Avinash Kaushik spent roughly 16 years inside Google, has held senior client-side roles at companies like Intuit and DirecTV, and now advises brands and agencies as chief strategy officer at Human Made Machine. So, when he tells marketers to renegotiate their agency contracts and expect savings of 25% to 75%, starting this month, it is worth stopping whatever else you are reading.
I have known Avinash for more than 20 years. He does not throw out numbers like that for shock value. He builds a model, shows his work, and dares you to argue with the math. In his latest Marketing < > Analytics Intersect newsletter, “Pay Less, Grow More, Agencies in an AI-Era,” he does exactly that, and the argument reaches well past the performance and creative agencies he names directly. Anyone running SEO, content, or GEO work through an outside partner should be doing this same math on their own contract this week.
Kaushik’s starting point is that three forces converged at once. AI got broadly, generally smart, not just smart at one narrow task. The ad platforms happen to also own the foundational AI models running underneath the campaigns those platforms sell, so intelligence showed up fast and is already wired into the tools agencies use daily. And every system in the stack now talks to every other system in real time, which does not make AI smarter in the abstract; it makes AI smarter about your specific account. Put those three together, and you get what Kaushik calls the “we are not in Kansas anymore” moment for every type of agency, media, creative, performance, brand, measurement, and yes, SEO.
The practical result is that work that used to justify a monthly retainer is now handled by the platform itself, and Kaushik says clients should expect 25% to 75% savings on existing scopes of work as that shift plays out, alongside 15% to 25% growth in fees for genuinely new work the old contract never anticipated. He is not arguing for a smaller relationship with your agency. He is arguing that the money should move to different work.
Where The Old Contract Stops Making Sense
Kaushik breaks the old agency scope into clusters, and the pattern across every one of them is the same. Account architecture, keyword and audience structuring, and campaign build-out once ran roughly a fifth of a typical contract’s cost, and he estimates that work can shrink by close to 80% now that platform algorithms handle segmentation and targeting better than a human team slicing campaigns “for control.” Manual bid and pacing adjustments told a similar story, with AI already outperforming human pacing decisions since late 2024, and Kaushik makes a sharper point here that deserves attention beyond his own newsletter. He argues that an agency jumping in to “rescue” a dip during an AI learning cycle is not helping, it is actively sabotaging the algorithm’s ability to learn, and every one of those manual rescues resets the clock.
Reporting tells the same story from a different angle. Weekly decks, twice-a-week status meetings, and hand-typed commentary on numbers that already live in a dashboard used to eat close to a third of a contract’s cost by Kaushik’s estimate, and he thinks 60% of that can go away now that AI-fronted data tools can explain what happened and why without a human translating a spreadsheet. None of this means the work disappears. It means the work moves to a machine, and the agency’s real job becomes deciding what the machine should optimize toward, not building or reporting on what the machine already does on its own.
The Contract Structure Kaushik Wants To Replace It With
Here is where the column stops being descriptive and starts being prescriptive, and it is the part I think SEO leaders will recognize fastest. Kaushik wants agency fees split three ways. A lean base retainer, roughly 40% to 50% of a now-smaller total, covers governance, steering, and data engineering. Project fees, another 30% to 40%, cover the work that still needs real human judgment, creative concepting, complex strategic analysis, portfolio strategy. And an outcome incentive, the remaining 15% to 25%, ties directly to incremental profit or verified revenue lift, never to a platform-reported metric like ROAS that the platform itself has every incentive to inflate.
That last point is the same argument I have been making about Citation Share of Voice and GEO measurement for the better part of this year. Kaushik’s warning against paying agencies a “percent of media spend” is structurally identical to the warning I would give any SEO team still paying an agency for deliverables, page counts, audits shipped, blog posts published, instead of paying for organic revenue actually generated. A contract built around activity always rewards more activity. It never rewards the judgment to do less and get more, and in an AI-mediated search environment, judgment is the only thing left that a machine cannot fully do for you.
What SEO Teams Should Actually Do With This
Kaushik’s own examples run through Performance Marketing and Google Ads, but the underlying math applies just as directly to SEO and content agency relationships, and here is how I would apply it.
First, pull your current SEO or content agency SOW and sort every line item into one of Kaushik’s three buckets. Anything that looks like keyword research busywork, manual rank tracking, or templated technical audits belongs in the “platform already does this” pile, because AI-driven crawlers, Google Analytics 4 anomaly detection, and automated technical monitoring have closed most of that gap. Be honest about how much of your current retainer is quietly paying for that pile.
Second, propose splitting your next renewal the same way Kaushik does, a smaller base retainer for governance and data ownership, project fees for the work that genuinely needs a strategist’s judgment, like entity building, content architecture for AI Mode and AI Overviews, or GEO strategy, and an outcome incentive tied to organic revenue or a verified Citation Share of Voice lift, not to a vanity metric like keyword rankings or deliverable counts that any AI tool can now generate on command.
Third, and this is Kaushik’s warning applied directly to SEO, make sure you own your own data before you renegotiate anything. GA4 access, Search Console, log files, and whatever AI citation tracking you are running need to sit in your hands, not your agency’s, or you have no leverage to ask for any of the above.
My Take
I do not think this means SEO agencies are becoming worthless, and neither does Kaushik about agencies broadly. I think it means the agencies that survive the next two years will be the ones that stop selling hours and start selling the judgment a machine still cannot supply, and the ones that keep billing for monthly rebuilds and manual reports are going to lose the client before they lose the argument.
Kaushik ends his piece with a line I keep coming back to. You can pay for the past, or you can pay for the present. After 20 years of watching agency-client relationships in this industry evolve, I think that is the correct frame, and I think most SEO teams are still writing checks for the past without realizing it.
More Resources:
- What Not To Automate With AI: The SEO Deskilling Trap
- 5 Key Enterprise SEO And AI Trends For 2026
- Well-Known SEO Explains Why AI Agents Are Coming For You & What To Do Now
Featured Image: Master1305/Shutterstock
