We've been in enough conversations lately to notice a pattern. An industrial company decides to modernize a function: HR, procurement, sales operations, whichever one feels the most overdue, and the plan is to buy an AI tool that handles it. That way, the company is “using AI.” A few months later, the tool is live, the invoice is paid, but the return isn't there. The people who are responsible for doing the work are frustrated because the tool doesn't understand their business the way they did. Leadership is frustrated because the thing they bought hasn't produced what it promised. The read most companies take from that experience is that the technology wasn't ready, or wasn't the right vendor, or needs another six months of configuration. Is this sounding familiar?
The read that it is the technology’s fault is wrong, and it's an expensive place to keep landing. If there is only one thing you take from this article, it is that AI should not be a thing you buy and install the way you'd buy a new machine for the plant floor. It's a capability, and capability isn't something a vendor can hand you finished. AI should be built, inside the organization, by people who understand the work well enough to direct AI correctly. A bolt-on tool arrives with none of that understanding. It is built by software teams who can ask the AI to build in what it thinks good HR practices are (as an example), but it doesn’t have the subject matter expertise on your business or your industry to hold up in practice. You have two options when it comes to AI adoption at scale in your organization:
To buy it as a bolt on piece of technology, and it will fail almost every time… remember your CRM implementation?
To take the subject matter experts in your company and teach them how to use AI. It is far easier to take a subject matter expert with 20 years of experience and teach them how to use AI. The basics are not complicated, incredibly powerful, and easy to teach. This approach is where you see the return on the LLM models you are paying for.
The fix isn't a better vendor. It's a different starting question. Instead of asking which tool solves the HR problem, or the sales ops problem, the more useful question is what it would take for your own team to solve this themselves, with AI doing the heavy lifting. That's a smaller, less flashy ask than replacing the function, and it's the one that compounds. A team that builds the skill to direct AI toward a real problem keeps that skill after the project ends. A team that buys a tool owns a subscription.
We've watched this play out with many of our clients (several coming to us because they are frustrated that their bolt on tool is 80% implemented and failing). We recently walked you through a case study of a pre-call and post-call AI native process. The pre-call and post-call tools we described didn't arrive as a finished product someone tried to bolt on. They got built around a specific sales methodology, tuned against real calls, and adjusted based on what worked and what didn't. That's capability, not technology, and it's the reason the result holds up once the novelty wears off. Better than even that, the capability allows for continuous improvement in a way that a tool simply can't.
Buying AI technology treats the symptom. Building AI capability treats the actual problem, and it's the only version of this that compounds instead of decaying.
If this was useful, the rest of the Commercial Intelligence Report is one click away. Subscribe to the Commercial Intelligence Report, a weekly analysis on how industrial commercial teams are turning AI into measurable ROI, and the discipline that makes it stick.