The AEO Discovery Call: Questions an Agency Asks Before Quoting AI Visibility Work

Most agency guides to answer engine optimization teach you how to pitch. This is the earlier, less glamorous conversation that decides whether the pitch is even worth building: the discovery call where you find out what the prospect actually needs, whether you can deliver it, and what it should cost. Get this call right and the proposal writes itself against a scope you can defend. Get it wrong and you either underquote a project that eats your margin or overquote one you lose to an agency that asked better questions. The discovery call is diagnosis, and the quality of your diagnosis is the first proof of competence the prospect will judge you on.

Diagnose before you interview

Do not walk into the call with only questions. Walk in with evidence. Before the call, run the prospect's brand and two named competitors through the AI engines your prospect's buyers actually use, ChatGPT, Claude, Gemini, and Perplexity, on ten to fifteen buyer-intent prompts drawn from their category. You now have a live snapshot of where they stand, which turns an abstract conversation about AI visibility into a specific one about their own invisibility. Showing a prospect an answer where a competitor is named three times and they are absent does more than any slide. The mechanics of running that snapshot fast are in how to run an AEO competitive teardown for a prospect in under an hour, and the read on why the engine chose the sources it did is the same forensic loop that powers how OnlyAEO works.

That pre-call diagnosis also protects you. It tells you whether the prospect is a cold start with zero citations or a brand with partial visibility that needs defending and expanding, and those are two different engagements at two different prices. You want to know which one you are quoting before the call, not after you have already named a number.

Question set one: what is the real goal

Prospects arrive asking for AI visibility, but visibility is never the actual goal. It is a proxy for something they can measure: pipeline, trial signups, brand defense against a competitor, or a board-level answer to "why are we not in ChatGPT." Your first job is to get behind the proxy.

Ask what would have to be true in ninety days for this to be worth it. Ask what triggered the search for an AEO agency now, because the trigger tells you the real stakeholder and the real fear. A CMO who just got asked by the board why a competitor shows up in AI answers has a different goal from a demand gen lead whose organic pipeline is flattening. Ask who else has to be convinced, because the person on the call is rarely the only signer. The way an agency frames a client's absence without sounding defensive is worth rehearsing here, and it is laid out in how an agency explains a client's AI invisibility without sounding like an excuse.

If the prospect cannot state a goal more specific than "show up in AI," slow down. A vague goal produces a vague scope, and a vague scope is where retainers go to die. You are better off spending the call narrowing the goal than racing to a number.

Question set two: the current AI footprint

Now use your pre-call diagnosis as a lever. Walk the prospect through what you found, then ask the questions only they can answer.

Ask whether they have ever measured their AI citation share before, and if so, how. The answer tells you their sophistication and whether they will trust your baseline. Ask which competitors they consider the real threat, because the two you tested may not be the two that keep them up at night. Ask whether they know if AI engines can even reach their site, since a surprising number of prospects are blocked at the crawler, the CDN, or the renderer without knowing it. Ask whether anyone has ever told them AI is stating something wrong about their brand, because fixing a factual error in model memory is a different task from earning a first citation.

These questions do two things. They surface scope you would otherwise miss, and they demonstrate that AI visibility has moving parts the prospect had not considered, which is precisely why they need an agency rather than a freelancer with a checklist.

Question set three: the content and technical baseline

The single biggest driver of your quote is what the prospect already has to work with. An AEO program that starts from a thin site with no answer-structured content costs far more than one that refreshes an existing library of ranking pages. So you have to inventory the raw material.

Ask how much published content they have and whether any of it already ranks on Google, because pages with existing authority are the fastest path to a first AI citation through a refresh rather than a net-new build. Ask whether they have structured data, an llms.txt file, or any AI-specific infrastructure, and be ready for the answer to be no. Ask who controls their site, because a marketing team that has to file a ticket with engineering for every schema change is a slower, more expensive engagement than one with CMS access. Ask whether their documentation, pricing, and comparison pages are public and crawlable, since those are the surfaces buyers ask AI about most and the ones most often locked behind logins or rendered in ways engines cannot read.

You can hand a prospect an immediate, no-cost win here by pointing them to the free llms.txt generator so they leave the call with a concrete artifact, and the ongoing version of that infrastructure is the AI Feed Engine that keeps the feed current as content ships.

Question set four: internal capacity and ownership

AEO is not a fire-and-forget deliverable. Someone on the client side has to approve content, provide subject matter input, and act on the reporting. If that person does not exist, your engagement will stall no matter how good your work is.

Ask who will own AI visibility internally and whether that person has time for it. Ask who approves published content and how long approval usually takes, because a two-week approval cycle changes your production cadence and your timeline promises. Ask whether they have in-house writers or expect you to produce everything, since that is the difference between a coaching retainer and a full production one. Ask what has failed before, because a prospect who churned through two SEO agencies is telling you about a decision-making problem you are about to inherit.

The honest read on capacity often reveals that the prospect needs a smaller, sharper engagement than they think. A 30-day pilot that proves the model before committing to a full retainer protects both sides, and the exact scope for one is in how to run a 30-day AEO pilot that wins the client retainer.

Question set five: budget and the definition of success

You cannot quote against silence. By this point in the call you have earned the standing to ask about money directly, and you should.

Ask what range they have in mind, and if they deflect, anchor with public benchmarks so the number is not a mystery. AEO agency retainers in 2026 commonly land between two thousand and twenty thousand dollars a month, with the mid-market B2B and SaaS sweet spot around four to twelve thousand, and standalone audits running from several hundred to several thousand dollars depending on scope. Sharing that range, backed by sources like an AEO agency pricing breakdown from Discovered Labs and The Answer Engine's pricing guide, signals confidence and filters out prospects who were expecting a few hundred dollars.

Then pin down success. Ask what they will look at in ninety days to decide whether to renew. Push for leading indicators, not just revenue, because revenue lags the work by a quarter or more on a channel where the AI answer that influenced the buyer leaves no referrer. The reason AEO gets funded despite that lag is that the traffic performs when it arrives, and buyers who click through from an AI recommendation convert well above organic, a pattern documented in AirOps' analysis of AI referral conversion rates. Set the expectation now that the early scorecard is citation share and retrieval presence, not closed deals, or you will lose the retainer at renewal to a metric that was never going to move in time.

Reading the red flags

The discovery call is also a qualification filter. Some prospects should get a polite pass rather than a proposal. Walk if they demand a guaranteed citation or a fixed ranking, because no agency controls what the engines cite and promising it is how you lose the relationship in month three. Walk if there is no internal owner and no willingness to appoint one. Walk if the budget and the goal are irreconcilable, since a prospect who wants category leadership in AI answers on a five-hundred-dollar budget is not a client, they are a future bad review. And be cautious with the prospect who has churned through multiple agencies and blames all of them, because the common factor may be a decision process you cannot fix.

A prospect who has no case study to point to is not a red flag, though, it is the normal state of this market, and closing them without proof is a solvable problem covered in how to land your first AEO client without a single case study.

From call to quote

By the end of a good discovery call you have five things: a specific goal tied to a stakeholder, a live picture of the prospect's AI footprint against named competitors, an inventory of their content and technical baseline, an honest read on internal capacity, and a budget range with a success definition. That is a scope. Price it against the effort the baseline actually requires, stage it so the prospect can prove the model before committing, and show them a documented arc from invisible to consistently cited, like the one in the FastTrackr AI case study. A proposal built on that foundation is defensible line by line, and it is the difference between winning work you can deliver profitably and winning work that quietly loses you money.