TL;DR
- Of the more than 25 million links AI engines cited in Muck Rack's May 2026 study, 84% came from earned media. The PR agency you hire now shapes a large part of what ChatGPT, Claude and Gemini say about you.
- Big PR firms and boutiques both sell AI visibility work in 2026. The differences that matter are who does the daily work on your account, what the agency counts as success, and whether PR, your site, community and content run from one plan.
- Our position, and it is a position rather than a finding: for a Series A or B B2B startup, a senior-led boutique that runs all four as one plan is the stronger bet. A big firm earns its fee when you need many markets and a lot of people at once.
- Take the four tests at the end into every pitch meeting. Each has a pass line. Run them on us too.
What is the difference between a big PR firm and a boutique?
Most of the big PR firms belong to a holding company, a group that owns many agency brands across advertising, media buying, health, branding and public relations. Inside the industry they're called networks. To a founder they're just the big firms, and that's the word we'll use. The biggest group got bigger last winter: on 26 November 2025, Omnicom completed its acquisition of Interpublic, putting both groups' agency brands under one roof, public relations among them. Not every large PR firm belongs to a holding company; some of the biggest are independent. For a startup, what matters is the size of the firm and how it builds the account team. Ownership is useful mainly because it predicts both.
A boutique is a small, usually independent firm with a short client list, where the senior people who sell the work also do it. Some boutiques are generalists. Others, like us, are built around one problem, which in our case is how B2B brands get found and cited in AI answers.
Two terms before we go further. Generative engine optimization (GEO) is the work of making your brand's presence findable and citable by AI systems like ChatGPT, Perplexity, Gemini and Google's AI Overviews. Answer engine optimization (AEO) is the narrower craft of structuring a page so an engine can lift a direct answer from it. Agencies of every size now use both words in their pitches, which is exactly why the words alone won't help you choose.
The two models at a glance
Every row describes a tendency, and any single agency can break it.
Why does the choice matter more now that buyers ask AI?
Your prospects now ask an AI assistant which vendors to shortlist before they ever fill in a form. What comes back is assembled from sources the engine trusts, and those sources are mostly other people talking about you.
The Muck Rack figure holds up over time. Its Generative Pulse team analysed links from ChatGPT, Claude and Gemini responses across 17 industries, and across three editions of the study going back to July 2025, earned media ranged from 82% to 89% of citations. Journalism alone was 27% in the May 2026 edition. Paid and advertorial content was 0.3%. You can't buy your way into those answers, which puts PR close to the centre of how AI describes your company.
The starting point for most B2B brands is low. Walker Sands' study of 45 million keywords across 828 companies found that 3.0% of relevant AI Overviews include the median enterprise B2B brand. That figure covers Google's AI Overviews only, and it describes enterprise brands, which are likely to have more coverage than a Series A company.
The answers also move. Researchers measuring AI search visibility found that answers vary across runs, prompts and time, so a single check is unreliable and visibility is better read as a distribution. That has a practical consequence for hiring: an agency that shows you one screenshot of ChatGPT mentioning a client has shown you very little.
So the hire comes down to whether the team can pick the sources engines actually cite, keep your site readable when an engine checks the details, and measure the result honestly.
Who actually does the work on your account?
This is the question that separates the two most clearly, and the one pitch decks answer least clearly.
How a big firm staffs an account
Large agencies are built to run many accounts at once, so they staff in layers. Senior leaders set strategy and show up at the pitch and the quarterly review. Account managers and executives handle the drafting, the media lists and the daily email. That structure is sensible when the work is broad: a public company launching in twelve markets needs bodies. It is less sensible when the work is narrow and changes weekly, because every change to the story has to travel up and back down the layers.
None of that makes a big firm's people less capable. The issue is time. In our experience, the senior partner in your pitch is carrying other clients too, and most of the hours your retainer buys belong to other people. You won't find that ratio in the deck, so ask for it.
How a boutique staffs an account
A boutique has fewer layers because it has fewer people. The person who scopes your strategy is usually the person writing the pitches and talking to editors. When your positioning shifts, the person who changes the pitch is the person who set the strategy.
Althea is built around its founder, Kara Silverman, who spent years in senior roles at global agencies and leading in-house teams before starting the firm. Clients on our ongoing programs get direct strategic access over Slack rather than a monthly check-in with someone they met once.
The trade-off is real and worth saying plainly. A small firm has key-person risk, finite bandwidth and fewer specialists on the bench. If your plan needs a crisis war room staffed around the clock or native speakers in six languages, a boutique will have to partner for it.
How does each model decide where to pitch?
A traditional program starts from a media list: the outlets that cover your category, weighted by reach and the relationships the team already has. That still matters. Journalism is more than a quarter of what AI engines cite, and relationships get stories written.
A program built for AI visibility adds a step in front of the media list. It starts from the questions your buyers actually type into ChatGPT or Perplexity, then looks at which outlets, community threads and reference pages the engines cite when they answer. Pitch targets follow from that evidence. A placement in a source that keeps showing up for your buyers' prompts can do more for you than several clips in outlets that never appear, and you only know which is which if someone has looked.
This is where we'd push back on a lazy version of our own argument. Large agencies can do this research, and some are helping define how it should be measured. The AMEC GEO Principles, published by the international measurement association for communications, were shaped with contributors from Ketchum and FleishmanHillard among others. The test is whether the team on your account will run that research every month and change the pitch list when the evidence changes.
For a deeper look at how this share of AI answers gets measured against competitors, see Part 3 of the Althea Score on share of voice.
What should an agency count as success?
The PR industry settled part of this argument years ago. The Barcelona Principles 3.0, the measurement standard AMEC updated in 2020, say measurement should identify outputs, outcomes and potential impact, and they state flatly that "AVEs are not the value of communication." Advertising value equivalents, the practice of pricing a clip as if it were an ad, still turn up in agency reports. If one appears in a pitch deck, you've learned something about the agency.
For AI visibility, the AMEC GEO Principles split measurement into three connected areas: upstream reputation (the earned, shared and owned signals that shape how you're understood), search and content readiness (whether your digital presence can be read by search engines and AI systems), and downstream AI outputs (how you appear in answers, including accuracy, framing, citations and omissions). The same page warns against vanity metrics, simplistic rankings and opaque scores. That is a useful checklist for any agency report, including ours.
We measure AI citation rates, SEO movement and a set of buyer prompts against a baseline, and we send a monthly visibility report showing progress against it. Because answers vary from run to run, we re-run a tracked set of prompts every month rather than relying on a single snapshot.
We don't claim pipeline attribution we can't see, and you should be wary of any agency that does. Connecting AI visibility to revenue takes your data. OpenAI's publisher guidance says referral links from ChatGPT carry utm_source=chatgpt.com, so those visits can be segmented in your analytics. Google Analytics lets you mark sign-ups or demo requests as key events, and your CRM can ask new leads where they first heard of you. An honest agency tells you which of those it can see and which it can't.
Can one agency run PR, SEO, community and content together?
It can, and for AI visibility we think it should, because each discipline produces half a signal on its own. A placement with no clear, crawlable page behind it gives an engine a claim it can't confirm. A thorough page with no independent coverage gives it a claim nobody has vouched for.
The site half is less exotic than some pitches suggest. Google says there are no additional requirements to appear in AI Overviews or AI Mode: the page has to be indexed and eligible to show a snippet. OpenAI's guidance says to allow its search crawler, OAI-SearchBot, if you want your pages to appear in ChatGPT's search summaries. That work is unglamorous, and it is often nobody's job when PR, SEO and content sit with three different vendors.
Community is the part we most often see left out of integrated programs. Reddit threads, developer forums and review sites carry the human proof engines lean on when they compare vendors. We wrote about that layer in Part 4 of the Althea Score on community consensus.
If you'd rather split the work across firms, you can. Insist on one shared research document, one prompt list and one report, or the pieces will be optimised against different goals. Our buyer's guide to choosing a PR and SEO agency for AI search covers how to keep separate vendors on one plan.
Which startups does each model suit?
When a big firm is the right call
Pick the big firm when the job is breadth. That means a launch across many countries at once, a crisis that needs people on shift, a regulated business with stakeholders in several markets, or an analyst-relations program spanning regions. It also makes sense when you already buy advertising and media through the same group and want one relationship. In those cases the layers are the product, and paying for them is reasonable.
When a boutique is the right call
Pick the boutique when the job is depth. A Series A or B B2B company in developer tools, fintech, adtech, martech or SaaS usually has one core market, a founder whose voice is the brand, and a positioning that is still moving. The work is getting a small number of trusted sources to describe you accurately and making sure your own site and community presence back that up. That rewards senior judgment and fast iteration over headcount.
If you're a Series A or B founder reading this, you're probably in the second group and being pitched by both. The deciding question is simpler than the decks make it: which team will actually be doing the work in month four?
How do you test any agency pitch for AI visibility?
Ask these four questions in the meeting, and write down the answers. An agency that passes three of the four with specifics, not promises, is worth a second meeting. If you're comparing several agencies side by side, the 100-point scorecard in that buyer's guide weights the same questions in more detail.
1. Who will do the work each week?
Ask for names, roles and the share of each person's time on your account, and ask whether those names can go into the contract. Pass: they agree to put the names from the pitch into the statement of work and tell you who covers when those people are away.
2. How will you measure AI visibility?
Ask which prompts they'll track, on which engines, how often they re-run them, and how they log cited sources separately from brand mentions. Pass: they show you a sample prompt list, a baseline method and a re-test interval, and they say the results are a range rather than a guarantee.
3. How will PR connect to our site, community and content?
Ask who owns technical SEO and structured data, who writes the pages that coverage points to, and who watches the communities your buyers use. Pass: one research document, one prompt list and one report cover all of it, whether one firm does the work or several.
4. What will you count as success, and what can't you attribute?
Ask what the monthly report contains and which numbers come from your systems rather than theirs. Pass: they name the metrics, the data source for each, and what they can't attribute. An answer built on reach or advertising value equivalents is a fail.
No one can promise you a citation, just as Google says no one can guarantee a #1 ranking. An agency that promises otherwise has failed the test before you ask a question.
Where does Althea fit?
Althea Labs is a brand presence and AI visibility agency that builds targeted visibility systems for B2B startups through integrated PR, SEO, and content programs.
Presence beats fame.
We run that work through PR Plus, which joins four components: Authority (traditional PR), Infrastructure (SEO and schema), Human Proof (community) and The Bridge (content). The work follows five steps in order: an Audit of your SEO, PR footprint and AI visibility; ICP and Messaging; Visibility Tracking, where we baseline 300 to 500 prompts pulled from real buyer conversations and your own customer interactions; Strategy, a 90-day editorial calendar across content, PR and community; and Execution, which ends each month in a visibility report against the baseline.
You see a baseline as soon as tracking is set up. Early wins typically show up in 60 to 90 days, and the gains compound over months as each placement and page makes the next one easier to earn.
We'd rather be judged on the four tests above than on our pitch. Our services page describes what each engagement includes.
FAQs
What should a retainer buy at Series A or B? Ask each agency what the fee pays for: whose hours, how many pieces a month, and which report. Compare those, not the headline number.
Can a traditional PR firm do AI visibility work? Yes, and some large firms helped shape the industry's measurement principles for it. What varies is whether the team on your account will do the prompt research, fix the site, work the communities and report against a baseline. Use the four tests above.
Can a boutique keep up if we scale fast? A good one will tell you where its limits are. Ask how staffing and deliverables change if you add markets or double your content plan, and which partners it brings in for work it doesn't do in-house.
What counts as an AI citation? A citation is when an AI answer links to or names a source, such as an article that covered you or a page on your site. A brand mention without a source is worth tracking separately, because engines differ in how they show sources.
How long before AI visibility improves? There's no fixed timeline, because it depends on your starting point, your category and how often engines refresh what they cite. We plan for early wins in 60 to 90 days, with gains compounding over months.
Do we need PR and SEO from the same firm? No, but they need one plan. Shared research, one prompt list and one report matter more than how many logos are on the contract.
Score us too
Before you sign with anyone, find out where you stand. Our audit measures your SEO, PR footprint and AI visibility against your category and gives you a baseline to hold any agency to, including us. Start with an audit.
CONTINUE READING
.png)
