Agencies That Know More Win More Retainers

TL;DR: AI hasn’t replaced agencies — it replaced what agencies traditionally charged for: content production and distribution. Operators running paid acquisition across high-CAC verticals need agency partners who bring proprietary intelligence to every call, not a team that pastes briefs into ChatGPT. The agencies still standing in 2028 will treat industry research as infrastructure, not a Friday afternoon task.

The Numbers Agencies Don’t Want to Talk About

December 2025: Omnicom cut 4,000 jobs and folded DDB, FCB, and MullenLowe into its Interpublic integration. A year earlier, Edelman let go of 330 people — 5% of its workforce — citing an expected 8% revenue drop. From December 2024 through May 2025, Ad Age tracked consecutive monthly declines in U.S. advertising and PR employment.

The easy narrative is that AI ate the headcount. The real story is sharper: AI ate the billable hours. Agencies built their margin on three things — access to distribution, capacity to produce content at scale, and the appearance of expertise. Two of those three are now free. Anyone with a laptop can publish anywhere instantly. Any operator can generate a 1,200-word draft in under a minute. What AI cannot do is walk into a client meeting with a tight, current, differentiated point of view that the client couldn’t source on their own. That gap is where agency value lives now.

For operators in high-CAC verticals — iGaming, forex, crypto, legal — this shakeout matters. If your agency’s value proposition is “we make content,” you are paying retainer fees for something that commoditized 18 months ago.

Why the Retainer Model Is Under Pressure

A retainer has always been a relationship contract, not just a deliverables contract. Clients supply business strategy, institutional knowledge, and industry context. Agencies supply expertise, capacity, and availability. That trade still holds in theory. In practice, the capacity side of the equation collapsed when LLMs became ubiquitous.

When a marketing director at a CFD brokerage can paste a brief into ChatGPT and get a usable draft in 45 seconds, the obvious question surfaces: what is the agency providing that this tool cannot? The answer cannot be “a polished version of what you just got for free.” It has to be something upstream of the draft — the intelligence, the angle, the competitive context that makes the draft worth reading in the first place.

Every time a client has to catch their agency up on something the agency should already know — a regulatory change in the jurisdiction they operate in, a competitor’s new offer, a shift in the media cycle — retention erodes. The client feels like they are managing the agency rather than the agency managing the account. That feeling kills contracts faster than a missed deadline.

Operators using paid media partners for lead generation need to hold their content partners to the same standard: show up informed or lose the seat at the table.

Content Intelligence Is the New Core Deliverable

The agencies that are still growing treat content ideation as a deliverable, not prep work. They walk into every client call with something new: a regulatory development the client hasn’t processed yet, a competitive move that deserves a direct response, a thought leadership angle the client’s leadership is uniquely positioned to own.

This is not about producing more content. Audiences want less content, not more. The web is already overloaded with generic AI output that is plausible for anyone and owned by no one. Volume is not the metric. Relevance and specificity are.

Six steps define a repeatable content intelligence process:

• Listen continuously — not vanity mention-tracking, but mining live conversations in industry forums, regulatory feeds, and competitor channels for gaps and angles.
• Identify the right signal — filter for what your client can say that nobody else can. High-impact content wins through contrast, not consensus.
• Repurpose across formats — a strong angle stretches into three to five formats (LinkedIn post, newsletter, solutions brief, outbound email) without diluting the point.
• Sequence across a calendar — spreading the formats over several weeks multiplies reach without requiring new research for each piece.
• Measure what landed — format and topic performance data closes the feedback loop and seeds the next content cycle.
• Repeat per client — the process only scales if the intelligence layer is organized, not haphazard.

Running this process manually for multiple clients simultaneously is where most agencies break down. The upstream work — research, signal identification, competitive mapping — is time-intensive and depends on whoever happened to see what, when they saw it, and whether they remembered to share it. Using AI at the draft stage while keeping the research layer manual is like putting premium fuel in a car with a broken engine. Optimizing upstream is where the efficiency gains compound.

E-E-A-T and Thought Leadership Are Not the Same Thing as “More Blog Posts”

Google’s E-E-A-T signals (experience, expertise, authoritativeness, trustworthiness) have become more consequential as AI-generated content floods every index. Thought leadership content — written from a specific person with demonstrable authority — carries the trust signals that both traditional search and AI-powered search engines weight heavily. Publishing consistent thought leadership from a named expert in a defined niche teaches AI systems who owns that subject.

For operators in regulated verticals, this matters operationally. A personal injury law firm that publishes consistent, authoritative commentary on mass tort developments from a named partner is building an asset that generic AI output cannot replicate. The same applies to iGaming operators publishing compliance and product differentiation content, or forex acquisition teams positioning around regulatory changes in their target jurisdictions.

Generic AI output is not thought leadership. It is plausible text. Those are different products with different market values.

What This Means for High-CAC Vertical Operators

If you are spending $10K or more per month on marketing in a vertical where the average cost to acquire a qualified lead runs $200–$2,000, content strategy is not a soft concern. It feeds your paid acquisition funnel, your landing page conversion rates, your retargeting sequences, and your organic search visibility.

Operators in crypto and web3 know how fast the conversation moves — a regulatory announcement or a protocol exploit can shift the entire content landscape in 48 hours. Agencies that are not plugged into those conversations in real time are producing content that is already stale when it publishes. Operators in legal marketing — particularly mass tort — know that intake volume is directly tied to how quickly a firm’s content responds to emerging case developments. Slow content equals missed cases.

For CDL recruitment operators, the window to hit a driver who is actively looking is narrow. Content that speaks directly to the driver’s specific pain points — routes, home time, pay structure — outperforms generic “join our team” messaging at every stage of the funnel. That specificity requires ongoing intelligence, not a quarterly content calendar built in a vacuum.

A structured marketing audit is a practical starting point to identify where your current agency’s intelligence process is breaking down and where content is lagging the conversation in your vertical. Combine that with precision targeting at the paid layer and you close the loop between what your content says and who actually sees it.

The Agency Standard That Operators Should Demand

The agency you retain in 2026 should be able to answer three questions without hesitation before every client call: What changed in your industry this week that affects your positioning? What is your closest competitor doing right now that you should respond to? What is the one angle your leadership is uniquely positioned to own that no one else in the market is running?

If the answer to any of those questions is “we’ll pull that together before the next meeting,” you are paying for capacity, not expertise. Capacity is cheap. Expertise — demonstrated continuously, not declared in a pitch deck — is what justifies the relationship and the cost.

Agencies losing retainers are not losing because they picked the wrong AI stack. They are losing because they charged for something that became free and never replaced it with something clients cannot get elsewhere. The replacement is intelligence — organized, continuous, and actionable before the blank page ever appears.

Originally reported by Search Engine Journal, July 2026.