Here’s a thing that’s happening right now, all across the agency world, and almost nobody is talking about it openly: your agency is very likely using AI tools to do a significant chunk of the work they bill you for every single month, and your retainer hasn’t budged a dollar. You’re almost certainly paying senior-consultant hourly rates for work that a $20-a-month subscription to ChatGPT, Jasper, or any number of other tools is now cranking out in a few minutes.
We’re not saying *every* agency is actively lying to you…some of them are using AI to do BETTER work faster and the economics make sense for everyone involved. But a lot of them are using it to do the same-old shitty SEO work faster, pocket the margin expansion, and not mention any of it to you.
We’ve spent years doing agency audits (originally for SEO and content) and going deep into what clients are actually getting versus what they’re paying for, and right now the AI question is the single most important thing to add to that checklist.
The Numbers Make It Obvious Why Agencies Love This
Let’s do some basic math, because it’s pretty clarifying. The average digital marketing agency runs on net profit margins somewhere in the 20% range, with specialized shops (SEO, PPC, content-heavy) sometimes pushing closer to 25-30% (we’ve seen department specific net margins at 50-75% in some agencies). Most expenses in an agency are labor, so labor efficiency is basically everything, which is why the AI productivity numbers agencies are seeing right now are, from their perspective, absolutely beautiful.
Harvard Business School and Boston Consulting Group ran a properly controlled experiment back in 2023 that put 758 BCG consultants on a range of knowledge-work tasks with and without GPT-4 access. The results were hard to ignore: consultants using AI completed tasks 25% faster, were 12.5% more likely to finish tasks successfully, and produced output that independent evaluators rated 40% higher in quality. For tasks that played to AI’s strengths, using it was almost strictly better in every measurable way. That study is called “Navigating the Jagged Technological Frontier” if you want to look up the full thing, and it’s worth reading.
Now, that study was looking at management consultants, but content writing, keyword research, SEO analysis, competitive research, and monthly reporting… that’s EXACTLY the kind of structured knowledge work where AI excels. If your agency’s team used to spend six hours writing a batch of blog posts and AI gets that down to 90 minutes of actual human time (prompting, editing, light QA), that’s a massive reduction in labor cost for that deliverable. Are they charging you proportionally less? I’ll let you answer that one yourself.
According to the 4As (American Association of Advertising Agencies) 2025 survey on generative AI inside U.S. marketing agencies, adoption is accelerating fast, with agencies using AI specifically across content, SEO, strategy, and reporting workflows. One notable finding: for most agencies surveyed, AI is currently described as a cost center internally… meaning they’re investing in it, getting the efficiency gains, but those gains aren’t being passed back to clients. The margin improvement stays on their side of the ledger. 😏
What ‘Quietly Running on AI’ Actually Looks Like
I want to be specific here because vague rants about AI aren’t useful to anyone. Here’s what it actually looks like in practice when an agency has AI-assisted a big chunk of their workflow without mentioning it to you:
Content that arrives suspiciously fast. A well-researched 2,000-word blog post with a real point of view used to be a 3-5 business day turnaround for research, drafting, and revisions. If it’s showing up in 24 hours, someone pressed a button. That’s not automatically bad work, but it’s worth knowing, and it should factor into what you’re paying.
AI loves a very predictable article skeleton: intro paragraph, a few H2 subheadings each with three or four bullet points underneath, thin conclusion with a call-to-action. If everything your agency delivers looks like that, down to the sentence rhythm and the way ideas are organized, you’re looking at unedited or lightly-edited AI output. Real writers go off-script. Real writers have weird tangents and opinions and things that don’t fit neatly into a list. Like pancakes…I mean, who doesn’t love a good pancake?
Reports that look great but say nothing actionable. Monthly reporting is one of the most automatable tasks in the whole agency workflow. Pulling performance data, writing narrative around trends, generating charts…a competent prompt engineer can automate 80%+ of a reporting workflow. If your reports feel copy-paste-y month to month, full of hedged language and observations that never actually recommend anything specific you should do, that’s a tell.
Or perhaps strategy docs that sound smart but feel hollow. This one’s harder to spot because AI can generate genuinely plausible-sounding strategy, but if the recommendations don’t reflect anything specific about your business, your competitors, your actual data…you’re reading a sophisticated-sounding generic document that could have been written for any company in your industry. Which it basically was.
We wrote more about the warning signs of a bad agency relationship in our post “Is Your SEO Agency Screwing You?”, and a lot of what used to show up as pure laziness or offshore outsourcing has just been replaced by AI outsourcing. Different mechanism, same basic outcome for you as the client. The agency world is built on churn-and-burn.
The Margin Math, and Why This Is More Than Just Annoying
Let me walk through a scenario that I think is going to feel pretty familiar…
You’re paying a $5,000 per month retainer for an agency handling your SEO and content. In their internal project setup, that retainer maps to some number of senior strategist hours, some writer hours, some account management time. You agreed to that arrangement because you were essentially buying those people’s expertise and time, and what they could produce with it.
Now the writer’s output is substantially AI-generated with some editing layered on top. The strategist’s monthly ‘research’ is a ChatGPT analysis they spent 20 minutes on instead of the half-day it used to require. The monthly report is a template with your data dropped in. Their actual human time on your account might be 40-50% of what it was 18 months ago, probably less, but the invoice is identical.
The agency’s costs just dropped significantly, their profit margin on your specific account went up, and you, sitting on the client side, have zero visibility into any of this. You don’t know what tools they’re using, you don’t know how much actual human time went into your deliverables, and most agency contracts sure as hell don’t include any clause about disclosing AI usage or adjusting rates based on efficiency gains.
The McKinsey State of AI report (November 2025) found that 88% of organizations are now using AI in at least one business function. The productivity gains are documented and real. If your agency isn’t using AI at all, honestly that would be a different kind of problem (they’re falling behind). The issue is the complete lack of transparency around it and the refusal to have an honest conversation with you about what your retainer is actually purchasing these days.
How to Find Out What’s Actually Going On
Okay, so what do you actually do?
Ask directly, and get the response in writing. This is the first move, and I know it feels awkward. Ask your agency: ‘What AI tools are you using on our account, and what percentage of our deliverables involve AI assistance?’ A good agency will (maybe) answer this honestly and walk you through their process. An agency with something to hide will give you a non-answer full of buzzwords about ‘AI-augmented workflows’ and ‘proprietary processes’ that sounds like it means something but doesn’t.
Read your deliverables critically. Pull a sample of the content they’ve produced for you over the last few months and just read it carefully. Is it specific to your business and audience? Does it make claims or observations that require actual knowledge of your industry? Is it actually GOOD? Or does it read like it was written about a company called [Client Name] with details filled in from a brief? You don’t even need AI detection tools to spot this, though those can help as a secondary signal.
Ask for a time breakdown. Request a simple summary of what hours went into your account last month. Any agency worth working with should be able to tell you roughly what they spent time on, even if it’s not a perfect timekeeping exercise. If they can’t or won’t provide anything meaningful, that’s an answer in itself.
Look at output volume versus team size. If your agency has dramatically increased the volume of what they’re delivering without increasing your retainer, without growing their team, or without explaining a process change… the math only works one of two ways. Either they got unbelievably more efficient (possible, and worth understanding how), or the quality and depth of what you’re getting thinned out significantly. You deserve to know which…though if you’re getting more output for the same price, at least they’re sharing the efficiency gains with you.
Honestly, if you want a systematic answer rather than just a vibe check, that’s exactly what our Agency Audit is designed to give you. We go through your actual deliverables, your contract terms, your billing, and your results data, and tell you plainly what you’re getting and whether it’s worth what you’re paying. No pitch baked into the findings. You can also reach out directly if you want to talk through your specific situation before committing to anything.
What to Actually Do About It
Say you go through the above and confirm your suspicions. Your agency is doing substantially AI-assisted work and your retainer has never been adjusted to reflect that. What are your real options?
At the very least, renegotiate. This is the most obvious one and also the most fair. If AI has genuinely reduced the labor input on your account, your retainer should reflect that in some way, whether that’s a lower price, more deliverables for the same price, or a clearer scope that reflects the actual work involved. A good agency will engage with this conversation. An agency that stonewalls you on it is telling you something important about how they think about the relationship.
Bring the automatable stuff in-house. If your agency is using AI to write your content, you could hire one smart person internally, equip them with the right tools and processes, and probably get comparable output for less than half what you’re paying the agency for that specific service line. The AI expertise that used to live only in agencies is not the moat it once was. The hard stuff, genuine strategy, technical SEO, media buying, creative direction, is (maybe) still worth paying for. The commodity output stuff? Nope nope nope.
Get smarter about what you’re buying. Going forward, the best agency relationships are going to be ones where you’re explicitly paying for expertise, judgment, and strategic oversight, not for raw output volume. An experienced SEO strategist who uses AI to research faster and produce more informed recommendations is worth every dollar. Paying $5,000 a month for AI-generated content with a light human edit on top, when you could replicate that for a fraction of the cost? That’s a different and much harder-to-justify proposition. We’ve been writing about this shift in our AI content optimization work, and it’s reshaping how smart companies think about outsourcing.
Explore what AI automation could do for you directly. This is worth a whole separate conversation (and we cover it in depth in our AI consulting practice), but the honest truth is a lot of what agencies charge you for is directly automatable if you have the right internal setup and guidance. That doesn’t necessarily mean abandoning agencies entirely, though it might, but at the very least it means using them for what AI actually can’t replace…yet.
The Bottom Line
The AI productivity wave is real, it’s running through the agency world at full speed, and 88% of organizations are now using AI in at least one business function according to McKinsey’s November 2025 data. Your agency is almost certainly using these tools. That’s actually fine and probably good. The question is whether you have any visibility into it, and whether your pricing honestly reflects the new economics of what they’re doing.
If you’re paying a premium retainer and getting AI-first output that was never discussed with you, never disclosed, never reflected in your pricing…you’re essentially subsidizing your agency’s margin expansion with no benefit flowing back to you. Agencies need to make money, obviously, I’m not suggesting they operate as nonprofits. But you need honest information about what you’re actually buying so you can make a rational decision about whether it’s worth it, and whether you should just take it in-house.
Get the information. Have the conversation. And if you want someone to go through your agency’s actual work and tell you straight what you’re getting for what you’re paying, that’s exactly what our Agency Audit is built for. We’ve been through tons of these, and we’ll give you an honest read, whether that means your agency is doing great work that’s worth every penny, or whether it’s time to renegotiate, restructure, or walk. You’re welcome. 😎
Sources
BCG / Harvard Business School “Navigating the Jagged Technological Frontier” (2023), HBS AI Institute: https://aiinstitute.hbs.edu/navigating-the-jagged-technological-frontier/
Full working paper on SSRN (Dell’Acqua, McFowland, Mollick, et al.): https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4573321
Harvard Crimson coverage of the BCG/HBS AI study: https://www.thecrimson.com/article/2023/10/13/jagged-edge-ai-bcg/
4As “The State of Generative AI Inside US Marketing Agencies, 2025”: https://www.aaaa.org/resource/the-state-of-generative-ai-inside-us-marketing-agencies-2025/
McKinsey “The State of AI in 2025” (November 2025): https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
Agency CPA “Average Agency Profit Margin in 2024”: https://agencycpas.com/profitability/agency-profit-margin-2024/
Digital Marketing Institute “10 Eye Opening AI Marketing Stats” (2025): https://digitalmarketinginstitute.com/blog/10-eye-opening-ai-marketing-stats-in-2025
Digital Applied “AI Marketing Statistics 2026”: https://www.digitalapplied.com/blog/ai-marketing-statistics-2026-adoption-data-points