Does labelling your agency 'AI-first' actually win pitches?

Not on its own. "AI-first" wins evaluations only when the pricing model, the deliverable list and the named case studies all shift with the label - without those three, procurement now deselects "AI-first" shops as fast as it used to shortlist them.

Twelve months ago, "AI-first" was a shortcut past the first shortlist call. In 2026 it is a filter that works in both directions. A rising share of client-side marketing leads - a group of them wrote about it on LinkedIn this week - now flag any agency whose feed reads as 100% AI opinion as a red flag rather than a differentiator. Ritson's Great Stay piece in The Drum landed the same week and pointed at the same thing: the market is starting to prefer agencies that clearly kept their people over agencies that clearly did not.

Meanwhile the AI-first cohort keeps growing. Directories like Humaine and features in TechRound now list dozens of UK "AI-first" shops. When the label is in a directory, the label has stopped being a differentiator.

What follows are the four questions procurement is now using to decide whether "AI-first" is real, and what a truthful answer looks like in each.

Has your pricing model changed?

If an "AI-first" agency still bills on person-days at broadly the same rate card, procurement will notice inside one call. The whole point of the shift, from the client side, is fewer hours at lower total spend for the same output. Agencies that promise faster delivery and same-margin billing get read as marketers, not operators.

The agencies converting on the label have moved to one of three shapes: outcome-based fees, tiered subscriptions with clear scope caps, or hybrid rate-cards where the AI-heavy deliverables are priced per artefact rather than per hour. All three are defensible in a procurement meeting. "We use AI internally to move faster and charge the same" is not.

Has your deliverable list actually changed?

The second question is what you are giving the client that a non-AI agency would not. If the answer is "the same monthly brand strategy report, but drafted faster", the label is decorative.

The strong answers here are specific. An AI-embedded PR shop is delivering weekly citation audits inside client dashboards, not quarterly PDFs. An AI-embedded creative agency is delivering a hundred iterations of a script for testing, not three. A B2B agency is running a monthly LLM-visibility report against ten priority buyer-persona queries. Any of those is a deliverable a traditional agency does not have on the price list.

Are the case studies named?

Every "AI-first" pitch deck opens with a workflow diagram. Very few open with a named client, a named metric and a signed logo. The signed logos are what convert.

Procurement now asks two follow-ups: which client is on record about the AI work, and can we speak to them. The agencies that pass have three or four references who will actually take the call. The ones that fail either name-check pilots that never went public, or hide behind NDAs the client would happily lift for a good conversation. If your case-study slide is stock imagery and abstract percentages, procurement assumes there is nothing behind it.

Does your public content match your positioning?

The last question is the LinkedIn one. When procurement checks the founder feed and the agency page, they are not looking for the number of AI posts. They are looking for a ratio.

The agencies winning on "AI-first" are running roughly a third AI content, a third client work, a third industry commentary. The agencies losing are running 100% AI content on the founder's feed and zero client work on the agency page. The signal that reads well to a marketing director is "this shop uses AI and also runs a full client book". The signal that reads badly is "this shop's entire public presence is opinions about AI".

That is a fixable content strategy problem, not a positioning problem.

Is your team still visibly hiring humans?

The fifth question is quieter but increasingly decisive. Ritson's Great Stay data landed the same week the AI-first directories filled up, and it pointed at a specific behaviour on the client side. Marketing directors who have watched their own graduate rung disappear are wary of agency partners who look like they cut theirs too.

The agencies that read as credible on AI-first now do one particular thing on their careers page and on LinkedIn: they show, on record, that they are hiring humans - specifically junior humans - into roles the AI is meant to have automated. A "senior AI-strategy lead" role paired with two "junior insight analyst" roles reads honestly. A careers page that has not been touched since 2024 reads as a shop that quietly stopped.

Procurement will not ask this question directly. They will notice.

The short version

"AI-first" is a positioning claim. Every claim now costs the same in procurement - a follow-up question, a reference call, a look at the feed. The claims that win are the ones with pricing, deliverables and named clients behind them. The claims that lose are the ones that are only claims.

If your agency has moved the pricing, changed the deliverables, and can name three clients who will take the call, keep the label. The market will reward you.

If none of those three are true yet, drop the label until they are. The market will punish you less for repositioning than for pretending.