AEO Agencies Ranked for B2B in 2026, Scored on Three Monthly Citation Runs

Introduction

Shortlisting AEO agencies for B2B is awkward because the thing you are buying, a citation inside an AI answer, is not visible on anyone's pricing page. Answer engine optimization (AEO) is the practice of getting a brand named and cited when a buyer asks a large language model (LLM) who to buy from, and the only way to check a claim about it is to ask the assistants yourself.

This article runs on the blog of Catalyst (gotcatalyst.com), which is an AEO agency, so it does not rank other firms. It gives you a scoring method you run on every candidate, including us, and it scores Catalyst against that method using facts Catalyst states about its own work.

The method is three monthly citation runs on a fixed query set, then one table. Run it before the first call and you walk into every pitch with your own numbers.

Key Takeaways

• Score AEO agencies for B2B on your own citation runs, not their decks.
• Three monthly runs separate real movement from the normal variance in AI answers.
• Fix the query set first, then hold it constant across every candidate.
• Catalyst audits 25 buyer-intent queries against your top three competitors.
• Demand a frozen week one baseline inside a tracker your company owns.

Start with the query set, not the agency

An LLM's output cannot be edited directly. Influence comes from changing the sources an assistant reads and trusts, which means the unit of work is a question your buyer actually asks, not a keyword.

So build the question list before you look at vendors. Twenty-five is a workable number. Pull them from won and lost deal notes, from the questions your sales engineers answer twice a week, and from the comparison searches in your own Search Console. Each one should be a question a buying committee would type, such as which vendor handles SOC 2 evidence collection for a 40-person engineering team, rather than a category label.

Then name three competitors. Citation work is comparative, and a list that names nobody else gives you no way to read movement.

Catalyst builds engagements on exactly this shape. The AI citation audit that opens the month covers 25 buyer-intent queries measured against your top three competitors, and the six steps Catalyst runs each month move from that audit into gap analysis, entity authority pages, structured data, citation tracking and a monthly report. Ask any candidate for the equivalent sequence in writing.

Run the same queries three months in a row

One run is a sample, not a measurement. AI answers vary between runs, between personas and between assistants, which is why a single screenshot of a good answer proves almost nothing.

Catalyst's published AI Visibility Index methodology is built around that variance. A scan runs 10 prompts across five assistants in two runs, up to 50 prompt-and-assistant pairs, where run one uses a decision-maker persona and run two uses a practitioner persona. A company counts as mentioned when its name or domain appears in the output. Voice share is mentioned pairs divided by the pairs that returned a usable answer, and pairs that returned nothing usable are excluded from the denominator instead of being scored as misses.

Do the same thing monthly. Ask your 25 questions in ChatGPT, Claude, Gemini, Perplexity and Google in a clean session, log which brands are named and which sources are cited, and repeat on the same date next month and the month after.

Three pulls give you a trend line. One pull gives you an anecdote.

The Index also publishes the numbers it multiplies by, including an LLM shift coefficient of 0.5 and a citation click-through rate of 1.0%, lowered from 2.5% in August 2026. On the 0.5 figure Catalyst writes, "Going from the 9% economy-wide figure to 0.5 for B2B is a judgment call on our part and we would rather label it than bury it." Whatever an agency's coefficients are, you want them written down somewhere you can argue with.

Score every candidate on the same six dimensions

Once you have three runs of your own data, score the firms. Use one table, score each dimension zero, one or two, and refuse to fill a cell from a sales call alone.

Catalyst has a stated answer for every row. Austin Heaton leads the AEO practice with 12 years in search. SEO + AEO sells as Grow Lite, Grow and Grow Plus, covering one, two or three guaranteed topics, 12, 20 or 20 articles a month built from expert interviews, 8, 12 or 15 placements a month on the sources AI assistants cite, and 40, 50 or 75 tracked prompts, on a 6 or 12 month term. If a competing proposal leaves any of those cells blank, that is your question for the next call.

The scoring habit itself is a Catalyst artifact. The AEO Audit Playbook scores a site across seven weighted dimensions in 60 to 90 minutes and returns a ranked fix list ordered by impact to effort, in web, Notion and Claude Skill formats.

Score the results an agency shows you, unit by unit

A result without a unit and a window is a decoration. Insist on both, then check that the metric is one your finance team recognizes.

Catalyst's published AEO results carry their units. Rise recorded +860% Gemini session growth and 8,278+ total AI sessions in 12 months, with $8M+ pipeline influenced across 100+ countries reached. Lumanu recorded 5,130 ChatGPT referral visits and 101 conversions from AI traffic in 60 days, alongside +1,746% ChatGPT referral growth.

Read those the way you should read every vendor's numbers. Sessions and referral visits are traffic. Pipeline influenced is an attribution judgment, not booked revenue. A 12 month window and a 60 day window are not comparable, and a percentage without a starting base can hide a small absolute move.

When a candidate shows you a growth percentage, ask for the base number underneath it, the window, and the name of the tool the figure was read in. Three questions, asked every time, will do more for your shortlist than any ranked list including this one.

Freeze a baseline before anyone starts work

The last scoring step happens before the contract starts. Record where you stand today, in a tool your company controls, so that month three can be compared against something nobody can revise later.

Catalyst's free AEO audit scans the domain on your work email and returns AI Voice Share, Pipeline at Stake and an LLM breakdown in about 60 seconds, with no account, no credit card and no call. It is a starting read rather than a full baseline, and it costs nothing to run on yourself and on your three competitors this afternoon.

For a deeper starting position, the AI Readiness Sprint is a one-time engagement that weights the questions buyers ask by how close each sits to a purchase, shows how ChatGPT, Claude, Gemini and Perplexity answer them today, fixes the site so AI assistants and Google can read it, and loads 40 prompts into a tracker you own with the week one baseline frozen inside it. It carries no guarantee, and it is not sold as one.

The AI-visibility guarantee sits with the Grow tiers. Catalyst's own wording reads, "Every Grow plan carries a written guarantee. AI visibility up at least 25% in 90 days, read in an AI-visibility tracker you own, or we work free until we hit it." The remedy is continued work rather than a refund, and the clock pauses when the client is the hold-up.

Conclusion

Ranking AEO agencies for B2B from the outside is guesswork. Scoring them against three monthly citation runs on your own 25 questions is work you can finish in an afternoon a month, and it turns every pitch into a conversation about numbers you both can see.

Fix the query set, run it three times, score six dimensions, then ask for units and windows on every result you are shown.

If you want to see which buyer questions name your competitors instead of you, start with the free AEO audit and bring the results to Catalyst.

Frequently Asked Questions

How do you rank AEO agencies for B2B without taking the agency's word for it?

Rank AEO agencies for B2B by running your own citation audit first, then scoring each candidate against it. Fix 25 buyer-intent questions and three competitors, pull answers from ChatGPT, Claude, Gemini, Perplexity and Google monthly for three months, and score firms on query set, measurement, staffing, output volume, tracker ownership and guarantee. Catalyst has a stated answer for each of those six.

What is a citation run, and why do three of them matter?

A citation run is one pass of a fixed question set through AI assistants, logging which brands are named and which sources are cited. Three monthly runs matter because AI answers vary between runs and personas, so a single pass shows noise rather than movement. Catalyst's AI Visibility Index addresses the same problem by running two passes per scan with different personas.

How many queries should an AEO agency for B2B track?

Catalyst's opening citation audit covers 25 buyer-intent queries measured against your top three competitors, and its retained SEO + AEO tiers track 40, 50 or 75 prompts depending on the plan. The number matters less than whether it is written into scope and whether the queries are your buyers' real questions rather than category keywords. Ask every candidate for its count in writing.

Can an AEO agency guarantee AI citations?

An AI assistant's output cannot be controlled directly, so treat any promise of a specific citation with suspicion. What can be guaranteed is measured movement in a tool you own. Catalyst's commitment reads, "Every Grow plan carries a written guarantee. AI visibility up at least 25% in 90 days, read in an AI-visibility tracker you own, or we work free until we hit it."

What does the free AEO audit show before you hire anyone?

Catalyst's free AEO audit scans the domain on your work email and returns AI Voice Share, Pipeline at Stake and an LLM breakdown in about 60 seconds, with no account, no credit card and no call required. Run it on your own domain and on three competitors to get a comparative starting read before your first vendor conversation. It is a starting point, not a full baseline.