May 17, 2026 · 3 min read
How I'd think about AI if I owned a 10-person business
An anchor post. Who's behind this notebook, three patterns I see going wrong in small-business AI adoption, the three questions I'd actually start with, and what's coming.
1 · Who I am and why this exists
For the last fifteen years my day job has been senior technology strategist across several different industries. While technical in nature, the role I was in needed to be in tight alignment with the business. In plain English: I'm the person they brought in to figure out which tools to buy, what to build, and how to keep it all working together.
That work involved a lot of vendor evaluation. A lot of integration planning. A lot of telling executives that the expensive thing they wanted wasn't going to do what the sales deck promised, and occasionally a lot of telling them that the cheap, boring solution was the right one.
I'm starting Runari because AI is moving fast, the marketing is loud, and small business owners are getting hit with sales pitches every week. They deserve someone who's done a few rounds of this at scale to help them think through it without an agency markup on the other end.
The day job stays. This is a notebook I happen to publish, plus a small advisory practice on the side. No team, no sales department, no pressure to upsell.
2 · Three patterns I see going wrong
I want to start by being concrete about what I see small businesses doing that wastes their money and time. Three patterns keep showing up.
Pattern A: "What can AI do?" is the wrong starting question.
Owners hear about AI, get excited (or anxious), and start trying to figure out what they could do with it. That's backwards. The right question is: what's eating my time or money right now? Then ask whether AI is the cheapest, dumbest way to fix that specific thing. Often it isn't.
Pattern B: Tool-first thinking.
Owner hears about ChatGPT, or Claude, or n8n, and tries to retrofit a use case onto the tool. The vendor's marketing is calibrated to make this happen. Start with the problem, pick the tool last. Better still: start with the problem, then ask whether a tool is even the right answer (sometimes the answer is "hire a part-time employee for four hours a week").
Pattern C: Buy everything, integrate nothing.
This one's the most expensive. Small businesses end up paying for eight SaaS subscriptions that don't talk to each other. The CRM doesn't know about the accounting tool. The calendar doesn't know about the intake form. The support inbox doesn't know about the customer history.
The wins in small-business automation are usually in the connections, not the new purchases. Quote → invoice → accounting. Intake → calendar → reminder. Support inbox → AI triage → human if needed. Connect what you already have and you'll often get more value than from any new $200/mo subscription.
3 · How I'd actually think about it
If I owned a 10-person business and someone asked me where to start with AI, here are the three questions I'd answer first.
1. What specifically is eating your time or money this week? Be specific. "Sales is hard" doesn't count. "I'm spending six hours every weekend writing follow-up emails to leads from last week's open house", that's specific. That's actionable.
2. What's the cheapest, dumbest, most boring way to fix that? Often this isn't AI. Sometimes it's a spreadsheet template. Sometimes it's hiring a part-time virtual assistant. Sometimes it's a $20/mo SaaS tool that's been around for ten years and just works. Start with the cheapest, dumbest answer. AI comes in only when it beats the dumb solution materially.
3. If AI is part of the answer, what does "good enough" look like in 30 days? You need a definition of success before you spend the money. Otherwise you'll either pay forever for a thing you can't measure, or you'll cancel something that was actually working. "Good enough in 30 days" forces a concrete bar.
The boring tools beat the exciting tools 80% of the time. The boring tools are also where senior judgment matters most, because nobody's marketing department is pushing them.
4 · How to reach me
If something here speaks to a specific situation you're trying to solve, I take a small number of advisory projects each year. Three shapes: a vendor & tool review, a strategy & roadmap, or implementation work. Each one is flat-fee and scoped in writing before any work starts.
If you just want to read along, the homepage will tell you when something new is up.
No ads. No course. No "scaling content." One person writing what he actually thinks.
Runari