What a 90-Day AEO Pilot Should Deliver Before You Sign a Retainer
Day 0-30, 31-60, and 61-90 deliverables that prove citation craft before you lock an AEO retainer.
Most AEO retainers in 2026 start with a deck, a content calendar, and no proof the vendor can move citation share on the queries that form your shortlist. A paid 90-day pilot exists to force that proof before you lock months of budget.
Here is the shape: what a pilot is for, the Day 0-30 / 31-60 / 61-90 deliverables, the go/no-go scorecard, and the red flags that should end the engagement early. Lock the delivery model first with AEO agency vs AEO consultant. Scope the commercial work on AEO for B2B SaaS or AEO for B2B. Retainer bands live on pricing.
G2's 2026 AI Search Insight Report, fielded among 1,076 B2B software buyers in March 2026, found 51% now start research with an AI chatbot more often than Google, and 69% chose a different vendor than expected because of chatbot guidance. The pilot has to move visibility on that surface, not ship blog volume.
What a 90-day AEO pilot is for (and what it is not)
Answer engine optimization (AEO) is the work of earning citations and mentions in AI answers across Google AI Overviews, AI Mode, ChatGPT, and Perplexity. A pilot is a paid proof on a fixed prompt set. It is not a discounted retainer, a free trial, or a content factory with a new label.
As of March 2026, Ahrefs' analysis of 863K SERPs and 4M AI Overview URLs found only ~38% of cited pages also ranked in the organic top 10 for the same query, down from ~76% in July 2025. Ranking KPIs alone will not tell you if the pilot worked.
Write the RFP and pilot gate together. See how to write an AEO RFP in 2026.
Day 0-30: baseline, ownership, and measurement stack
Week one should produce assets you keep even if you fire the vendor on day 91.
- Fixed prompt set: 20-40 commercial queries (category, comparison, alternatives, pricing intent) across AI Overviews, AI Mode, ChatGPT, and Perplexity.
- Baseline citation audit: share of voice vs a named competitor set, by engine, on day 0.
- Ownership map: who owns the prompt library, entity map, CMS edits, and dashboards after the pilot.
- Measurement stack: Search Console Generative AI performance report for Overviews and AI Mode (rolled out worldwide as of Aug 31, 2026, per Google Search Console Help), plus GA4 referral views and a weekly prompt log for assistants.
The Free AEO Report Card is Austin Heaton's diagnostic of citation gaps and money-page readiness. Use a baseline like it so day-90 results compare to a real day-0 number, not a slide.
Want a second set of eyes on the pilot SOW before kickoff? Book a 30-minute call.
Day 31-60: money pages and entity cleanup
Months two should change extractable commercial pages, not fill a blog calendar.
- 3-5 money-page rewrites built for citation: comparison, alternatives, category, and pricing-adjacent pages with clear claims, dates, and entities.
- Entity consistency pass across site, LinkedIn, directories, and any profiles engines already treat as sources.
- Schema and snippet eligibility: Google's AI features guidance is blunt: indexed and snippet-eligible pages can appear as supporting links. There are no extra proprietary technical requirements.
- Weekly scorecard: citation share by engine on the fixed prompt set, plus which URLs engines cited.
If the vendor is still debating tools while your money pages stay untouched at day 45, you do not have a pilot. You have a discovery project.
Day 61-90: citation movement and the go/no-go gate
The last 30 days exist to decide the retainer, not to invent new scope.
- Citation trajectory: day-90 share of voice vs day-0 on the same prompts and engines. Direction matters more than a guaranteed rank.
- Named operators still in Slack: the people who shipped month one should still be doing the work.
- Owned artifacts: prompt library, entity map, rewritten money pages, and dashboards in your accounts.
- Honest gaps: which engines moved, which did not, and what the retainer would fund next.
As of Q3 2026, Search Console's Generative AI report shows impressions only for Google AI features. It does not cover ChatGPT or Perplexity, and it does not show competitor share. Treat GSC as one surface inside a wider scorecard.
If you need help turning this into a day-90 scorecard your CMO will trust, book a 30-minute call.
The go/no-go scorecard before you sign the retainer
| Gate | Pass signal | Fail signal |
|---|---|---|
| Baseline | Day-0 prompt log + competitor set delivered in week 1 | No fixed prompts; we will track later |
| Craft | 3-5 money pages shipped and indexed | Blog posts only; commercial pages untouched |
| Measurement | Weekly citation share by engine | Sessions and rankings only |
| People | Named operators still in Slack at day 90 | Pitch team vanished after kickoff |
| Ownership | You hold prompts, entities, dashboards | Vendor-locked tools with no export |
Fail two or more gates and do not convert. Pay for the pilot. Walk. The downside of a hard gate is fewer retainers start. The upside is you stop funding content mills that renamed SEO as AEO.
Red flags that should end the pilot early
- Ranking guarantees for AI Overviews or ChatGPT
- No fixed prompt set by day 14
- Content volume as the primary KPI
- Refusal to name who does the work in the SOW
- Dashboards you cannot export
- Identical SEO reporting with AEO find-and-replaced
Austin Heaton's take: pay for proof, then buy the retainer
Austin Heaton's take: a 90-day pilot is the filter after the RFP. If a vendor cannot baseline citation share, rewrite money pages, and show trajectory on a fixed prompt set, they are not ready for a retainer. That position has a cost: you may burn a quarter and restart. For teams where G2's 51% already start in a chatbot, starting without proof is the larger risk.
If you want the operator model for SaaS or broader B2B, start on AEO for B2B SaaS or AEO for B2B, then check pricing.
Lock the prompt set this week. Insist on day-0 baselines. Convert only when the scorecard passes. Book a call if you want Austin Heaton to pressure-test the pilot SOW.
Frequently Asked Questions
What should a 90-day AEO pilot deliver?
A fixed 20-40 prompt set, day-0 citation baseline, 3-5 money-page rewrites, entity cleanup, weekly citation share by engine, and owned dashboards before any retainer.
How is an AEO pilot different from a retainer?
A pilot proves craft on a fixed scope and prompt set. A retainer funds ongoing citation ops after the go/no-go gate. Do not start with a 12-month lock if you have no day-0 baseline.
Which KPIs prove the pilot worked?
Citation share vs a named competitor set by engine on the same prompts, money pages shipped, and operator continuity. As of March 2026, Ahrefs finds only ~38% of AI Overview citations also rank top 10 for the same query.
Should you pay for the pilot?
Yes. Free pilots attract mills. Paid pilots fund real baselines and money-page work. Retainer terms live on pricing.
When should you refuse to convert to a retainer?
If two or more gates fail: no baseline, blog-only output, ranking-only reporting, missing operators, or locked dashboards.
Where does the Free AEO Report Card fit?
The Free AEO Report Card is Austin Heaton's diagnostic of citation gaps and money-page readiness. Use it, or an equivalent baseline, as day 0 so day 90 is comparable.
