What does an AI SEO agency do?
Most agencies selling AI SEO are running a traditional SEO programme with a GEO slide added. “AI SEO,” “generative engine optimisation (GEO),” “AEO” and “AI search optimisation” get used interchangeably by agencies with very different capabilities. Meanwhile, B2B buyers use AI tools for vendor research, and generative search increasingly shapes shortlists before a sales conversation begins.
Key takeaways
- AI SEO aims for citation in AI-generated answers, alongside the SEO foundation.
- A typical programme combines question mapping, technical groundwork, content, third-party presence and tracking.
- In our experience, early citation gains often appear within 8 to 12 weeks.
- Outside partners perform about a third of digital marketing at surveyed US companies.
We run AI search programmes for B2B SaaS brands, and the pattern repeats: strong Google rankings, few appearances in the AI answers buyers read. In this guide, we’ll explain what an AI SEO agency does, how it differs from traditional SEO, what the first few months involve, and how to judge whether an agency understands the channel.
How AI SEO differs from traditional SEO
Traditional SEO optimises for position in a ranked list of results. AI search optimisation optimises for citation: whether a generative model names, quotes or recommends your brand when it builds an answer. The signals overlap, but AI search adds work classic SEO doesn’t cover: mapping the questions buyers ask AI, and building presence on the third-party sites these tools draw from.
Ranking well doesn’t guarantee citation. Ahrefs’ AI Overview study found that 38% of AI Overview citations come from pages in Google’s top 10. A separate Ahrefs study found only 12% of links cited by ChatGPT, Gemini and Copilot appear in the top 10 for the same prompt, against 28.6% for Perplexity.
A brand can therefore rank well in Google and still be absent from the AI answers in its category. We treat the gap as a source-set problem: getting into the sources a model trusts and can extract from, then structuring what we publish so it can be quoted. Want the full breakdown of how the two disciplines overlap and where they split? Read our GEO vs SEO guide.
Why keyword targeting and link building are not enough
Keyword research and link building still have a place in the SEO foundation, where technical health, topical authority and third-party trust are the same signals generative engines evaluate. Ranking pages still supply a share of AI Overview citations. An agency that delivers keywords and links as the main AI SEO output, though, is mostly running a standard SEO programme under a new name.
Buyers now put conversational, multi-part questions to AI instead of typing a keyword, the shift covered in keywords to conversations. A strong AI search service builds for that, with prompt-level measurement, third-party presence and entity clarity, so AI engines understand who the brand is.
How AI-generated answers change buyer search behaviour
When a buyer asks an AI tool to recommend vendors, it returns one synthesised shortlist, with no ranked list of links to scan. A brand left out of that answer may never be considered at the research stage, whatever its position in traditional results.
In the categories we track, the brands cited in AI vendor comparisons are the ones already talked about across third-party sources. A buyer may also meet a brand in a ChatGPT answer and then confirm it with a Google search, a pattern covered in AEO vs SEO, so AI visibility needs tracking alongside rankings.
What an AI search optimisation service includes
At FirstMotion, AI search optimisation is five things run as one programme:
- Question mapping: the questions buyers ask AI at each stage of their journey
- Technical groundwork: comparison tables, FAQ schema and answer-first structure that help AI engines find and quote you correctly
- Content: expert-led content and briefs built from real buyer questions and sales calls
- Third-party presence: building visibility on the sites and publications AI engines already trust and pull from
- Tracking: citation rate, share of voice and mention position, by engine
Evidence can guide the content. An academic study of generative engines, the GEO research paper, found that adding citations, quotations and statistics raised source visibility by up to 40% on its own benchmark, with results varying by domain. Third-party presence targets the publications and community sites AI engines retrieve from for a brand’s category prompts, as covered in this digital PR guide.
Measurement across AI platforms
Because ChatGPT, Perplexity and Google AI Overviews rarely move together, we track citation rate and share of voice daily, against named competitors and broken down by platform. That covers ChatGPT, Claude, Gemini, Copilot, Perplexity and Google AI Overviews. The data lives inside our ContextualJourney™ platform, next to traditional traffic and conversion reporting.
Why a traditional SEO approach falls short
A traditional SEO programme optimises for Google rankings, backlink profiles and keyword coverage. Those still matter, but they don’t show where a brand appears in AI answers. In our view, the measurement gap is the clearest difference. A team that can’t say which prompts cite you, or which competitors win the answers you miss, is working without the data this channel runs on.
An agency that moves straight to content briefs without that data is probably running an SEO programme with new terminology.
Technical groundwork for AI search
Technical work in AI search covers structured data, entity clarity and internal linking, so AI systems can find a brand’s pages, understand who it is and extract answers cleanly. Schema is groundwork. In a controlled Ahrefs study of 1,885 already heavily cited pages, adding it produced no major uplift in AI citations on any platform. Our entity authority guide covers the entity signals AI systems use.
A UK discretionary wealth manager shows what this looks like in practice. AI engines were confusing the firm with an unrelated company, third-party sources were calling it a robo-advisor, and an early audit found it accounted for almost none of its category’s tracked AI citations. We separated the entity, reviewed its schema, and worked with the third-party sources behind the inaccurate description.
We also built answer-first content around the questions prospective clients ask, with comparison tables, named reviewers and FAQ schema. The firm reached #1 share of voice among its tracked competitors at 51%, and the share of chats citing its site rose from 1.7% to 9.8%. Read the wealth management case study.
What the first few months look like at FirstMotion
There’s no industry standard for how an AI search engagement begins, so this is how we run it. It starts with a free 20-minute call and an agreed scope, then moves into a standalone strategy phase that usually takes four to five weeks. Scope sets both the timeline and the fee: the number of countries a brand operates in and the number of existing pages to review each move.
The phase covers audits (AI search and visibility, technical SEO and AEO), prompt mapping across 150+ buyer prompts, customer and ICP insights, competitor analysis, content-gap, backlink and off-site authority analysis, keyword research, and entity and brand consolidation. It ends in a full SEO and AI search roadmap that the client keeps whether or not they continue with us, to run in-house or with us in whatever arrangement suits them.
Ongoing Growth work then follows the roadmap: content published, technical fixes implemented and visibility tracked by platform, as set out below. You can see how we work in full.
What ongoing Growth work includes
Question mapping happens in the strategy phase. Ongoing Growth work at FirstMotion then covers four areas:
- Content: every month, our team writes, edits and publishes answer-first content, rewrites existing pages based on content-gap findings, and ships new landing pages
- Technical fixes: technical SEO and AEO fixes implemented directly, including schema, metadata and internal linking
- Off-site: opportunity review and priority targets drawn from competitor citation and backlink gaps, with placement as an add-on (our guide to how brand mentions drive citations explains the thinking)
- Tracking and reporting: a live dashboard benchmarked against competitors, regular review meetings, a dedicated Slack channel, and reporting on leads, pipeline and revenue alongside citation data
How long AI search results take
Timelines vary by platform, category and how competitive the prompts are. Some changes can show up quickly, while broader gains take longer to build. In our experience, early citation gains often appear within 8 to 12 weeks of a strategy phase finishing, and share of voice tends to compound over three to six months. That growth depends on reviews and third-party coverage, which sit outside a brand’s control. Earning that presence is the focus of our Digital PR and Reddit marketing work.
Platforms also change how they work constantly, so we monitor updates every week and look for competitor overtakes early. This guide to measuring AI share of voice explains the method.
We don’t guarantee specific rankings or citations, so treat any timeline as an estimate. For a worked example, see how a legal technology client held the #1 share of voice position for 12 months in this case study.
Should you hire an AI SEO agency or build in-house?
In The CMO Survey 2026, sponsored by Duke University’s Fuqua School of Business, Deloitte and the American Marketing Association, 308 marketing leaders at US for-profit companies reported that external agencies, partners and services perform 33.6% of their digital marketing activity on average, up from 31.6% in 2022.
| Measure | Finding (latest survey) |
|---|---|
| Approach to developing capabilities | 59.5% of emphasis on building, through training and hiring |
| Partnering | 38.5% of emphasis on partnering, including 15.5% on agencies |
| Top gap named | Existing capabilities lacking the people, time or budget to function (22.3% of responses) |
| Training and development budgets | 3.8% of marketing budgets on average, down from a pre-pandemic high of 5.8% |
| AI search | 41.5% of companies use GEO; 11.6% of responses name managing GenAI, GEO and bots as a lacking capability |
Building is the dominant approach, yet the single most cited gap is resourcing (22.3% of responses), and training and development budgets have fallen from their pre-pandemic high.
Two limits apply. The sample is US-only and covers marketing in general, not AI search, and the report doesn’t compare cost or results between internal and agency delivery. The routes can also be combined, since the roadmap from our strategy phase can be run in-house.
Who we built this for
We specialise in Series A and B SaaS and software brands, where the buyer journey and budget realities are specific to that stage. We see the most impact where buyers research vendors before engaging sales.
Across the brands we work with, the pattern at programme start is high Google visibility and near-zero AI citation for category prompts. The GEO agency article explains how to evaluate if an agency is genuinely building for that gap.
AI search work is likely to matter less for brands that sell mainly through direct relationships, enterprise procurement or local search channels. It’s also premature for pre-product brands, because third-party presence depends on a product in market with real users who generate reviews and coverage.
Start with the data, not the deck
Every FirstMotion engagement starts with a fixed-scope strategy phase that maps where your brand appears across the prompts your buyers use, which platforms return it, and which competitors win the answers you’re missing. A free 20-minute call comes first, with no obligation, and we’ll tell you honestly whether we can help. Book a free 20-minute call