I’ve spent the last eighteen months inside agencies that are making real workforce decisions around AI. Not the headlines, the actual budget meetings. The truth is simpler and harder than most coverage suggests: AI-related headcount changes aren’t happening because the technology suddenly works. They’re happening because agency economics have been broken for years, and AI gives leadership a reason to finally redesign workflows that were already unsustainable.

If you run a marketing or creative agency, you know the math. Revenue per employee has been flat or declining since 2019. Client budgets haven’t kept pace with the volume of asks. Your account managers are spending 40% of their time on reporting instead of strategy, and your content team is buried in requests that all pay the same flat rate regardless of complexity. The typical agency in the $2M-10M range is leaking $60K-$180K annually to work that should be automated but isn’t, because until recently the cost of custom automation was higher than the cost of hiring another mid-level person.

That calculation just flipped. And the agencies making moves right now aren’t replacing people with AI because the AI is good. They’re redesigning how work flows through the business, then using AI to handle the parts that don’t require judgment. The difference matters, because if you approach this as a technology problem you’ll make expensive mistakes. If you approach it as an economics and workflow problem, you’ll find margin you didn’t know you had.

The Real Driver: Margin Pressure and Workflow Bloat

Let’s talk about what’s actually happening in your agency right now. Your account managers are capped at six to ten accounts each, not because they lack skill but because each account generates 15-20 hours of non-billable work every month. Monthly reports, performance decks, client emails, internal status updates, the endless Slack threads to pull numbers from five different platforms. One agency partner told me his AMs were spending two full days per account per month on reporting alone, which meant the only way to grow revenue was to hire more AMs and accept the margin hit.

Your content team is in the same trap. The volume of requests has doubled in three years. Clients want more assets, more formats, more channels. But the per-asset rate hasn’t moved, and your team is starting every piece from a blank page. The cost per piece is going up because your senior people are doing work that doesn’t require their judgment, and you can’t hire junior people fast enough to absorb the overflow without tanking quality.

This is the part most AI coverage misses. The layoffs and restructures you’re reading about aren’t happening because ChatGPT can suddenly write better copy than your team. They’re happening because leadership finally has a tool that lets them separate the judgment work from the production work, and the economics of that separation are too compelling to ignore. If you can get your AMs back to strategy and client relationships, and your content team back to creative direction instead of first-draft production, you don’t need to hire three more people to hit your growth target. You might not need to hire anyone.

What Workflow Redesign Actually Looks Like

Here’s what I see working in agencies that are doing this well. They start by mapping every repeating task in the business, then they ask two questions: does this task require human judgment, and does this task directly generate revenue or client value? If the answer to both is no, it’s a candidate for an agent.

Take monthly reporting. Your AM pulls data from Google Ads, Meta, LinkedIn, your CRM, and whatever analytics platform the client uses. They drop it into a template, write a summary, draft an email, and schedule a call to walk through it. The entire process takes four to six hours per account, and 90% of it is data assembly and formatting. The judgment part is the three-paragraph summary and the recommendations section. Everything else is production work.

A Reporting Agent handles the production work. It connects to every platform, pulls the data on a schedule, populates the template, and drafts the email summary in your AM’s voice. Your AM reviews it, edits the recommendations, and sends it. The four-to-six-hour task becomes a 30-minute review. Your AM just got 15 hours back per month, which means they can handle eight or nine accounts instead of six. You didn’t replace the AM. You redesigned the workflow so the AM does the work only they can do.

The same logic applies to content production. Your team gets a brief, researches the topic, writes the first draft, revises it, formats it, and hands it off. The creative judgment happens in the revision and the strategic choices about tone and angle. The first draft is production work. A Content Production Agent takes the brief, pulls relevant background from your knowledge base, and produces a first pass that’s on-brand and on-format. Your writer edits instead of starting from zero. A piece that used to take three hours now takes 90 minutes, and the quality is higher because your writer is spending their time on the parts that matter.

I’m not describing theory. I’m describing what we’ve built for agencies in the last twelve months through the AI audit for marketing and creative agencies. The workflow redesign comes first. The agent comes second. If you skip the redesign and just bolt AI onto your existing process, you’ll automate waste and wonder why your margin didn’t move.

The Tasks That Benefit from AI vs. the Tasks That Don’t

This is where most agencies make their first mistake. They assume AI is good at creative work and bad at repetitive work, because that’s the narrative in the press. The reality is the opposite. AI is excellent at high-volume production work with clear templates and rules. It’s terrible at judgment calls that require context, client history, and strategic intuition.

Your AM knows when a client is about to churn because they’ve been in the relationship for 18 months and they can read the tone shift in emails. An agent can’t do that. But an Account Health Agent can watch the data, flag when engagement drops or when a campaign underperforms, and draft the message your AM should send. The judgment about whether to send it and how to frame the conversation stays with the AM. The monitoring and the first-pass drafting moves to the agent.

Your senior copywriter knows when a piece needs a complete rewrite because the angle is wrong. An agent can’t make that call. But the agent can produce the first draft so your writer isn’t staring at a blank page. The creative direction and the revision stay human. The production work moves to the agent.

The agencies that are restructuring successfully are the ones that have done this task-by-task audit. They’re not asking “what can AI do?” They’re asking “which parts of this workflow require judgment, and which parts are just production?” Then they’re moving the production work to agents and redeploying their people to the judgment work. The result isn’t fewer people. It’s higher output per person, which means better margin and more capacity to grow without hiring.

One trades-adjacent agency in our network described the shift this way: “We used to think we needed to hire two more AMs to hit our growth target. After we built the reporting and account health agents, we realized our existing AMs could each handle 40% more accounts. We didn’t hire. We gave everyone a raise and kept the margin.” That’s the economics of workflow redesign.

Why the Audit Comes Before the Build

If you’re reading this and thinking “I need to figure out which tasks in my agency are agent-ready,” you’re asking the right question. The next step isn’t to start building agents. The next step is to map the work.

We run a 60-minute Omni Audit for agencies that walks through exactly this process. You bring your workflows, we map the repeating tasks, and we identify the three to five highest-value opportunities for agents in your business. You leave with a priority list, a rough ROI estimate for each agent, and a build roadmap. No deck, no sales pitch. Just the map.

The reason the audit comes first is that most agencies don’t have a clear picture of where their time is going. Your AMs know they’re busy, but they can’t tell you how many hours per month they spend on reporting versus strategy. Your content team knows they’re underwater, but they don’t have data on how long each asset type takes or which clients generate the most revision cycles. Without that map, you’re guessing. And if you’re guessing, you’ll build the wrong agents and waste six months on automation that doesn’t move margin.

The agencies that get value from AI are the ones that start with the audit. They identify the specific tasks that are eating margin, they quantify the time cost, and they build agents that target those tasks. Then they measure the result and iterate. Book a 60-min Omni Audit and we’ll walk through your workflows in detail.

The Three Agents Most Agencies Build First

Once you’ve done the audit, the build priority usually shakes out the same way. Agencies almost always start with one of three agents, because these are the tasks that hit every agency and the ROI is immediate.

The Reporting Agent is the most common first build. It connects to your ad platforms, analytics tools, and CRM, pulls performance data on a schedule, and drafts the monthly report and email summary. Your AM reviews and edits, but the data assembly and formatting are handled. For an agency with ten accounts, this agent saves 40-60 hours per month. At a blended AM rate of $75/hour, that’s $3K-$4.5K in recovered capacity every month. The agent pays for itself in the first billing cycle.

The Content Production Agent is the second most common build. It takes a brief, pulls relevant context from your knowledge base, and produces a first-pass draft in your brand voice and format. Your writer edits instead of starting from scratch. For an agency producing 50-80 pieces per month, this agent cuts production time by 30-40%, which means your team can handle more volume without adding headcount. One agency we worked with went from three content writers to the same three writers handling 60% more output, which let them take on two new retainer clients without hiring.

The Account Health Agent is the third build. It monitors client accounts daily, flags risk signals like engagement drops or budget pacing issues, and drafts the proactive message your AM should send. This agent doesn’t replace the AM’s judgment about how to handle the situation. It replaces the manual monitoring and the “what do I say?” moment when the AM realizes something is off. For agencies with 15-30 accounts, this agent typically catches two to three issues per month that would have otherwise slipped through. The revenue protection alone justifies the build.

These three agents show up in almost every agency audit because they target the tasks that are universal, high-volume, and easy to template. If you’re not sure where to start, start with one of these. You’ll see ROI in the first 60 days, and you’ll have a working model for how to build the next agent.

What This Means for Your Workforce Planning

Let’s get specific about what this means for your hiring and restructuring decisions over the next 12 months. If you’re planning to grow revenue by 30% this year, the old model says you need to hire at least two more AMs and probably another content person. The new model says you need to build two or three agents and redeploy your existing team to higher-leverage work.

The math is straightforward. If your AMs are capped at six accounts each and you want to add 12 new accounts, you need two more AMs at $80K-$100K each, plus benefits and overhead. That’s $200K-$250K in new cost to generate the revenue. If you build a Reporting Agent and an Account Health Agent, your existing AMs can each handle eight to nine accounts. You add the 12 new accounts without hiring, and your margin on that new revenue is 40-50% higher because you didn’t add the headcount cost.

This is the calculation that’s driving the workforce changes you’re reading about. It’s not that AI is replacing people. It’s that the economics of growth have shifted, and agencies that redesign their workflows around agents can grow faster and more profitably than agencies that keep hiring. If you’re still planning headcount the old way, you’re going to lose margin to competitors who figured this out six months ago.

The agencies that are handling this well are being transparent with their teams about what’s changing and why. They’re not cutting people. They’re retraining people to work with agents, and they’re redeploying capacity to higher-value work. Your senior AM who used to spend 40% of their time on reporting is now spending that time on strategy and upsells. Your content team is producing more and better work because they’re editing instead of drafting from scratch. The people who adapt to this model are more valuable, not less. The people who resist it are going to struggle.

The Missteps to Avoid

I’ve also seen agencies make three common mistakes when they start building agents, and all three are expensive.

The first mistake is building an agent without redesigning the workflow first. You take your existing reporting process, which is a mess of manual data pulls and inconsistent templates, and you try to automate it. The agent inherits all the inefficiency and adds a layer of complexity. You spend three months building it, and when you turn it on, your AMs don’t trust it because the output is inconsistent. You wasted the build budget and you still have the original problem. The workflow redesign has to come first. Clean up the process, standardize the templates, then build the agent to execute the clean process.

The second mistake is assuming the agent can handle judgment calls. You build a Content Production Agent and you expect it to write final-draft copy that goes straight to the client. It can’t. It’s going to miss tone, it’s going to make factual errors, and it’s going to produce generic output that doesn’t match your brand. Then you lose client trust and you pull the agent offline. The right model is agent-produces-human-edits. The agent handles the first draft, your writer handles the creative judgment and revision. If you set that expectation from the start, the agent works. If you expect the agent to replace the writer, you’re going to be disappointed.

The third mistake is building agents in isolation without connecting them to your existing tools. You build a Reporting Agent that lives in a separate dashboard, and your AMs have to log into a new system to review the output. They don’t. They keep doing it the old way because the new way adds friction. Agents have to integrate into the tools your team already uses. The reporting output should land in Slack or email. The content draft should appear in Google Docs or your CMS. If you add a new tool to the stack, adoption will be slow and the ROI will take longer to show up.

These mistakes are avoidable if you start with the audit and you design the workflow before you build the agent. That’s why we structured Omni for marketing and creative agencies the way we did. The audit identifies the right tasks, the workflow redesign removes the inefficiency, and the agent build targets the clean process. You avoid the missteps and you get to ROI faster.

The Next 60 Minutes

If you’re still reading, you’re probably thinking about which tasks in your agency are eating the most margin and whether agents can help you recover it. The next step is to map it. You can do this internally, but most agencies don’t have the time or the objectivity to do it well. You’re too close to the work.

The Omni Audit takes 60 minutes. You walk me through your workflows, I ask questions about time cost and repetition, and we build the priority list together. You leave with three things: a map of the highest-value agent opportunities in your business, a rough ROI estimate for each one, and a build roadmap that shows you what to do first. No deck, no sales pitch. Just the map and the next steps.

We’ve run this audit with 40+ agencies in the last year, and the pattern is consistent. The tasks that are killing your margin are the ones you’ve stopped noticing because they’re just “how we do things.” The audit makes them visible, and once they’re visible, the decision to build agents is obvious. Book my Omni Audit and we’ll map your workflows in detail.

The workforce changes you’re reading about aren’t coming from some sudden AI breakthrough. They’re coming from agencies that finally have a tool that lets them separate judgment work from production work, and the economics of that separation are too good to ignore. If you’re still planning growth the old way, you’re leaving margin on the table. If you’re ready to redesign the workflows and build the agents, the next 60 minutes will show you exactly where to start.

You can also explore more about how we approach AI strategy and implementation or dive into Omni Ops to see how agents integrate into your existing stack. The tools exist. The question is whether you’re going to use them to redesign your business or whether you’re going to keep hiring your way to growth and watch your margin compress. The agencies that figure this out in the next six months are going to have a structural advantage that’s hard to catch. The audit is the first step.