Two years ago, if you'd told me that a single person could run data builds for a seal manufacturer, email campaigns for a print provider, telemarketing oversight for a golf networking business, live dashboards for a software company, and content production for half a dozen clients simultaneously, I'd have said you were describing a nervous breakdown, not a business model.
And yet here we are.
AA2 is one person. Me. One founder, no employees, no contractors sitting in a serviced office pretending to be a team. The work gets done, the clients get results, and nobody has to sit through a status meeting that could have been an email. The secret isn't working 18-hour days. It's building a tech stack that does the heavy lifting while I do the thinking.
This is the most honest piece I've written for this column. No client references to hide behind. Just the reality of running a one-person agency in 2026 and what made it possible.
Every morning starts the same way. Coffee, then a glance at the monitoring dashboard. Overnight, automated workflows have already done things that would take a junior marketer half a day: verified new prospect data, sent scheduled email sequences, updated CRM pipelines, flagged any campaign anomalies, and compiled engagement reports.
The backbone is n8n, an open-source workflow automation platform running on a VPS I manage myself. It connects everything. When a prospect opens an email, that event ripples through the system: updating their engagement score, adjusting their position in the pipeline, potentially triggering a follow-up action. When a new contact gets added to a data build, automated verification checks the email, confirms the company details, and scores the lead before it ever reaches a human eye.
Then there's the AI layer. I use AI for data research, email draft generation, content outlining, report compilation, and dashboard automation. Every piece of AI output gets reviewed by me before it goes anywhere near a client. The AI proposes. I decide.
The CRM ties it all together. Client sub-accounts, each with their own pipelines, automations, and reporting. I can see every client's entire marketing operation from a single screen.
And then there's the content production pipeline. Blog posts, email campaigns, social content, white papers. AI generates first drafts based on detailed briefs I've built for each client. I rewrite, refine, and approve. The output is human. The speed is not. A monthly email campaign that would take a copywriter two days to research and draft takes me half a morning, because the research and structural work is already done before I sit down to write.
None of this is theoretical. Every component is live, running, and producing results for real clients right now. The stack evolves constantly. Something that took me three hours six months ago takes forty-five minutes today, because I've automated another step or refined another workflow. That compounding efficiency is the real engine of the model.
There's no typical day, which is sort of the point. But here's a recent Monday, unedited.
07:00. Health check email arrives. All 40-odd automated workflows ran overnight without errors. One client's email campaign hit a 48% click rate. Another client's data enrichment batch verified 340 contacts. All good.
08:00. Strategy call with a warehouse management software company. Thirty minutes reviewing their campaign performance, adjusting targeting for next month, discussing a new sector they want to enter. This is the part AI can't do. Listening to a client describe their market, picking up on the nuances, translating that into a plan.
09:00. Write two email campaigns. AI generates the first drafts based on the client brief, previous campaign performance, and the tone guidelines I've built for each client. I rewrite roughly 40% of it. Sometimes more. The AI gets the structure right but misses the subtlety. A phrase that sounds fine in isolation reads wrong when you know the client's audience.
11:00. Data build review. A prospect list for a new client has come back from the enrichment pipeline. 600 contacts, each with verified emails, LinkedIn profiles, company details, and engagement scores. Two years ago, this would have taken a researcher three weeks. The automated pipeline did it in four hours. My job is to check the quality, remove any obvious mismatches, and approve it for campaign use.
12:30. Lunch. Actually eating lunch, not working through it. This matters more than people think.
13:00. Dashboard updates. Two clients need their monthly reporting refreshed. The data pulls automatically from the CRM and email platform. I add context, write the narrative, highlight what's working and what isn't. Clients don't want a spreadsheet. They want someone to tell them what the spreadsheet means.
14:30. Telemarketing review. I don't make the calls myself. I work with a calling partner who's been doing director-level B2B telemarketing for nine years. My job is providing the intelligence: who to call, what they've engaged with, what to say. The caller does the rest. We debrief weekly.
15:30. Content production. Blog posts, social content, LinkedIn articles. Some for clients, some for AA2. Again, AI handles the first pass. I handle the voice, the specifics, the bits that make it sound like a human wrote it. Because a human did.
17:00. Admin. Invoices, proposals, the unglamorous machinery of running a business. Even this is partially automated. Xero handles the accounting. Automated workflows handle the invoice generation and payment tracking.
That's one person's day. Six clients touched, campaigns running, data verified, content produced, strategy delivered. No team. No office. No overhead.
Here's where I'm supposed to say "and it all works perfectly." It doesn't. There are genuine limitations to this model, and pretending otherwise would be dishonest.
Capacity has a ceiling. I can serve around eight to ten active clients well. Beyond that, the strategy work suffers because there aren't enough hours for deep thinking. I've turned down work rather than dilute quality, which is the right call but doesn't feel great at the time.
Illness is a real risk. If I'm out for a week, automation keeps the campaigns running, but nobody's steering. I've built enough redundancy that a short absence doesn't cause problems, but a long one would. That's the trade-off of a one-person operation.
Some clients want a team. They want to see three people on a call, want to know there's a project manager and a designer and a strategist all working on their account. I'm honest about what AA2 is. Some prospects hear "one person with AI" and see risk. Others hear it and see efficiency. The right clients find me. The wrong ones don't, and that's fine.
And sometimes I get it wrong. Last year I took on too many projects in the same month and the quality of my strategic thinking suffered. Not the execution, the automation handles that reliably, but the part that requires me to sit and think about a client's market for an uninterrupted hour. I course-corrected, but it was a reminder that automation doesn't eliminate human constraints. It just moves them to a different part of the process.
The counterintuitive truth is that a one-person agency with the right technology can often deliver better results than a small team.
No handoff errors. When information passes between team members, things get lost. The brief gets summarised. The nuance disappears. The email that goes out doesn't quite match what the client described. In a one-person operation, the person who takes the brief is the person who writes the copy is the person who builds the campaign is the person who reviews the results. Nothing falls between chairs because there's only one chair.
Speed. There's no waiting for a designer to finish another project. No scheduling a meeting to align on next steps. If a client calls at 10am with a campaign idea, the first draft can be in their inbox by lunchtime. Not because I'm rushing. Because there's nobody to wait for.
Consistency. Every client gets the same quality of thinking because it's all coming from one brain. There's no junior account exec handling Tuesday's emails while the senior strategist is in another meeting. The person the client hired is the person doing the work. Every time.
Accountability. This one's underrated. When something goes wrong, and things do go wrong, there's no finger-pointing between departments. No blaming the data team or the creative team or the account team. It's me. That level of direct accountability changes how carefully you approach every piece of work. There's no hiding in a team when you are the team.
I've spent 800 words talking about tools and automation, and I want to be clear: the technology isn't the point. The point is what the technology makes possible, which is a business model that simply didn't exist five years ago.
Before AI and automation matured to this level, one person could be a freelancer or a consultant. You could advise, or you could execute, but doing both at scale was physically impossible. The tech stack changes that equation. It doesn't replace the strategist. It gives the strategist the capacity to also be the executor.
I didn't set out to build a one-person agency. It happened because the tools caught up with the ambition. Three years ago, I would have needed a data researcher, a copywriter, a CRM administrator, and a campaign manager to deliver what I deliver today. Those roles haven't disappeared. They've been absorbed into a workflow where technology handles the repeatable parts and I handle the parts that require a brain.
The model isn't for everyone. It requires a willingness to learn technical systems, a comfort with working alone, and an honest assessment of your own limitations. But for the right person with the right stack, it changes what's possible. Completely.
Over the past two years of writing this column, most of what I've covered comes back to one idea: marketing works when the thinking is good and the execution is consistent. The one-person agency model doesn't change that principle. It just proves you don't need a payroll to deliver on it.
If you're a business owner reading this and thinking "that sounds too good to be true," I'd understand the scepticism. The best thing I can offer is the results. The campaigns are real. The data builds are real. The dashboards are live. And the person behind all of it is one bloke in Wiltshire with a decent internet connection and a very long to-do list.
Martin Dugan, AA2
