A retainer buys hours. A lab is a standing instrument.
An AEO agency sells you hours: audits, content, schema rollouts, monthly reports. Mid-market retainers run $2,000 to $10,000 a month, enterprise $10,000 to $25,000+, and the technical recommendations land in your engineering backlog, where they wait[01][02]. Bull AI Labs is a lab, not a retainer: a standing instrument that probes your site as the AI agents themselves, ships the engineering fix, and re-probes until it verifies. A retainer buys effort. The loop buys outcomes.
By Sai Narendran · Updated 24 Aug 2026 · Sources dated at the foot
What a retainer actually buys.
- Priority query list and prompt research
- Citation and competitor matrix
- Entity pages and FAQ hubs
- JSON-LD rollout across templates
- Listicles and third-party placements
- Reddit and Quora seeding
- Weekly rank tracking and a monthly report
The 2026 pricing guides put mid-market retainers at $2,000 to $10,000 a month and enterprise engagements at $10,000 to $25,000+[01][02]. To be fair: good agencies do real work, especially off-site. Reddit presence, PR, and third-party placements move citations, and we do not sell any of it.
The question is not whether the work is real. It is whether the work touches the layer that is actually broken.
Three problems a retainer cannot outwork.
01 · No persistent instrument.
Agency measurement is periodic manual checks, and the industry itself concedes there is no universal standard for quantifying AI visibility[01]. A monthly screenshot is not an instrument. Our probes run on cadence and on demand: 10 canonical crawler user agents, every score anchored to a persisted run[05].
02 · The backlog handoff.
Technical fixes get recommended, not shipped. The recommendation becomes a ticket, the ticket joins the backlog, and nobody re-fetches the page as ChatGPT-User afterward to confirm the shipped change altered what the agent receives. A fix that is never re-probed is an assertion, not a result.
03 · The incentive.
A retainer bills recurring content. Fixing retrieval once shrinks the invoice, so the invoice prefers content. If your site 403s ChatGPT-User - and in 802 persisted probe runs, it did 96 times against a browser agent's 7[05] - ten listicles will not help.
If your invoice is mostly schema, you bought the wrong layer.
Many agencies still sell JSON-LD rollouts and llms.txt as citation levers. The largest controlled study says otherwise: Ahrefs, May 2026, 1,885 treated pages against 4,000 matched controls. AI Mode +2.4%, ChatGPT +2.2%, AI Overviews -4.6%. The first two are statistically indistinguishable from zero[03]. llms.txt accounts for roughly 0.1% of AI bot requests in our production observation[04].
So we capped schema at MEDIUM severity and 1 impact point, hard-capped in code, and we keep llms.txt as a check with zero citation weight. If your AEO invoice is mostly schema deployment, you bought the wrong layer.
What a standing instrument buys instead.
Probe, Persist, Generate, Ship, Re-probe, Verified. Two clocks govern the outcome, and we label them clearly: retrieval eligibility moves in 24 hours to 7 days; answer share takes 3 weeks to a quarter. No retainer changes those clocks. Shipped and verified fixes are what move them.
These figures have decayed since the engagement ended, which we publish, because the decay is the argument for a standing instrument rather than a one-off report[06].
Before the next retainer renewal, watch the fetch.
Paste a URL and the two-fetch diff runs live: the same page fetched as a browser and as an AI agent, in the same second.
That was a demo domain. Run it on yours. Free, no signup, full diagnosis.
Retainer vs lab.
| Dimension | Typical agency retainer | Bull AI Labs |
|---|---|---|
| What you buy | Hours: audits, content production, schema rollouts, monthly reporting. | A standing instrument: live probes, shipped fixes, and re-verification on cadence. |
| Measurement cadence | Periodic manual checks, typically monthly. The industry concedes there is no universal standard for quantifying AI visibility. | Persisted probe runs on cadence and on demand: 10 crawler user agents, raw versus rendered. |
| Technical fixes | Recommended, then handed to your engineering backlog, where they wait. | Shipped as sprint tasks: render parity, crawl access, extraction structure. |
| Verification | Asserted in the next report. Nobody re-fetches as the agent to confirm the change landed. | A fix closes only when the re-probe confirms what the agent receives actually changed. |
| Content | The core of most retainers: listicles, entity pages, Reddit and PR seeding. Real work, especially off-site. | The second lever, generated from measured gaps. The fetch is the first. |
| Incentive structure | Recurring billing on recurring content. Fixing retrieval once shrinks the invoice. | Proof is free. You pay for the remedy, and the re-probe is the receipt. |
| First proof | Weeks to an audit deck; months to a measurable delta. | Minutes. The free audit shows the two-fetch diff on your own pages before you pay anything. |
| Typical cost | $2,000 to $10,000/mo mid-market; $10,000 to $25,000+/mo enterprise, per 2026 pricing guides. | Free diagnosis. Paid plans start when you want fixes shipped and continuously re-verified. |
Which one is yours.
Pick an agency if
- You need off-site citation building at scale: Reddit, PR, listicles. That is real work, and we do not sell it.
- You have zero content capacity and need a done-for-you content program.
- You want a team to own your AI visibility reporting end to end.
Pick Bull AI Labs if
- The technical layer has never been probed: no one has fetched your pages as ChatGPT-User.
- Agency recommendations keep dying in your engineering backlog.
- You are paying monthly for reports that restate the same findings.
- You want the fix verified, not asserted.
The two compose. Several customers run us for the fetch layer and an agency for off-site citation building. Once the pages verify clean, the agency's content has something to land on.
Questions buyers actually ask.
How much does an AEO agency cost in 2026?
Per 2026 pricing guides, mid-market AEO retainers run $2,000 to $10,000 a month, and enterprise engagements run $10,000 to $25,000 or more. That typically buys audits, content production, schema rollouts, and monthly reporting. The Bull AI Labs diagnosis is free; you pay only for shipped, verified fixes.
What does an AEO agency actually do?
The typical deliverable list: priority query research, a citation and competitor matrix, entity pages and FAQ hubs, JSON-LD rollouts, listicles and third-party placements, Reddit and Quora seeding, and weekly rank tracking with a monthly report. Good agencies do real work, especially off-site. The technical fetch layer is usually recommended, not shipped.
Should I hire an AEO agency or use a platform?
It depends on the bottleneck. If your gap is off-site presence and content capacity, an agency is the right buy. If your gap is that AI agents cannot fetch or read your pages, no amount of content fixes that, and a probe will show it in minutes. Many teams run both: an agency for off-site, Bull AI Labs for the fetch layer.
Do AEO agencies fix technical issues?
Usually they identify them. Technical recommendations land in your engineering backlog as tickets, and nobody re-fetches the page as an AI agent afterward to confirm the fix changed what the agent receives. Bull AI Labs ships the fix and closes it only when the re-probe verifies the change.
Is AEO just SEO with a new label?
No. SEO assumes a crawler that renders JavaScript and indexes what it renders. Agent-class AI crawlers fetch raw HTML and do not execute JavaScript, so a page can be perfectly indexed by Google and unreadable to ChatGPT-User. The two-fetch diff makes that gap visible in one test.
How fast can AEO show results?
Two clocks govern the outcome. Retrieval eligibility, whether agents can fetch and read your pages at all, moves in 24 hours to 7 days once fixes ship. Answer share, how often you are cited, takes 3 weeks to a quarter. Be suspicious of anyone who sells the second clock on the first clock’s timeline.
Before you renew the retainer, see what the agent can read.
Sources · numbered and dated
- 01 ·Digital Elevator, 2026 AEO agency pricing guide. Mid-market retainers $2,000-$10,000/mo, enterprise $10,000-$25,000+/mo. Concedes there is no universal standard for quantifying AI visibility.
- 02 ·Stackmatix, 2026 AEO pricing bands. Consistent retainer ranges across mid-market and enterprise tiers.
- 03 ·Ahrefs, 11 May 2026. Schema study, n=1,885 treated pages vs 4,000 matched controls: AI Mode +2.4%, ChatGPT +2.2%, AI Overviews -4.6%.
- 04 ·Bull AI Labs production observation, 2026. llms.txt accounts for roughly 0.1% of AI bot requests; retained as a check, zero citation weight.
- 05 ·Bull AI Labs production corpus. 802 persisted probe runs, 17 Aug 2026. ChatGPT-User received a 403 on 96 runs; a browser user agent on 7.
- 06 ·Freeletics 12-week sprint engagement. Figures shown as measured at engagement end; they have decayed since, and the decay is noted wherever the figures appear.