In-house vs agency — for AI visibility work

Why AI visibility work behaves differently than traditional SEO

Search visibility used to mean one thing: rank on page one of Google and the clicks would follow. That link is breaking. In the United States, Google searches ended without a click 68.01% of the time in the first four months of 2026, according to SparkTorós clickstream analysis of Similarweb data — meaning a majority of searches now resolve inside the search results page itself, often inside an AI-generated summary, rather than sending a visitor to any website. Separately, Ahrefs analyzed 300,000 keywords and found that when a Google AI Overview appears above a page's top-ranking result, that page's click-through rate falls by roughly 34.5% compared to the same position without an AI Overview present (Ahrefs, 2025).

That shift changes what "visibility" work actually requires. Traditional SEO optimizes a page to rank in a list a human scans. AEO (answer engine optimization) and GEO (generative engine optimization) optimize content, structured data and brand entities to be selected, synthesized and cited by an AI model that is answering on your behalf — inside ChatGPT, Google's AI Overviews, Perplexity, or Copilot. The skills involved overlap with SEO but are not identical: they lean more on structured data and entity clarity, on writing that AI systems can lift and cite cleanly, and on monitoring tools that did not exist three years ago.

This is the backdrop against which the in-house-versus-agency decision should be made. It is not simply "who can rank a page" — it is who can build and maintain the newer, more specialized muscle of AI-citation visibility, on a timeline that matches how fast the underlying platforms are moving.

When in-house makes more sense

In-house ownership of AI-visibility work makes the most sense when a company already has functioning technical SEO and content operations to extend, rather than build from zero. Content Marketing Institute 2025 B2B research found that 76% of B2B marketing organizations already maintain dedicated in-house content resources, though more than half of those teams (54%) are small — just two to five people. For a team already in that position, AEO/GEO work is often an extension of what they do, not a new department: the same people who understand the product, the buyer language and the existing content library can absorb AI-visibility practices — structured data, entity consistency, answer-first content formatting — into their existing workflow.

In-house also wins on institutional memory and direct access. A staff SEO or content lead sits in the same channel as product, sales and engineering; they can get a schema change deployed, confirm a claim with a subject-matter expert, or adjust messaging the same week a product changes. As the Forbes Communications Council piece on this decision notes, a company that already employs "a marketing team with copywriters, web developers and other digital professionals" is often better served adding a dedicated specialist to that existing team than routing the work externally.

In-house makes the least sense when a company is starting from nothing — no technical SEO capability, no content operation, no data/analytics muscle — because building all three from scratch, while also learning a newer discipline like GEO, is a multi-quarter undertaking most businesses cannot afford to wait out.

When an agency makes more sense

An agency makes more sense when the gap isn't headcount, it's specialization and speed. AI-visibility work draws on a narrower, newer skill set than general marketing — reading how large language models retrieve and cite content, structuring entity and schema markup so AI crawlers parse it correctly, and running the AI-citation tracking tools that show whether a brand is actually appearing inside ChatGPT or AI Overview answers. Building that expertise from a standing start, inside a generalist marketing team, takes time most companies don't have while competitors are already being cited.

Gartner's research on marketing spend gives a useful signal here: even as CMOs cut budgets broadly, agency relationships aren't disappearing — they're being consolidated toward the ones delivering specialized value. In Gartner 2025 CMO Spend Survey, 39% of CMOs said they planned to cut agency budgets, largely by eliminating underperforming or redundant relationships rather than agency use altogether, while agency spend still accounted for roughly 20.7% of total marketing budgets. In other words, CMOs are getting pickier about which agencies earn a seat at the table — favoring those with a specific, hard-to-replicate skill, which is exactly the position a dedicated AEO/GEO agency occupies relative to a generalist in-house team.

An agency also makes sense when AI-visibility work needs to move across an entire fragmented landscape at once — ChatGPT, Google AI Overviews, Perplexity, Copilot — each with different retrieval behavior. Search Engine Land GEO coverage frames this as fundamentally a governance and consistency problem across the enterprise, not a single content tactic, which is easier for a team that has already solved it for other clients than for one solving it for the first time.

What each option actually costs for AEO/GEO work

Cost comparisons between in-house and agency work are often oversimplified into "salary vs. retainer," but the fuller picture matters more for a discipline as new as AEO/GEO, where tooling costs are still shifting.

An in-house hire's true cost is rarely just base salary. It includes benefits, payroll taxes, training time, the software and AI-visibility tracking tools that specialist needs, and the ramp-up period before they're fully productive — the Forbes Communications Council piece notes that building an internal team means absorbing "wages, benefits, taxes, payroll, training, paid time off, and more," costs an agency retainer effectively pools across many clients. An agency retainer folds specialist salaries, tooling licenses and cross-client benchmarking into one line item, but it's an ongoing cost with no equity built in the business — stop paying and the active work generally stops.

The direction of travel also matters. Gartner 2025 CMO Spend Survey found that 39% of CMOs planned to reduce labor costs in the same year that 39% planned to cut agency spend — both levers are being pulled at once, which suggests the real decision for most marketing leaders isn't "in-house instead of agency," it's which specific roles and relationships earn their cost, in-house or external. For AI visibility specifically, that means asking a narrower question: is the cost of building AEO/GEO capability in-house, from scratch, actually lower than paying for expertise that already exists — once ramp-up time and tooling are counted?

The skills and tooling gap most teams underestimate

The skills gap in AI-visibility work is not hypothetical — the data on generative AI adoption inside marketing teams shows a wide split between using AI tools and having a mature, repeatable process for them. Content Marketing Institute found that while 81% of B2B marketers now use generative AI tools, only 19% have integrated AI into their daily workflows in a structured way; 54% describe their approach as ad hoc experimentation. The same research found real capability gaps inside marketing tech stacks more broadly: 47% of B2B marketers say they lack efficient lead-generation processes and 45% say they cannot yet support data-driven decision-making with their current tools.

Gartner 2026 CMO Spend Survey puts a sharper number on the readiness question specifically for AI: only 30% of CMOs report mature or fully developed AI readiness, and 70% acknowledge their internal marketing processes are not yet mature enough to implement and scale AI effectively — even as they plan to keep increasing AI investment.

For AEO/GEO specifically, that maturity gap shows up in tooling most in-house teams haven't budgeted for yet: AI-citation tracking (monitoring what ChatGPT, Perplexity or AI Overviews actually say about a brand), structured-data auditing tools, and prompt-level visibility monitoring. None of this is exotic once you know it, but very few generalist marketing hires walk in already fluent in it — which is the real argument for treating this as a build-or-buy decision on a specific skill set, not a referendum on the team's overall quality.

How fast should you expect results, and from whom

Timeline expectations should be set separately from the in-house-vs-agency decision, because AI visibility work is inherently slower to compound than a one-time technical fix. AI Overviews, ChatGPT and Perplexity are trained and re-indexed on cycles the platforms control, not the marketer — a structured-data fix or a rewritten answer-first page can take weeks to show up in citation tracking, and results tend to build gradually as a brand's entity signals become more consistent across the web, not in a single release.

What differs between in-house and agency isn't the ceiling on results, it's the ramp-up before the first real signal. An agency that has already run AEO/GEO audits for other clients typically has a repeatable diagnostic — which pages already get cited, where structured data is missing, which competitors are winning AI answers — that it can run in the first two to four weeks. An in-house team starting from zero usually spends that same window building the diagnostic itself, plus setting up AI-citation tracking that may not already exist in the tool stack.

That gap narrows fast once a team is running, though. HubSpot research shows marketers report growing confidence in adapting to AI-driven search change — 70.2% now say they can adapt their strategy to organic-search shifts like AI Overviews, and marketers' self-reported understanding of how to measure AI's impact rose from 48% to 67.5% year over year. The skill is learnable — the question is simply how much of that learning curve you want to pay for versus absorb.

A practical way to decide

Rather than treating this as a permanent identity choice, it helps to answer four concrete questions.

Do we already have technical SEO and content production capability to extend? If a content team and a technical SEO function already exist and are performing well, adding AEO/GEO as a specialized skill on top of that team is usually faster and cheaper than starting an external relationship from zero.

Do we have, or are we willing to buy, AI-citation tracking and structured-data tooling? Without visibility into what ChatGPT, Perplexity and AI Overviews are actually citing, neither an in-house team nor an agency can prove impact. This is a non-negotiable line item either way.

Is this a one-time project or an ongoing discipline? AI-visibility work is not a launch-and-done SEO audit; retrieval patterns shift as platforms update their models. A one-time engagement — a technical AEO audit, a structured-data cleanup — suits a project-based agency relationship. Ongoing monitoring and iteration suits either a retained agency partnership or a dedicated in-house owner, but rarely a one-off internal project squeezed into someone's existing job.

How fast do we need the first credible read on where we stand? If leadership needs a benchmark against competitors inside the next month, an agency with existing frameworks will usually get there faster than a team building its diagnostic process from scratch.

None of these questions have a universally "correct" answer — they're inputs, not a formula. But answering them honestly, before comparing rate cards, is what actually determines fit.

The hybrid model: why many teams end up blending both

In practice, a growing number of teams don't pick one model — they split the work by what each side does best. A common pattern: keep content production, brand voice and day-to-day publishing in-house, where product knowledge and speed matter most, while bringing in an agency or specialist for the technical AEO/GEO layer — structured data architecture, AI-citation tracking, and platform-specific optimization across ChatGPT, Google AI Overviews and Perplexity — where narrow expertise compounds fastest.

There's a data signal behind this shift, too. Gartner 2025 CMO Spend Survey found that 22% of CMOs say generative AI has already let them reduce their reliance on external agencies specifically for creativity and strategy work — meaning some of what agencies used to be hired for is increasingly handled in-house with AI tools, while the harder-to-automate specialized layers (technical implementation, cross-platform strategy, measurement) are where outside expertise still earns its cost.

The mix that works is rarely 50/50, and it shouldn't be treated as a permanent split either — the right blend today, while AEO/GEO is still a maturing discipline most teams are learning in real time, may look different in two years once in-house teams have absorbed more of the skill set themselves. The honest way to run a hybrid model is to name, in writing, exactly which deliverables sit where, and revisit that split on a fixed schedule — quarterly is reasonable — rather than letting it drift by default.