You know you need to modernize. You've heard it from your vendors, your accountant, your kid who just graduated with a business degree. The pressure is real. Your competitors are doing something (you're not entirely sure what), and you're starting to wonder if you're falling behind.
Then someone says the magic word: AI.
And suddenly every software salesperson within a hundred miles is at your door promising transformation. Efficiency. Intelligence. The future.
Here's a perspective you don't hear nearly enough: for most small and mid-sized businesses, AI is the wrong answer to the right question.
The question is real. You do have processes that are eating up time. You do have manual work that's prone to errors. You do have bottlenecks that are costing you money. But the assumption that AI is the solution? That's where the market is lying to you.
Let me be direct. Ninety-five percent of enterprise AI pilots fail to deliver measurable ROI. That's not a typo. MIT's 2025 State of AI in Business report found that the vast majority of companies throwing money at generative AI are getting nothing back. Meanwhile, S&P Global found that 42% of companies scrapped most of their AI initiatives in 2025, up from 17% the year before.
Those aren't small companies with small budgets. Those are enterprises with dedicated IT teams and millions to spend. And they're still failing.
Now ask yourself: if companies with those resources can't make AI work, what makes you think the salesperson promising you a plug-and-play solution has cracked the code?
There's a reason AI projects fail at this rate. AI systems require clean data (which most businesses don't have), clear objectives (which most projects don't define), ongoing tuning and oversight (which nobody budgets for), and tolerance for a technology that will confidently tell you something completely false and call it a Tuesday.
That last part isn't a joke. AI hallucinations are a documented, persistent, and unsolved problem. Air Canada's chatbot promised a customer a bereavement discount that didn't exist. A court ruled the airline had to pay. New York City deployed a municipal chatbot that gave citizens advice that was not just wrong but actually illegal. Attorneys have been sanctioned for filing court documents with AI-generated case citations that were entirely fabricated. The AI made them up, formatted them correctly, and presented them with complete confidence.
Now imagine that happening inside your business. Your customer service bot promising something you can't deliver. Your automated reporting pulling numbers from thin air. Your "intelligent" system making a decision at 2AM that you discover at 8AM when the damage is already done.
This is the part where I'm supposed to tell you the alternative is to do nothing. It's not.
The alternative is automation that actually works. Rule-based automation. Workflow automation. RPA (Robotic Process Automation). These aren't new. They're not sexy. And that's exactly why they work.
RPA doesn't learn. It doesn't adapt. It doesn't hallucinate. It does exactly what you tell it to do, every single time, without variation. For the kinds of problems most growing businesses face (invoice processing, data entry, report generation, customer onboarding paperwork, moving information between systems that don't talk to each other), that's not a limitation. That's a feature.
The numbers back this up. McKinsey Digital found that RPA delivers ROI between 30 and 200 percent in the first year. The Institute for RPA reports labor cost savings of 25 to 40 percent. Deloitte's Global RPA Survey found that 78% of companies that implemented RPA scaled their usage within 12 months because the results were that clear.
Compare that to the 95% failure rate for AI pilots.
This isn't about being anti-technology. I've spent 25 years in IT leadership, including eight years as CIO for a global manufacturer. I've evaluated every new technology wave that's come through. Some are real. Some are hype. The difference between the two usually comes down to one question: does it solve a problem you actually have, or does it create new problems while promising to solve hypothetical ones?
For most small and mid-sized businesses, the problems are concrete. You have a process that takes too long. You have a task that someone is doing manually that could be automated. You have systems that don't integrate and require someone to copy data from one screen to another. These are solved problems. The solutions have been around for years. They don't require a data science team or a six-figure implementation budget or a tolerance for your system occasionally making things up.
Here's what this looks like in practice. You have an accounts payable person who spends four hours a day entering invoice data from PDFs into your ERP. An RPA bot can do that same work in minutes, with zero errors, every day, without asking for a raise or calling in sick. You have a sales coordinator copying customer information from your CRM into your quoting system because the two don't integrate. That's a workflow automation problem with a workflow automation solution. You have an accounting team manually reconciling purchase orders, receiving documents, and invoices (the classic three-way match) before approving payment. That's a rule-based process with clear logic: if the numbers match within tolerance, approve; if they don't, flag for review. RPA handles that without breaking a sweat. No machine learning required. No training data. No risk that the system will decide to get creative with the numbers.
The market wants you to believe that if you're not implementing AI, you're falling behind. What's actually happening is that companies are spending money they don't need to spend, on technology they don't need, to solve problems that have simpler solutions.
If you want to modernize, start with the boring stuff. Map your processes. Identify the repetitive, rule-based tasks that eat up your team's time. Implement automation that does exactly what you need it to do and nothing more. Get the ROI. Then, if you have a genuine use case for AI (and there are some), you'll be implementing it from a position of strength rather than desperation.
Or you can chase the hype, throw money at pilots that won't deliver, and wake up in a year wondering why your "intelligent" system just sent a quote to a customer with numbers it invented.
Your call.
If you're running a business between M and M and you're not sure which of your processes are worth automating (or whether AI even makes sense for you right now), that's exactly the kind of question I help business owners answer.
I offer a free strategy session where we look at what you're dealing with, identify the low-hanging fruit, and figure out whether you need AI, rule-based automation, or something more simple as "boring" RPA, or simply just cleaner processes. No pitch. No pressure. Just a straightforward conversation about what actually makes sense for your situation.
Book a Free Strategy Session →
