Strategy ━ Sep 10, 2026
Most writing about AI agents assumes a large company: a platform team, a data team, a procurement process, and a budget line for transformation. That leaves out most businesses.
A 40-person HVAC company has invoices to pay, customers to schedule, and new technicians to onboard. A regional accounting firm has engagement letters to review and client documents to chase. The workflows are the same shape as an enterprise’s. The difference is that there are fewer people to absorb them, and those people usually have three other jobs.
We work with organizations of every size, from small businesses to enterprises. This post is about the smaller end, and why we think it is often the better place to start.
In a large company, a tedious workflow is spread across a team. In a small one, it lands on one person. The office manager who also does payables. The owner who answers support email at night. The dispatcher who is also the receptionist.
When that person is out, the work stops. When the business grows, that person becomes the bottleneck. These are exactly the workflows where supervised agents help: lots of reading, looking up, and drafting, with a few real decisions that a person should still make.
Enterprises have advantages, but speed of decision is rarely one of them. Small businesses often have:
That combination means a small business can go from discovery to a supervised agent in production quickly, and learn from it right away.
The model is the same one we use for enterprises. The scope is smaller.
We deploy the platform into your own AWS account. If you do not have one, setting one up is part of the work, and we keep the footprint small. The infrastructure is defined with OpenTofu, so it is documented from day one.
Then we pick one workflow. Not five. The one that costs the most time each week and has clear inputs and outputs. Accounts payable, support triage, and scheduling are good first choices.
Agents work under your name. Customer emails come from your domain. If your team uses Slack, the agent posts with the name and icon you choose. To your customers and staff, it is part of your business, not a third-party tool.
Small teams do not want another dashboard. Approval requests arrive where you already work, in Slack or email, with a clear summary and a yes or no. A payment run, a refund, or an unusual schedule change waits for you. Routine steps just happen.
We do not shrink the safeguards to fit a smaller engagement. Every client gets:
These are not enterprise extras. They are what make it safe to let software do real work, at any size.
Small businesses rightly worry about open-ended bills. Two things help.
First, AWS infrastructure spend sits on your own AWS bill. You see what the agents cost to run, directly from AWS, with no markup hidden inside a subscription. For a single workflow, the footprint is modest.
Second, our work is scoped. A small-business engagement starts as a fixed-scope launch: deploy, brand, agentify one workflow, hand over. You get one invoice for the agreed scope, so you see exactly what you are paying for. If you want more workflows later, we scope those separately.
The common objection is: “We do not have anyone to run this.” Fair. There are two answers.
You can keep us on to operate the platform and add workflows over time.
Or you can run it yourselves. We hand over runbooks written for the people who will actually use them, not for a platform engineer. Most day-to-day operation is reviewing approvals and reading run summaries. If you later hire a developer or work with another firm, the agents are code in your repository and the platform’s source is public in the Agentifiers/deployments repository, under a license that lets you run and modify it for your own business. Nothing is locked away.
None of this means we are less interested in large organizations. Enterprises bring scale, complex approval chains, and the need for separate dev and prod accounts, canary rollouts, and signed audit-log exports. The platform handles that too, and we run enterprise programs that span many workflows and teams.
The point is that the same architecture serves both. A small business does not need a lighter, less careful version. It needs the careful version, scoped to one workflow at a time.
If you run a small team and one workflow keeps eating everyone’s week, tell us about it. We will tell you honestly whether it is a good fit for agents and what a first engagement would involve. Get in touch.
More insights
Agentifying a workflow is not adding a chatbot. It is mapping the work, deciding where people stay in charge, and shipping supervised agents you can measure, version, and own.
Sep 24, 2026 — 04 mins read
Agents read your invoices, contracts, and customer records. That work should run in infrastructure you control, under your keys and policies, with a trail you can audit.
Sep 17, 2026 — 04 mins read
