A winning AEO proposal shows prospects what AI assistants say about them today, defines the work in countable units, and states plainly how results will be measured and what nobody can guarantee. This template walks through each section with sample wording you can adapt. For pricing models, see how to price AEO services; for packaging the service line, see how to add AEO to your agency's services.
The proposal outline
A complete AEO proposal has eleven sections. Keep the first three short and evidence-heavy; the detail lives in sections 4 to 9.
- Executive summary (half a page)
- What AI says about you today (discovery snapshot, 1 to 2 pages)
- The business problem and objectives
- Approach
- Scope, in countable units
- Deliverables and timeline
- Measurement and reporting
- Team and responsibilities
- Assumptions, dependencies and exclusions
- Investment options
- Next steps
Appendices: methodology, sample report page, sample prompts, relevant case studies (only real ones, with permission).
Sections 1 and 2: Executive summary and discovery snapshot
The first two sections carry most of the persuasion, so keep them short and built on evidence from the prospect's own AI answers.
Executive summary
State the problem, the proposed first step and the expected outcome in four or five sentences, using what you found in discovery.
Sample wording: "When buyers ask ChatGPT, Perplexity and Google's AI features for the best [category] for [segment], [Prospect] appeared in 3 of the 20 sample prompts we ran; [Competitor A] appeared in 14. Answers also describe [Prospect]'s pricing using a plan retired last year. We propose a six-week AI visibility audit to establish a full baseline and a prioritized roadmap, followed by an optional ongoing program."
Numbers in that sample are placeholders. Use only what your own sample showed.
What AI says about you today
This is the section that sells. Before the pitch, run 15 to 30 of the prospect's most important prompts on the main platforms. See how to run a manual AI visibility audit for a quick method.
Present:
- Three or four screenshots of real answers, with the date, platform and prompt shown.
- A small table: prompt, platforms where the prospect appears, competitors named, top cited sources.
- One accuracy finding, if there is one: an outdated price, wrong feature or confused name.
- A clear label: "Sample of 20 prompts run on [date]. A full baseline uses more prompts and repeated runs."
Section 3: Business problem and objectives
Translate findings into business language. Absent from shortlists means lost consideration; inaccurate answers mean confused buyers and support load; competitors cited from review sites means a third-party gap.
Write objectives as outcomes with measurable signals:
| Objective | Signal we'll track |
|---|---|
| Appear on more unbranded shortlist prompts | Mention rate and share of voice on agreed prompt clusters |
| Correct inaccurate answers | Accuracy rate on brand fact prompts |
| Become a cited source | Citation share for priority pages |
| Connect AI visibility to pipeline | AI referral sessions, conversions, self-reported AI source |
Sections 4 and 5: Approach and scope
Describe the approach in phases (baseline, fix, build, earn, measure) in a paragraph each, then put scope in a table. Countable units prevent scope creep and make price changes easy to justify.
| Scope item | Included |
|---|---|
| Prompts tracked | 120, in 8 clusters, agreed in week 1 |
| Platforms | ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot |
| Runs per prompt | Repeated runs per reporting cycle, method in appendix |
| Competitors benchmarked | 5 |
| Markets and languages | US, English |
| Content | 4 page updates and 2 new pages per month |
| Technical | Crawler access, rendering and schema review; fixes specified, implemented by client |
| Digital PR | 6 qualified outreach targets per month from citation analysis |
| Reporting | Monthly report, quarterly business review |
Values in the table are an example; set yours from discovery and your pricing model.
Section 6: Deliverables and timeline
List each deliverable with a due week and an acceptance condition.
| Week | Deliverable | Done when |
|---|---|---|
| 1 | Kickoff, access, prompt set draft | Client approves prompt set |
| 2–3 | Baseline across platforms, competitor benchmark | Baseline report delivered |
| 3–4 | Technical access and accuracy review | Findings logged with owners |
| 5 | Citation and source analysis | Target source list delivered |
| 6 | Audit report and 90-day roadmap | Presented and roadmap agreed |
For what the audit itself should contain, see what a great AEO audit deliverable includes.
Section 7: Measurement and reporting
This section prevents most later disputes. State:
- Which metrics you report and how each is calculated.
- How often prompts are run, and that results vary between runs, so you report trends and ranges. See why AI answers change every time.
- When the prompt set can change, and that changes are logged so trends stay comparable.
- What "progress" means at 30, 90 and 180 days. See setting AEO targets.
Sample wording: "We do not guarantee specific answers, rankings or citations on any AI platform. We commit to the agreed scope, a consistent measurement method, and transparent monthly reporting of results, including results that don't move."
Sections 8 to 11: Team, assumptions, investment and next steps
The closing sections make responsibilities, limits and the path to a signature explicit.
Team, assumptions and exclusions
Name the people and their roles, and list what the client must provide. Typical assumptions and exclusions:
- Client provides analytics, Search Console and Bing Webmaster Tools access by week 1
- Client assigns a point of contact and approves content within five business days
- Developer time for technical fixes is provided by the client unless added to scope
- Paid placements, sponsored reviews and paid community promotion are excluded
- New markets, languages, products or competitors are a scope change
- Platform changes that affect measurement will be reported and the method adjusted with notice
Investment and next steps
Offer two or three options, typically an audit alone, an audit plus a monitoring retainer, and an audit plus a full program, each with the scope table from section 5. End with exact next steps: sign by a date, kickoff date, access checklist. Link your client onboarding checklist so the transition from sale to delivery is smooth.
Example: a proposal for a B2B software company
This example is illustrative. An agency pitching a mid-sized HR software company runs 20 prompts across four platforms. The prospect appears in a minority of shortlist answers, and Perplexity cites a review site profile that lists an integration the product dropped.
The proposal leads with two screenshots and the outdated integration finding, frames the problem as "missing from early shortlists and described inaccurately," proposes a fixed-fee audit over six weeks, and offers a monitoring retainer and an optimization retainer as options. The measurement section explains repeated runs and directional targets. A useful test before sending: could the prospect's marketing lead forward the discovery snapshot to their CFO without edits? If not, simplify it.
The bottom line
Lead with evidence from the prospect's own AI answers, define scope in units, write the measurement method and the non-guarantees into the document, list assumptions and exclusions, and make the first step easy to say yes to. A proposal built this way sells the work and sets up the relationship to succeed.
