How to choose an AEO agency in 2026

Pick an AEO agency that starts from your buyers' real questions and shows how it measures AI answers (questions, engines, re-checks). It should tie results to pipeline, publish its price and refuse to promise rankings or citations. Then test it on a 90-day scope with written goals.

Your buyers now ask AI what to buy. In a G2 survey, 51% of B2B software buyers start research in an AI chatbot more often than Google. The same survey found that 69% say a chatbot led them to a different vendor than planned. If your brand is missing from those answers, an AEO agency can help. Picking the wrong one costs you a quarter.

This guide works the same whether you hire us or someone else. It walks through six steps: brief, shortlist, questions, proof, price and a 90-day test.

Do you need an AEO agency at all?

You need one if buyers in your category ask AI for advice and you have no one in-house to run the work. You do not need one if you lack a clear product and buyer, or if no one on your side can approve content each week.

Signs you should hire:

• A buyer told sales they found a competitor through ChatGPT.
• Your board asks what AI says about you, and no one knows.
• Organic traffic is flat and paid costs keep rising.
• A competitor keeps getting named in AI answers for your category.

Signs you should wait:

• You are still finding product-market fit.
• No one can review drafts or approve fixes on your site.
• You need leads in the next few weeks. AEO compounds; it is not a paid ads switch.

Step 1: How do you write the brief?

Write one page before you talk to anyone. A brief makes every agency answer the same question, so you can compare them.

Put these in it:

1. Who buys from you: role, company size, sales motion (self-serve or sales-led).
2. The questions they ask: copy them from sales calls, demos, support tickets and reviews.
3. Who AI names today: ask ChatGPT and Perplexity your top buyer questions and note every brand.
4. What success means: demos, trials or pipeline from AI and search, not traffic alone.
5. Budget and term: a monthly range and how long you can commit.
6. Who approves work: the person on your side who signs off content and site changes.
7. Start date and any deadline: a launch, a rebrand or a migration.

Step 2: How do you build a shortlist?

Look in three places: directories, lists and referrals. Then check each name against your brief.

Each AI engine leans on different sources. From our platform, in September 2026, 45% of the links ChatGPT cited were directories and review sites like Clutch and G2. Of the sources Perplexity read, 16% were agency-written "best of" lists. So the names AI gives you depend on which engine you ask.

Read agency lists with care. On 1 October 2026 we read five "best AEO agency" lists written by agencies. All five ranked their own author first: NoGood, First Page Sage, Optimist, Omnius and Unified Platforms. That does not make them wrong. It means you should judge each name on its own facts.

Step 3: Which questions should you ask?

Ask every agency the same 10 questions. The answers tell you more than any deck. Here is what a good answer and a red flag sound like.

On question 8, Google says no extra requirements or special optimisations are needed to appear in AI Overviews and AI Mode. And llms.txt is still a proposal by Jeremy Howard (2024), not a standard the engines have said they use.

Step 4: How do you check an agency's proof?

Ask for three things behind every result: the baseline, the time window and how it was measured. A number without those is a claim, not proof.

In our read of 52 published agency case studies, about 10 of the 47 real engagements report leads, signups or revenue. Most report traffic or visibility.

Watch for the same result told two ways. One Discovered Labs case is quoted on Optimist's list as 4x, from 550 to 2,300+ AI-referred trials. On HubSpot's case-study roundup it reads 6x, from 575 to 3,500+ (both as of 1 Oct 2026). Ask the agency for the primary source.

Be careful with AI screenshots too. Answers change from one ask to the next and between engines. Our platform found that only about 1 in 8 of the companies named (12%) appeared in both ChatGPT's and Perplexity's answers. One screenshot shows one answer on one engine. Ask which questions and engines sit behind it, and how often each was re-checked.

Step 5: What should you pay, and on what terms?

Published AEO and GEO agency retainers run from about USD 3,000 to over USD 15,000 a month, as of 1 October 2026. Many agencies publish no price at all.

A few published examples, checked 1 October 2026:

• Optimist: advisory from USD 3,000 a month, full service from USD 4,000 a month, month to month with 30 days notice.
• LoudFace: from USD 5,000 a month, 3-month first engagement.
• DerivateX: USD 5,000, 8,000 or 12,000 a month, plus a citation budget, after a 90-day pilot.
• Discovered Labs: EUR 6,995, EUR 10,995 or from EUR 15,995 a month, month to month with 30 days notice; EUR 5,995 or 8,995 with a 3 or 6 month commitment.

Put these in the contract:

• The deliverables per month (pages, fixes, mentions, reports).
• The prompt set you will track, by engine.
• How and when results are reported.
• Who owns the content, data and accounts.
• The notice period and what you keep if you leave.
• Disclosure rules for any community or mention work.

Step 6: How do you run a 90-day test?

Agree written goals at kickoff, then review at day 30, 60 and 90. A fixed test shows how the agency works before you commit to a long retainer.

Split goals into two kinds. Goals the agency controls: pages shipped, fixes done, prompts tracked, AI channel set up in analytics. Goals it can only influence: being named in AI answers, demos and pipeline. Judge the first kind hard. Track the second kind honestly, over time.

Expect early signs, not full results. Agencies state their own timelines: Discovered Labs says first results in 1 to 2 weeks and full impact in 3 to 4 months. Optimist says results in 1 to 3 months, compounding over 6 to 12.

What are the red flags?

Walk away from any of these:

• Promised rankings or citations. AI answers change from run to run. Discovered Labs says no one can honestly promise citations, and LoudFace tells buyers to walk away from anyone who does.
• Schema or llms.txt sold as the fix. Both are hygiene.
• Undisclosed Reddit or community posting. It breaks community rules and puts your brand at risk.
• Traffic-only reports. You need demos and pipeline, not just visits.
• One vanity score. Ask for share of voice by topic and engine, with the questions behind it.
• No baseline. If they cannot show where you start, they cannot show what changed.

Where does Noudar AI fit?

We turn AI visibility into pipeline. Noudar AI is an AI-native SEO and AEO agency for B2B SaaS and tech companies, based in Portugal. AI does the research, checks and first drafts. SEO and AEO experts make every call.

We are new and have no client case studies yet. So we run a Design-Partner batch: 3 spots, open until 31 October 2026. It costs EUR 1,500 at the start and EUR 1,500 at day 90. If we miss the goals agreed at kickoff, you get a full refund. A case study with us is part of the deal.

Ask us the 10 questions above. Hold us to the same answers.

Get a free AI visibility audit, results in 12 hours. Or see pricing.