Traffic & revenue

Your pillar scores tell you how ready you are for AI agents. The dollar figure tells you what that readiness is worth. We take the gap between you and the category leader, size it against how much your category is being asked about, and translate it into a monthly USD range using typical conversion rates. It's always a range, never a single number — the range narrows as you connect real data sources (Stripe, Google Search Console, App Store analytics).

Scores measure readiness. Traffic and revenue estimate impact. We translate the gap between your readiness and the category leader's into a slice of the agentic demand in your category, and map that through category conversion behaviour to a USD range. Every input is visible as a concept; the band narrows as you connect more data sources.

Pillar scores roll up into the Appkitekt AI Index. The Index and per-agent breakdown drive your mindshare — your share of agent answers in your category. Mindshare, combined with category demand and the quality of your handoff, projects agentic visits. Those visits sit alongside human impressions from the downstream funnel. Together they resolve to a low–high USD revenue-at-stake band.

Each step is a separate, diagnosable stage. A high score with low mindshare means the structural fixes worked but agents haven't re-crawled yet. A high mindshare with low conversion means the agents recommend you but the handoff (AXO) is leaking.

Mindshare is the share of category prompts where your app appears in the agent's answer, weighted by rank. It's the agent-era equivalent of share-of-voice. We use mindshare — not the raw Index — to project traffic, because two apps with the same Index can have very different visibility in the prompts users actually run.

The Index tells you how fixable the gap is. Mindshare tells you how visible you are right now. Read both together. See Reading the numbers for the band definitions.

Agentic visits are sessions that originated from an AI agent: a click from a ChatGPT / Claude / Gemini / Perplexity / Copilot answer, an App Intents deep-link, or an MCP tool call that resolved to your surface. We estimate volume two ways:

• Coarse (ARO only) — a baseline range per category, scaled by your mindshare gap. Used on the ARO audit when no other data is connected. Always shown as a wide band.
• Refined (Agent Intelligence) — actual agent-referrer logs from your origin, GSC impressions for AI-overview queries, and App Store / Play impressions when relevant. The band narrows; confidence flips from coarse to partial or refined.

Agents don't replace humans; they sit in front of them. Many users see an agent answer, then click through, then land on your site or store listing. Human impressions (GSC for web, App Store / Play analytics for app) are the downstream volume that the agentic layer feeds. We surface them alongside agentic visits so you can see the full funnel — not just the agent slice.

The revenue band is the dollar version of the mindshare gap. We size the agentic demand in your category, take the share that sits between you and the leader, and translate it into a USD range using conversion behaviour and order value for that category. Defaults come from category benchmarks and are replaced by your own numbers whenever a connector (Stripe, store analytics, GSC) is live.

The output is always a low–high band, never a single number. The band narrows as inputs sharpen — see the confidence section below.

On the Agent Intelligence dashboard we show two USD bands side-by-side:

• Brand monthly at stake — your slice of the agentic pool, sized by your mindshare gap vs. the category leader.
• Market monthly at stake — the total agentic pool in your category. Useful for board decks and roadmap prioritisation.

Read together, the two bands separate "is the category big enough?" from "are we winning our share of it?". Optimizer's job is to push your share toward the leader.

Every revenue tile carries a confidence label. We never hide it:

• Coarse — audit only. Wide band, directional. Use for "is this worth fixing?".
• Partial — audit + at least one data connector (GSC, store analytics, or agent-referrer logs). Band narrows meaningfully.
• Refined — all relevant connectors live, with enough history. Band narrows again and is safe to share externally.

Revenue numbers are a lagging indicator. The honest sequence is: run an ARO audit (coarse USD band), push fixes in Optimizer, re-run the audit ~24h later (band should move), then watch Agent Intelligence for the refined band to shift over 2–4 weeks. Single-cycle moves of ±10% on the USD band are within noise; treat sustained two-cycle movement as signal.

• Not a guarantee. The band is a planning aid, not a forecast.
• Not attribution. We can't prove a specific sale came from a specific agent answer — no one can, yet.
• Not comparable across tools. Other products use different prompt sets, models, and assumptions; their numbers will not match ours.

For the underlying score logic, see Scoring & weighting. For the per-metric definitions, see Reading the numbers.