitive gap analysis on the prompts they win
- Build pages for the comparisons buyers ask AI about
- Accelerate authority-building on the sources engines cite
What Tools Do You Need to Measure AI Visibility?
You need three things to instrument AI visibility. Start with the manual baseline and automate from there.
Manual Baseline (Week 1)
Free but time-intensive. Run your top 20 buyer prompts on every engine your buyers use, three times each, and record every mention, citation and recommendation, and how accurately each answer describes you. This baseline is what every later number is judged against.
Platform Automation (Month 2+)
Choose a tool on criteria, not on the demo: it holds your prompt panel fixed, reports each engine separately, re-runs prompts and shows run-to-run variance, separates mentions from citations, and lets you export the raw answers. For current price bands, see the enterprise SEO pricing guide.
Attribution Integration (Month 3+)
Connect AI visibility to revenue in three spans: a GA4 channel group for AI-referral sessions, branded and direct demand trended beside Share of Model, and a "how did you hear about us" field that names AI tools. The pipeline bridge shows how to report each.
How to start
- Step 1. You can start by writing 20 buyer prompts in your buyers' words and running them on every engine three times.
- Step 2. Then, record Share of Model, citation rate and accuracy per engine. That is your baseline.
- Step 3. Next, you should put the five numbers on one slide with one sentence on top, and quote nothing that rests on less than two weeks of data.
- Step 4. Implement one fix per prompt you lose: a visible answer on the page, a self-contained claim, or a mention on a site the engines trust. Apply the fixes to the money prompts first.
- Step 5. If you want the agents to run the panel and the fixes, contact us to schedule a walkthrough.
What Will Your CFO Ask About AI Visibility?
When you present these KPIs to the executive team, prepare for these questions.
"How does this connect to revenue?"
Through three spans, reported separately: AI-referral sessions and the pipeline they open, branded demand that rises with Share of Model, and deals whose "how did you hear about us" answer names an AI tool. Illustrative arithmetic: if 5% of your pipeline touches AI search, a 20% improvement in that slice lifts total pipeline by 1%.
"What's the investment required?"
Platform cost plus the team time to act on what it finds: content, technical fixes and outreach. Current market price bands are in the enterprise SEO pricing guide; the real variable is how much of the fixing your team can absorb.
"What's the payback period?"
Payback is program cost divided by monthly incremental revenue, measured against a pre-program baseline. The AI visibility ROI guide walks through it; expect revenue to lag visibility by at least one sales cycle.
"What if we wait?"
You lose the baseline. AI answers shift week to week, and without a trend line that starts now you cannot tell a competitor's gain from noise, or show the board what your own work changed.
Frequently Asked Questions
What KPIs should a CMO use for AI search visibility?
Five: Share of Model against the top two competitors, citation rate on the money prompts, the accuracy of how AI answers describe the brand, AI-referred and AI-influenced sessions, and the actions shipped that moved them. Measure each per engine on a fixed panel of buyer prompts; the formulas are in the AI search metrics playbook.
How do I explain AI visibility trends to executives?
Show the trend, not a snapshot. One slide carries the five numbers, each with its prior value and direction, under one sentence: "Our brand appears in X% of the AI answers our buyers see, up from Y%, and here is what we shipped to move it." Quote nothing that rests on less than two weeks of data, because single AI answers vary from run to run.
Which KPIs matter most in a board-level AI visibility report?
Share of Model against named competitors and citation rate on the money prompts lead the report, because they show whether buyers see you and whether the answers send buyers to you. Accuracy, AI-referred and AI-influenced sessions, and the actions shipped complete the slide; per-prompt and per-engine detail belongs in the appendix.
How should CMOs interpret AI visibility metrics alongside traditional performance KPIs?
Read them as leading indicators next to the lagging ones you already report. Put the citation line beside the click line and impressions beside clicks, so a page that is quoted but not visited shows up. When clicks fall while citations hold or rise, the search surface has usually changed; that is not proof the content failed.
How do companies measure AI visibility as a KPI alongside SEO rankings?
They report rank and citation as two separate lines for the same set of queries, with separate owners. A page can rank and not be cited, or be cited from page three, so neither number explains the other; the AI search metrics playbook shows how ranking and citation have come apart.
What frameworks help CMOs compare AI brand visibility across brands or markets?
Use one structure everywhere: a fixed prompt panel per market and language, run on the engines buyers use in that market, with the same number of runs per prompt. Compare Share of Model and citation rate per engine rather than raw mention counts, because panel size and engine mix change the totals, and keep a separate baseline for each brand and market.
Make AI SEO Agents Your Unfair Advantage
Key takeaway: five numbers and one sentence, measured per engine on a fixed prompt panel; quote nothing built on less than two weeks of data.
The brands winning AI visibility aren't guessing. They're measuring these 5 KPIs weekly and acting monthly. Get your baseline and start building the dashboard that matters.
Related Reading
5 KPIs measure visibility. The 3 legs explain what to fix when KPIs slip.
Read The 3-Legged GEO Stool: Why Brand + Technical + Content Wins AI Search →