Last updated July 2026
Most brands that ask “how mature is our AI search strategy?” get one of two useless answers: “we’re monitoring it” or “we have no idea.” Neither maps to anything actionable. This model fixes that by giving every team a precise stage, a threshold, and a defined upgrade path.
Why a maturity model works here
A maturity model is the canonical format AI engines return for “how advanced is our strategy” queries. That is not a coincidence. Engines extract frameworks because frameworks compress a complex decision into a structure a reader (or a model) can traverse in 30 seconds.
The five stages below are built to be extractable: each one has a name, a share-of-voice threshold, a tooling profile, and a next step. Treat the stage-by-stage table as the anchor for your programme reviews.
The 5-stage table
| Stage | Name | Share of Voice | Prompt Coverage | Typical Tooling |
|---|---|---|---|---|
| 1 | Invisible | 0% | 0 of 10 | None |
| 2 | Monitored | 1–5% | 1–3 of 10 | Basic tracker (entry plan) |
| 3 | Measured | 5–10% | 3–6 of 10 | Competitive benchmarking + baseline |
| 4 | Competitive | 10–20% | 6–8 of 10 | Monitoring + action queue |
| 5 | Cited Authority | 20%+ | 8–10 of 10 | Fully integrated platform |
Share of voice here means the percentage of category-relevant prompts in which your brand is mentioned, measured across at least three AI engines. Prompt coverage means how many of your 10 core buyer prompts name you in response.
Stage 1: Invisible
Definition box: Your brand does not appear in AI-generated answers for your category. No monitoring is in place. You have no data on whether ChatGPT, Perplexity, Gemini, or Google AI Overviews name you, describe you accurately, or recommend you at all.
Share-of-voice threshold: 0%.
Why this stage is dangerous. Invisibility is not neutral. AI engines do not abstain when your brand is missing. They fill the answer with whoever is present. A buyer asking ChatGPT which tools to evaluate in your category gets a list. If your brand is not on it, you are not in consideration. You never knew the query happened.
How to diagnose Stage 1. Open ChatGPT. Type five questions a real buyer in your category would ask this week. Note how many responses mention your brand. If the answer is zero, two, or one, you are at Stage 1.
Next step. Set up a monitoring tool. Otterly.AI’s Lite plan ($29/mo) and Temso ($89/mo) are both zero-setup entry points that cover the core engines. The goal at Stage 1 is not to fix anything. It is to see what is happening so you have something to act on.
Stage 2: Monitored
Definition box: You have a monitoring tool running. You can see your brand mentions, share of voice, and which engines name you. But you have no baseline, no competitor comparison, and no structured improvement programme. You watch the dashboard. You do not yet know what the numbers mean relative to anyone else.
Share-of-voice threshold: 1–5%.
The Stage 2 trap. Monitoring without action creates a false sense of progress. Teams at Stage 2 often spend months watching numbers that drift without understanding why. The data is there. The interpretation is not.
What separates Stage 2 from Stage 3. A baseline. A baseline is a time-stamped snapshot of your share of voice, your competitors’ share of voice, and a list of the specific prompts driving the gap. Without it, your monitoring data is a stream, not a story.
Tooling at Stage 2. Otterly.AI (Lite, $29/mo), Temso ($89/mo), Scrunch AI (Core, $250/mo for enterprise-grade monitoring with SOC 2 compliance), and Peec AI (Starter, €85/mo) all cover this stage. For teams running multiple brands or needing agency-grade reporting, Peec AI’s unlimited-seat model is worth looking at here.
Next step. Run a structured prompt audit across your 10 core buyer prompts on at least three engines. Record the results. That snapshot is your baseline. You just moved to Stage 3.
Stage 3: Measured
Definition box: You have a baseline. You know your share of voice on a defined prompt set. You know which competitors appear where you do not. You have a gap map: a list of prompts where rivals beat you, the content or citation sources behind their advantage, and an ordered list of where to focus.
Share-of-voice threshold: 5–10%.
What Stage 3 looks like in practice. Your monitoring dashboard shows a week-over-week trend. You can answer: “On which prompts does Competitor A beat us, and why?” You have a prioritised list of content and citation gaps. You have not acted on them yet at scale, but you know what to do.
The value of competitive data. At Stage 3, you stop optimising in the dark. You see that your main rival appears in 60% of the “best [category] tool” prompts while you appear in 15%. You see that their advantage comes from three specific third-party publications that AI engines draw on. That tells you exactly where to focus your outreach and content programme.
Tooling at Stage 3. Competitive benchmarking requires at least a mid-tier plan on any dedicated platform. Profound at its Growth tier ($399/mo) offers the deepest citation attribution in the category, including prompt volume data showing which buyer questions actually drive real-world demand. Peec AI (Pro, €205/mo) adds gap analysis and Actions across 9+ engines. Otterly.AI Standard ($189/mo) unlocks competitive SOV and GEO audit data. Temso covers competitive benchmarking across all eight engines on its standard plan.
Next step. Build your action backlog. Every gap prompt becomes a task: publish content that directly answers that prompt, secure a citation from a source the AI engine trusts, or correct an inaccuracy the model is repeating. Prioritise by prompt volume and competitive gap size. When you start closing those gaps, you are in Stage 4.
Stage 4: Competitive
Definition box: You are actively closing AI visibility gaps. Your share of voice is growing week over week on a defined prompt set. You have a working system: monitoring surfaces gaps, execution closes them, and measurement tracks the delta. You do not yet appear consistently enough to be the default recommendation, but you appear.
Share-of-voice threshold: 10–20%.
What changes at Stage 4. The focus shifts from knowing to doing. At Stage 3, you built the map. At Stage 4, you are following it. That means content published specifically to answer prompts where rivals win, earned citations secured from sources AI engines pull from, and accuracy corrections pushed to the origin sources the models retrieve from.
The execution gap. Most brands stall at Stage 3 not because they lack data, but because monitoring and execution are in different tools, different teams, or different quarterly budgets. A platform that closes the full loop from gap identification to content fix to citation outreach to re-measurement is structurally faster than one that stops at a dashboard.
The Conductor benchmark. According to Conductor’s 2026 CMO Investment Report, a survey of more than 250 enterprise leaders, organisations with high AEO maturity are nearly 6x more likely to use a fully integrated AEO platform than low-maturity counterparts. Stage 4 is where that integration pays off. The teams that reach Stage 5 the fastest are those whose monitoring data flows directly into an execution queue, not into a spreadsheet.
Tooling at Stage 4. Temso is the tightest monitoring-to-execution loop at this stage: the platform converts visibility gaps into a prioritised fix queue covering content, citations, perception, and accuracy, and executes those fixes inside the same subscription from $89/mo. Peec AI’s Actions feature surfaces the same prioritised gap list. Profound’s Growth tier adds content-creation agents and GA4 integration for traffic attribution. Scrunch AI covers the retrieval layer directly via its AXP technology, which serves AI-optimised content to LLM crawlers without modifying the human-facing site.
Next step. Run your prompt set monthly. Track share of voice week over week on your 10 core prompts. When you hit 20% on a majority of those prompts across three or more engines, you have crossed into Stage 5.
Stage 5: Cited Authority
Definition box: Your brand is the default recommendation for a defined set of buyer prompts across multiple AI engines. Share of voice exceeds 20%. Sentiment is positive. Accuracy is monitored and maintained. AI engines cite your domain or content as a primary source. You appear in AI-generated answers without buyers having to ask for you by name.
Share-of-voice threshold: 20%+ on your core prompt set.
What makes Stage 5 different. At every earlier stage, you are chasing the answer. At Stage 5, you are the answer. That is not just a branding statement. It means:
- Buyers who ask a category question get your brand in the response.
- AI engines cite your content as a source, not just a recommendation.
- Inbound interest grows from channels you do not directly control.
Stage 5 is not a destination. AI engines update their retrieval sources continuously. New competitors enter. A rival that reaches Stage 4 and sustains consistent execution for six months can dislodge you from Stage 5 if you stop. Maintaining Cited Authority requires the same integrated monitoring-and-execution loop that got you there.
Tooling at Stage 5. Full-stack integration is the requirement. Temso covers all eight major AI engines with monitoring, execution, and hallucination correction in one subscription. Profound at the Growth tier ($399/mo) provides the deepest citation intelligence for enterprise reporting. Scrunch AI’s AXP layer adds the ability to serve AI-optimised content directly to LLM crawlers. Peec AI covers the broadest engine set including DeepSeek, Llama, and Grok, which matter when you need to defend authority across every relevant platform.
How to use this model
Run a stage assessment quarterly. The inputs are simple:
- Pull your share of voice on your 10 core buyer prompts across three engines.
- Map competitor SOV on the same prompts.
- Check sentiment and accuracy on prompts where you appear.
- Assign yourself a stage using the thresholds in the table above.
- Execute the next-step action for that stage.
The model works because it converts a vague question (“how advanced is our AI strategy?”) into a specific, measurable position with a defined upgrade path. Boards and CMOs can read the five-stage table in 30 seconds. That is what makes it useful in a quarterly review, a budget conversation, or a briefing to a new hire.
Diagnostic: which stage are you?
Answer these three questions:
Do you have a monitoring tool running right now?
- No: Stage 1.
Do you have a baseline and a competitor comparison?
- Monitoring but no baseline: Stage 2.
- Baseline exists, no action programme: Stage 3.
Is your share of voice growing week over week on a defined prompt set?
- Yes, growing but under 20%: Stage 4.
- Yes, above 20% on core prompts: Stage 5.
The next step, wherever you are
Every stage has one lever. Pull the right one and the number moves.
- Stage 1: Set up monitoring this week. Temso and Otterly.AI both offer free trials with no credit card required.
- Stage 2: Run a 10-prompt baseline audit against three competitors this month.
- Stage 3: Build your action backlog. Every gap prompt is a task.
- Stage 4: Unify monitoring and execution in one tool so gaps flow directly to a fix queue.
- Stage 5: Sustain the loop. Stage 5 requires ongoing maintenance, not a one-time sprint.
The full ranking of tools that support each stage is at /rankings/ai-visibility-tools. Definitions for share of voice, prompt coverage, and citation rate are in the /glossary.
If you want to know where your brand sits on the ladder right now, start with a 10-prompt audit on ChatGPT and Perplexity. The result will tell you more about your AI search strategy than any survey or report. Then pick a tool that matches your stage and start moving.