What I Covered at the AI Workshop

Yesterday I ran "Finally, AI That Makes Sense" - a 90-minute session for Concord Chamber members on what AI actually does for small businesses and how to use it without making your workflow worse.

I'm writing this the day after because a few people asked me to put the main points somewhere they could reference later. This is that post.

If you missed the workshop, this is most of what we covered. If you were there, consider this a written version of the session you can share.

The one thing I kept coming back to

AI is a capable person who knows nothing about your business.

That's the mental model I started with, and I think it's the most useful frame for small business owners who are trying to figure out where this stuff fits. If you hand a capable person a task with no context - no background on your clients, your voice, your industry - they'll produce something generic. Not because they're bad at the job. Because you didn't brief them.

Most AI frustration I hear from small business owners comes from this gap. They try a tool, get a mediocre result, and conclude the tool isn't useful. Often the tool is fine. The setup was missing.

The corollary: when you brief AI well, the results are genuinely good. The work shifts from producing content from scratch to reviewing and refining something that's mostly right. That's a meaningful time savings.

What actually works for small businesses

I focused on three use cases that have a real payoff and hold up in daily use. Not demos - things I've watched work for people running small businesses.

Email drafts. Writing the first version of a professional email is one of the most time-consuming things a small business owner does. It's not hard work - it's just slow. AI handles first drafts well if you give it context: who you're writing to, what the relationship is, what you're trying to accomplish, and roughly what you want to say. You still edit. But editing a good first draft takes a fraction of the time it takes to write from a blank page.

Summarizing and organizing information. If you have a long email thread, a set of meeting notes, or a pile of documents that need to be pulled into one clear summary - AI does this quickly and reliably. I use this constantly. It removes a category of low-skill, high-time work.

Answering questions you'd otherwise search for. Not factual lookups (use a search engine for those). But things like: how do I structure a client proposal? What should I include in a service agreement? How do I approach a difficult conversation with a vendor? AI gives you a starting point that's usually better than what you'd piece together from a search.

The voice problem - and the fix

The most common complaint I hear: "AI doesn't sound like me."

That's correct. Out of the box, it doesn't. It sounds like AI - polished, generic, slightly corporate. If you publish that without editing, people notice.

The fix is straightforward: you have to teach it how you communicate.

I showed the room a simple version of this. Before asking AI to write anything, give it a brief description of how you write. Something like: "I write the way I talk. Short sentences. No jargon. I don't say 'please don't hesitate to reach out.' I say 'feel free to email me.'" That's it. Five minutes of setup that changes every output you get from that point forward.

You can go further - paste in examples of emails you've actually sent, let it identify the patterns, then use those patterns as a style guide. But even the basic version makes a noticeable difference.

What I told people to skip

I spent some time on this because I think it's underserved in most AI content.

There are AI use cases that get enormous coverage because they look impressive in demos. They don't hold up the same way in practice for small businesses.

AI-generated social media posts, published as-is. The quality is generic unless you put real work into the prompts and editing. If you're going to edit heavily anyway, you might as well write it yourself - or at least use AI for the structure and rewrite from there.

AI-generated proposals and reports with no human review. AI can draft these. But the quality that matters in a proposal - specificity, relevance, knowing what the client actually cares about - requires your knowledge of the situation. AI doesn't have that unless you spend the time putting it in. Which is fine, but understand that's the trade you're making.

Chatbots for client communication. For most small businesses, direct personal communication is a competitive advantage. Replacing it with a bot trades something valuable for something mediocre. Automating your responsiveness is worth doing. Automating the relationship is not.

The filter question

I closed the session with the question I apply before adopting any new tool: does this solve a problem I actually have, and does it solve it in less time than it costs to learn and maintain?

For AI, I'd add: does it produce output I'm comfortable putting my name on, with a reasonable amount of editing?

If the answer to both is yes, it's worth keeping. If not, it doesn't matter how impressive the demo was.

What's next

The workshop was for Concord Chamber members, and the interest was high enough that I'm already thinking about what a follow-up session looks like.

If you came yesterday, thank you. If you want to talk through where AI fits specifically in your business - not in general, but in yours - that's a conversation I'm glad to have. Book a free Small Business Efficiency Checkup and we can look at the whole picture, AI included.

If you'd like to know about future workshops, the best way is to follow along here or connect with me through the Concord Chamber. No mailing list, no sequence. Just announcements when something is scheduled.