Every agency owner I talk to eventually asks the same question. Should we build our own AI tools, or just buy something off the shelf? It usually comes up after a client asks why the agency isn’t “doing more with AI” yet, or after a competitor starts talking up their AI-powered reporting in a pitch.

The honest answer for most agencies between $1M and $25M in revenue is simple. Buy it. Don’t build it. And the reason has nothing to do with whether your team is technical enough. It’s about what building actually costs you, measured in the work you’re not doing while you build.

The build vs buy question is really a time question

A recent piece on the AI agent landscape for businesses framed this well: the decision isn’t about capability anymore, most off-the-shelf agents can do what a custom build would do. The decision is about where you want your best people spending their hours. For a 20-person creative shop, that’s not an abstract tradeoff. It’s the difference between your senior strategist writing prompts and testing pipelines for three months, or that same person sitting in front of a client explaining a campaign pivot.

Agencies this size don’t have a dedicated engineering team sitting idle. The people capable of building a custom AI workflow are the same people who bill out at your highest rates and run your biggest accounts. Every week they spend building internal tools is a week they’re not doing the work clients actually pay for. That’s the real cost of “build,” and it rarely shows up on a budget line.

Where the hours actually go

Before you can decide what to buy, you need to know what you’re buying it for. In most agencies we talk to, the time drain concentrates in three places.

Reporting and client comms. Account managers spend a huge chunk of their week, often somewhere between 30% and 50% depending on account load, pulling numbers from ad platforms, formatting decks, and writing the Slack or email summary that goes with them. None of that is strategy. It’s data entry with a nice font.

Content production cost. The volume of content requests keeps climbing every year, more platforms, more formats, more variations for testing. But the price you can charge per asset hasn’t kept pace. Somewhere in there, the per-piece cost of production quietly becomes the thing eating your margin.

The account scaling ceiling. Most AMs cap out somewhere between 6 and 10 accounts before quality drops. Past that point, the only lever an agency has to grow is hiring, and hiring is exactly what compresses margin. You’re adding headcount to handle volume, not to grow strategic value.

These three pains are why the build vs buy question matters at all. If you’re only dealing with one small workflow, building might make sense. But agencies dealing with all three at once need something that works now, not something that’s ready in six months.

What “buy” looks like in practice

This is where the abstract debate gets concrete. At Enterprise DNA we build and deploy a set of named agents inside our Omni system specifically for agency operations, and I want to walk through what three of them actually do, because “AI agent” gets thrown around so loosely it’s stopped meaning anything.

The Reporting Agent

The Reporting Agent connects to every platform your accounts run on, ad platforms, analytics tools, CRM, whatever you’ve got. Once a month, it pulls the performance data, builds the client report in your template, and drafts the AM’s email summary in their voice. The AM reviews it, tweaks a line or two, and sends. What used to take a day and a half per account now takes twenty minutes of review.

This is the single highest-leverage place to start for most agencies, because reporting is pure hours-in, no-strategy-out work. It’s also the easiest one to point to on a P&L. If your AMs are spending a third of their week on reporting and you’ve got six AMs, that’s roughly two full-time salaries worth of hours going into work a system can do faster and more consistently.

The Content Production Agent

The Content Production Agent takes a creative brief and produces a first-pass draft, on-brand, on-format, ready for a human to edit rather than write from a blank page. It’s not replacing your creative team. It’s replacing the blank page. Editing a draft that’s 70% right takes a fraction of the time that starting from nothing does, and it means your writers and designers spend their time on the judgment calls that actually require a human, not on typing out the fourth variation of a social caption.

For agencies dealing with rising volume and flat per-piece pricing, this is the lever that protects margin without cutting quality. You’re not asking clients to pay more per asset. You’re changing how much human time goes into each one.

The Account Health Agent

The Account Health Agent runs in the background every day, watching account activity, spend pacing, engagement trends, whatever signals matter for that client. When something looks off, or when there’s an opportunity worth flagging, it drafts the next-step message before the AM even has to go looking for the problem. This is the one that quietly raises the ceiling on how many accounts a single AM can run well, because the AM isn’t spending time monitoring, they’re spending time acting on flags that are already teed up.

None of these three took our team building from scratch to stand up for a client agency. That’s the point of buying versus building. The infrastructure already exists. What takes time is configuring it correctly for your specific accounts, platforms, and brand voice, which is a very different project than building a model or an agent framework from zero.

Why building custom rarely pays off at this size

I want to be direct about this because agency owners get pitched “custom AI builds” constantly right now, usually by dev shops that would very much like a six-figure project.

A custom build means you’re now responsible for the infrastructure, the maintenance, the platform integrations breaking when an ad platform changes its API, and the ongoing work of keeping the model current as tools evolve. You need someone on staff, or on retainer, who understands this stuff well enough to fix it when it breaks at 11pm before a client deadline. Most agencies under 50 people don’t have that person, and hiring one costs more than most off-the-shelf agent platforms would cost for years.

There’s also a slower cost that’s easy to miss. Building takes months. Buying takes weeks. Every month you spend building is a month your competitors are already running lean on reporting and content, quoting tighter timelines, and protecting margin you’re still losing. The AI agent landscape moves fast enough that a six-month build risks landing outdated the week it ships.

None of this means “buy anything with AI in the name.” It means buy something built for your specific workflows, from a partner who understands agency operations, not a general-purpose chatbot wrapper. That distinction matters more than build versus buy itself.

The dollar math, not the hype

Let’s put this in terms that matter more than “AI adoption.” If you run a 20-person agency with six account managers, and each one is spending a third of their week on reporting instead of strategy or growth work, you’re paying full salary for roughly two heads’ worth of hours that go into a task an agent can do in minutes. That’s before you count the content production inefficiency or the ceiling on how many accounts you can profitably run per AM.

Multiply that across a full year and you land in the range most agencies your size are actually leaking, somewhere between $60,000 and $180,000, depending on account count, AM salary levels, and how manual your current reporting and content workflows are. That range isn’t a scare number. It’s what we typically find when we sit down with an agency owner and actually map the hours against the payroll.

The fix isn’t “hire more AMs” or “cut content quality to save time.” It’s removing the manual layer from work that was never strategic in the first place, so the people you already pay well are doing the work that actually justifies their rate.

What this looks like started, not theorized

If you’re weighing build versus buy right now, the fastest way to get a real answer isn’t another vendor deck. It’s an audit of your own operation. We run something called the Omni Audit, a 60-minute session where we map your actual account load, reporting hours, and content pipeline against what a set of configured agents would take off your plate. You walk away with three specific outputs, no generic slide deck, no sales pitch dressed up as a “strategy session.”

You can see Omni for marketing and creative agencies to get a sense of how the audit is structured before you book anything. It’s worth doing even if you’re not sure you’re ready to move, because you can’t make a good build-versus-buy decision without knowing your actual numbers first.

For agencies specifically wrestling with reporting load, our ops-focused agent work covers the Reporting Agent, Content Production Agent, and Account Health Agent in more depth, including how they get configured for an individual agency’s platforms and brand voice. If you want the broader picture of how AI agents fit into agency operations beyond these three, our insights collection has more write-ups on specific use cases we’ve built for agencies in this size range.

Booking the audit

The 60-minute session is deliberately short. We’re not trying to sell you a strategy engagement. We map your current account load, identify which of the three pains, reporting, content cost, or the scaling ceiling, is costing you the most, and show you what a configured agent setup would look like for your specific accounts. You leave with numbers, not a proposal.

Book a 60-min Omni Audit and bring your account list. That’s the only prep required.

The real decision in front of you

Build versus buy isn’t really about technical capability anymore. Off-the-shelf agent platforms, configured well, can do what a custom internal build would do, faster and without the maintenance burden. The real decision is whether you want your senior team spending the next two quarters building internal tools, or spending it on the client work that actually grows the agency.

For most agencies under 50 people, the math isn’t close. Buying frees up the hours that building would consume, and it does it in weeks instead of months. If you’re still weighing it, look at your own numbers first. Check the AI audit for marketing and creative agencies and see where your reporting hours, content costs, and account ceiling actually land. You might find the $60K to $180K range isn’t hypothetical for your shop, it’s just been sitting there unmeasured.

If you want to talk through what that looks like for your specific accounts before committing to anything, book my Omni Audit and we’ll walk through it together. No deck, no pitch, just your numbers.