Five Drivers of Value Creation, One Framework: 80/20 for PE-Backed Companies
80/20 gives lower-middle-market management teams a repeatable way to cut...
Read More
In 1996, Bill Gates wrote in The Road Ahead something that feels more relevant today than it did thirty years ago.
“Automation applied to an efficient operation will magnify the efficiency. Automation applied to an inefficient operation will magnify the inefficiency.”
He wrote that before smartphones existed. Decades before AI. And yet, it describes almost exactly what we’re watching play out across businesses right now.
A 2025 MIT study, The GenAI Divide: State of AI in Business, put a number on it. Despite $40 billion invested in AI initiatives, 95% of businesses have seen zero measurable return. Not modest returns. Zero. MIT researchers found that most companies are pouring AI budgets into sales and marketing. The visible and exciting stuff, while the real ROI sits quietly in back-office operations that nobody’s paying attention to.
We’d take that finding one step further and suggest the problem isn’t just where companies are deploying AI. It’s that many are using it to get faster at things they shouldn’t be doing at all.
There is a framework we use at Strategex called USa — Understand, Simplify, automate. We sit down with management and map every single step in a given process. No shortcuts, no assumptions. Every handoff, every approval, every data entry point, etc., is laid out in full. Then we sort each step into one of three buckets:
Value-add tasks:
Activities your customers are willing to pay for. These create real customer value and are worth doing exceptionally well.
Necessary, but not value-add tasks:
Steps customers would never pay for directly, but that are prerequisites to delivering what they do pay for. Think compliance checks, internal approvals, and status updates. Unglamorous, but essential.
Non-value add:
Tasks that are none of the above. They exist because they always have, or because no one stopped to question them. They create no value, they’re not required, and they’re not a prerequisite for anything that matters.
Here’s where most AI implementations go wrong: companies are spending real time and money on the non-value add bucket because those tasks tend to be easiest to automate. And the ROI? Gates predicted it, and MIT validated it: zero. Congratulations, you’ve just gotten faster at doing something you shouldn’t be doing in the first place. The right answer isn’t automation. It’s elimination.
The better opportunity is in that middle bucket. Necessary, but not value-add tasks, are exactly where AI can make a meaningful difference, but not as an end in itself. Efficiency is not the only goal. It’s the means to something more important.
When you use AI to clear the operational burden of those necessary but not value-add tasks, you free up your best people to spend more time on the work that actually matters, which are the value-add tasks your customers care about and will pay a premium for.
That’s a fundamentally different ambition for AI than shaving cost. It’s about becoming world-class at the things that set you apart.
And here’s why that distinction matters more now than ever: as AI compresses time and lowers the cost of doing almost everything, companies that compete on efficiency alone will face relentless price pressure and margins will erode. The logic is simple. If your edge is that you do things faster or cheaper, then everyone else will too, and eventually AI will erase that advantage entirely.
But companies that use AI to become better (not just faster) at the work customers value most are playing a different game. They’re building the kind of expertise, judgment, and relationships that customers will pay for regardless of what AI does to the broader market. Value-based pricing holds, or improves, precisely because the value being delivered is greater than ever. Cost goes down, quality goes up, and profitability follows.
That’s the real case for AI investment. Not the efficiency win on its own, but what you do with the time and capacity you get back. Simplify first. Automate what’s left. Then pour everything into being world-class at the work that matters.
-
