What You Could Do Versus What You Should Do With AI

A midsize technical services company we worked with raised its close rate by 7% and improved margins by 1.5% using AI. What's interesting is that they never set out to do anything with AI. They set out to fix their sales pipeline.

Sales performance was already a top priority for the leadership team, and the information they needed was scattered across their systems. The ERP held years of customer and transaction history. The CRM tracked opportunities moving through the pipeline. And the sales calls themselves held another layer of insight into what customers were saying and how deals were really progressing, most of which never made it into the structured data. AI let the company bring those signals together and see its pipeline far more accurately. The close rate went up, pricing discounts went down and margins improved across the portfolio.

Those are impacts a CFO doesn't have to search for. They tie directly to revenue and margin, two measures the company was managing long before anyone mentioned AI. And I think that's the most useful observation for any leadership team trying to decide where AI deserves real investment. The best opportunities usually aren't new things you suddenly care about because AI exists. They're new ways to improve things you've cared about for years.

Time saved isn't money saved

Contrast that with where most companies have started, which is employee productivity. Give people an AI assistant and they can draft, summarize, research and analyze in a fraction of the time. Those gains are real for the individual, and better tools are worth providing. But unless that time turns into lower costs, more output or more revenue, it never reaches the income statement. The company has created capacity without capturing value.

The sales pipeline project I mentioned earlier worked differently because it started from the other end. It began with an outcome that already had an owner, a measure and a known economic value, and then asked whether AI could move it. Revenue growth, margin, customer retention, forecast accuracy, working capital, inventory, quality and downtime all fit that description. When you start there, you already know what a meaningful improvement is worth, and the question becomes whether AI can amplify or accelerate value. That's a far easier investment to evaluate than trying to translate individual productivity gains into business value.

The 2027 question

This matters now because most companies are building their 2027 plans, and the list of things AI could do is only getting longer. Models are getting more capable, agents are taking on more complex work, and vendors are building AI into software you already own. Every one of those possibilities will compete for the same money, leadership attention, data resources and employee capacity. Even a healthy AI budget won't cover all of them.

That means making choices. Which opportunities are tied to outcomes that matter strategically or financially? What would have to change for the investment to pay off? What will it take to get there, and what are we choosing not to fund instead? These aren't new questions for CFOs. What's new is the sheer number of AI opportunities competing for investment, which makes the discipline of capital allocation more important than ever.

The companies that get the most out of AI will be the ones that get good at telling the difference between what they could do with AI and what they should do.

About Root Idea

Root Idea helps CFOs protect the business from AI decision risk. We work directly alongside finance teams to map AI influence, establishes decision governance controls that hold up to board scrutiny, and delivers training and change management to make governance stick.

If your organization is scaling AI and governance hasn't kept pace that's exactly the conversation we're built for. Learn more at rootidea.ai.