Answer-Engine Visibility (AEO / GEO)
Fixes: the business is invisible when a buyer asks an AI. More people now ask ChatGPT, Perplexity, Google AI Overviews, and Gemini “who’s the best ___ near me / for ___” instead of scrolling ten blue links. If the AI never names you, you don’t exist for that buyer — and unlike a Google rank you can’t even see it happening. This is the successor to local SEO: an effectiveness fix (lost demand) that compounds. Pairs with review-reputation-engine.md and Google Business Profile.
The metric that makes it sellable
You can’t sell what you can’t measure. This play is anchored on three numbers, tracked per engine over time: - Visibility Score — how often you’re surfaced for the prompts your buyers actually type. - Citation Share — of the sources the AI cites, how many are you vs. competitors. - Visibility Rank — where you sit vs. the named incumbents.
Baseline these in week 1 and the whole engagement has a scoreboard the client watches move.
The stack (with alternatives)
| Job | Recommended | Alternatives | Why / when to switch |
|---|---|---|---|
| Measure + act on AI-search visibility | J&M’s own aeo-engine (built 2026-07-06, ~/projects/aeo-engine) — reverse-engineered from Profound’s published formulas: daily prompt-set runs across Perplexity/ChatGPT/Gemini surfaces via OpenRouter, Visibility Score / SoV / Citation Share / Rank per engine, client HTML reports, the source-attack to-do list, and a one-shot baseline mode wired into the prospecting pipeline (pipeline.py --aeo). Runs on the VPS daily (6:15 cron → vault). Cost: ~$0.02/query, ~$30-70/mo per tracked category |
Profound (“Aim”) ($99 ChatGPT-only / $399 Growth real entry / enterprise custom — verified 2026-07-05); Peec AI, Otterly.ai, Scrunch; manual prompt runs | Self-owned engine is the default: zero seat cost, our data, feeds the pitch pipeline. Consider Profound only at client scale where their 10-engine coverage, panel-based Prompt Volumes (1.5B licensed prompts — the one piece we can’t self-build), and agent marketplace justify $399+/mo. |
| Execute the fixes it recommends | Claude skills (content, comparison pages, schema, FAQ) + earned-media outreach | Manual copywriting, a freelance writer | The tool surfaces what to fix; the lift comes from actually shipping the pages and getting cited. |
Agency-model note: the leverage move is J&M holds one Profound seat and services many clients through it, rather than each SMB buying their own — that’s how the tool’s cost pencils against an owner-operator budget. Confirm licensing allows it. Buy-trigger (opinionated): do NOT carry the $399/mo pre-revenue. Baseline prospects with the manual prompt-set audit (it doubles as the gift-first “oil-change test drive” artifact); buy the Growth seat when the first AEO retainer signs — one $1.5k/mo client covers the seat ~4×. Watch the credit math: ~400 credits/mo shared across clients caps how many prompt sets one seat can track — verify per-client burn before stacking a third client on it.
AOA build steps (Audit → Optimize → Automate)
- Audit — define the 20–50 prompts a real buyer would ask an AI in this category/geo. Run them across the tracked engines. Capture baseline Visibility Score, Citation Share, Rank, and the incumbents who own the recommendation slot and why they’re cited (which sources).
- Optimize — close the citation gaps: fix on-site facts and entity data, add schema, build the comparison / “best X for Y” / alternatives pages the engines pull from, and get placed in the earned-media sources they already cite (roundups, directories, trade newsletters, review sites). Fix the process before automating it.
- Automate — let the background agent monitor daily, flag anomalies (a competitor overtaking you, a new prompt you’re losing), and queue content tasks; review weekly. Log the deltas.
Time to ship
Baseline report week 1. First visibility-lift tasks live within the first month. Meaningful movement compounds over 60–90 days — set that expectation; this is not a same-week win like missed-call text-back.
Compliance / cautions
- No SMS/consent surface. The only rule is factual accuracy — never fabricate claims, credentials, or reviews to chase a citation; AI engines (and regulators) punish it and it burns trust.
- Earned-media outreach is real PR work, not spam — pitch genuine value.
What to log (renewal mechanism)
Per-engine Visibility Score / Citation Share / Rank deltas, pages shipped, placements earned, and the estimated demand captured. “You went from #7 and 13% to #2 and 48% visibility in this category” is the renewal story.