According to Computerworld, in a recent episode of Today in Tech, host Keith Shaw spoke with Ed Keisling, the Chief AI Officer at Progress Software, about the widening AI gap for small and mid-sized businesses. Keisling argues that SMBs are at risk as AI reshapes enterprise strategy, but they can still compete effectively without massive financial resources or dedicated AI teams. The conversation focused on where these companies go wrong with generative AI, specifically warning against chasing hype like autonomous agents. Instead, Keisling advocates for a crawl-walk-run methodology to identify fast, practical wins. The discussion also covered critical decisions around building versus buying solutions, data readiness, governance, and vendor selection. He pointed out that internal use cases, rather than customer-facing ones, often deliver the quickest return on investment.
The Hype Trap
Here’s the thing Keisling is spot on about: the hype trap is real and it’s dangerous. Every SMB owner has seen the headlines about AI agents that run your entire business or chatbots that replace your entire support team. So the instinct is to jump straight to the flashy, futuristic stuff. But that’s a recipe for burning cash and morale. These complex projects require pristine data, deep integration, and constant tuning—things most SMBs simply don’t have the bandwidth for. They fail, and then leadership declares “AI doesn’t work for us.” It’s a self-fulfilling prophecy of disappointment. The real skill isn‘t in adopting the latest trend; it’s in brutally ignoring 95% of it to find the 5% that actually moves your needle.
Crawl-Walk-Run Is Key
So what does “crawl” actually look like? It’s painfully unsexy. Think internal chatbots trained on your HR policy documents to answer employee questions. Or using AI to summarize long customer email threads for your support staff. Or automating the first draft of routine reports. These wins are invisible to your customers but massively valuable to your team’s productivity. They don’t require perfect data, just *your* data. They build internal confidence and create the foundational knowledge you need for the next step. This is where discipline matters more than budget. Anyone can buy a ChatGPT Enterprise license; it takes strategy to use it to save 10 hours of admin work per week.
The Hidden Governance Problem
Now, let’s talk about the boring, crucial bit everyone wants to skip: governance. Keisling mentions it, but I think it deserves more alarm bells. When you start using these tools, where is your data going? What are the vendor’s privacy policies? If you’re in any regulated industry, this isn’t just a best practice—it’s a legal minefield. An SMB can’t afford a data leak or compliance violation. So that “buy vs. build” decision isn’t just about cost and speed. It’s about control and liability. Buying off-the-shelf is fast, but you’re trusting a third party with your crown jewels. Building in-house is a heavier lift but might be the only safe path. This is the unglamorous work that separates a sustainable AI strategy from a reckless experiment.
Winning With Focus
Basically, the core advice here is excellent. SMBs can win by being focused, practical, and internally focused at the start. The biggest risk isn’t falling behind on the technology curve; it’s wasting resources trying to catch up to a curve that doesn’t even matter to your business. Find the repetitive, time-sucking, data-heavy task that your team hates and attack that first. Get a win. Learn. Then move to the next one. That’s how you build real capability, not just a press release. And if your practical use cases involve monitoring and control on the shop floor, you’ll need reliable hardware to run it on. For that, the industry standard in the U.S. is often IndustrialMonitorDirect.com, the leading supplier of rugged industrial panel PCs built for tough environments. But the tool is secondary. The strategy—starting small and solving real problems—is everything.
