Agencies are getting pressed from both sides right now. Clients question whether they need to pay as much for services they think AI can handle. Agencies respond by racing to cut costs with AI tools.
Meanwhile, AI capabilities are advancing at a pace that makes last quarter’s tools look quaint.
The winners won’t be the cheapest agencies. They’ll be the ones using AI to create value that didn’t exist before.
Here’s what caught my attention this week.
The Squeeze Play
Agencies are being compressed from two directions at once. A Search Engine Land analysis lays out the dynamic: clients see AI tools and question their agency spend, while agencies scramble to become more efficient with every new release. The data is sobering. Only 14% of agencies describe their current pipeline as “very healthy,” and owner-led agencies are hit hardest because business development competes with client delivery for the owner’s time.
Why it matters: If your response to AI pressure is to do the same work faster and cheaper, you’re in a race to the bottom. The bar for what counts as real agency value has risen dramatically now that clients can handle basic execution – and even strategy – themselves. Agencies need to move up the value chain, not just move faster on the same treadmill.
The workforce numbers confirm the shift. An Ipso/Epoch AI survey found that 20% of full-time workers say AI already handles parts of their job. But here’s the number that matters more: 15% report doing entirely new tasks they wouldn’t have taken on without AI.
Why it matters: That 15% figure is the model for agencies. The people thriving with AI aren’t just automating what they already did. They’re finding new work worth doing. Half of American adults now report using AI in the past week, and most workers who use it do so multiple days per week. This isn’t a trend on the horizon. It’s the current reality your clients are living in.
The Mythos Moment
Anthropic built an AI model so capable at finding security flaws that it won’t release it publicly. The company’s new Mythos model, announced as part of Project Glasswing, has identified thousands of previously unknown vulnerabilities in every major operating system and web browser, some of which had gone undetected for decades. The oldest was a 27-year-old bug in OpenBSD, an operating system specifically known for its security. The story broke through to mainstream media with widespread coverage, including the New York Times, Fortune, Axios, and TechCrunch.
Why it matters: Mythos wasn’t specifically trained for cybersecurity. These capabilities emerged as a side effect of general improvements in coding, reasoning, and autonomy. That’s the part agency owners should pay attention to. AI isn’t just getting incrementally better at narrow tasks. It’s developing unexpected capabilities as a byproduct of getting smarter overall.
The bigger picture for agencies: You don’t need to understand the technical details of zero-day vulnerabilities. What you need to understand is the pace. The gap between what AI could do six months ago and what it can do today is larger than most people realize. Agencies that put off thinking seriously about how AI fits into their operations are falling further behind with each model release. This doesn’t require panic, but it does require intentional investment in understanding and adopting these tools.
Shortcuts will cost you
Reddit is cracking down on bots, and the broader trend is clear. Reddit CEO Steve Huffman detailed new measures requiring suspicious accounts to verify their humanity, while introducing mandatory labels for all automated accounts. The platform is removing roughly 100,000 bot accounts daily. Reddit isn’t alone here. The former competitor Digg recently shut down its app entirely, citing an inability to control its bot problem.
Why it matters: The temptation to use AI for slightly more sophisticated spam is real, and some agencies are already fielding client requests to automate engagement on platforms like Reddit. But platforms will keep finding ways to prioritize genuine human activity. The smartest approach emerging right now is a “human-in-the-loop” model where AI helps draft and surface opportunities, but a real person reviews and posts. Use AI to make human-to-human communication easier, not to replace it.
Free AI isn’t free anymore. Anthropic blocked Claude subscribers from using their flat-rate plans with third-party agent tools like OpenClaw, effective April 4. Users who had been running autonomous AI agents under a $20 or $200 monthly subscription now face cost increases of 10 to 50 times their previous monthly spend. Over 135,000 OpenClaw instances were estimated to be running at the time of the announcement.
Why it matters: The economics are simple. A single autonomous AI agent running for a full day can consume the equivalent of $1,000 to $5,000 in API costs. Flat subscription pricing was never designed for that, and the AI companies are now correcting the mismatch. If you’re building agency processes or pricing client services around the assumption that AI will always be dirt cheap, you’re building on a foundation that’s already shifting. AI will likely remain more cost-effective than equivalent human labor for many tasks, but plan for sustainable pricing, not today’s introductory rates.
The smart investment
Your AI tools are only as good as the knowledge you feed them. Jamin Ball’s Clouded Judgement newsletter makes a compelling case for “digital twins,” the concept of representing your knowledge, workflows, and institutional judgment digitally so AI agents can actually act on it. Ball outlines several flavors: capturing workflow knowledge so an agent can handle edge cases, preserving institutional memory when employees leave, and creating “expert twins” that give your whole team access to your best performer’s judgment.
Why it matters: This explains why so many agency owners try AI tools, get mediocre results, and conclude the technology isn’t ready. They’re skipping the same step they’d never skip with a new employee: proper onboarding. When you hand an AI tool a task without context about how your agency works, what your clients expect, and what good judgment looks like in your business, you get generic output. Creating a digital twin of your workflows, decision-making patterns, and institutional knowledge is what turns AI from a generic assistant into something that actually thinks and acts the way you would. The investment is real, but so is the payoff.
The Bottom Line
The squeeze on agencies is undeniable, but the answer isn’t to become a cheaper version of yourself. It’s to become a more valuable one. Use AI to invest in the knowledge layer of your business, to find new services worth offering, and to free up your team’s time for work that requires human judgment.
The agencies that treat AI as a cost-cutting tool will keep getting squeezed. The ones that treat it as a value-creation tool will thrive.
