Why isn’t your AI investment showing up in the numbers?

Most leadership teams can tell the same story about AI. The pilots are done, he tool is in people’s hands, but the spend line keeps rising. Then someone in the board meeting asks the obvious question: where is it in the numbers?

For a lot of businesses, the honest answer is that it isn’t there yet, and new research suggests that waiting longer won’t fix it on its own.

The adoption plateau

In Harvard Business Review this month, researchers from Boston Consulting Group looked at more than 600 US public companies. Roughly three-quarters have invested in AI and moved beyond pilots. Yet they haven’t earned a shareholder-return premium over their industry peers. The authors call this the adoption plateau. Adding more tools and widening deployment doesn’t get you off it.

The companies pulling ahead are doing something different. They pick one of three paths to value:

1. Doing existing work more efficiently
2. Improving current products
3. Creating new offerings

Then they build the capability that path needs. That might be redesigned workflows, broad AI fluency across the workforce, or specialist talent. The point is that the strategy and the capability have to match.

Not more AI. Not faster AI. AI pointed at a clear outcome, with the right foundations underneath it.

What the three paths look like in planning

In Anaplan, each path already has a practical starting point:

• Doing existing work more efficiently. CoModeler turns build requests into plans and flags model inefficiencies, while PlanIQ produces a faster forecast baseline. We covered both earlier in this series, in building with AI agents and explaining AI forecasts.
• Improving what you already deliver. Sharper forecasts and earlier anomaly spotting give the business a better plan, sooner, so commercial and supply chain teams can respond before the quarter is lost.
• Creating something new. Fast scenario modelling lets you test a new market, product or pricing model before committing capital to it.

The tools are there. What decides whether they pay back is what sits underneath them.

Why planning is where the plateau shows first

All three paths run through planning. Doing work more efficiently means faster forecasts and fewer manual reconciliations. Improving products means sharper demand signals. Creating new offerings means testing scenarios before you commit capital.

That makes planning the place where AI either starts paying back or quietly stalls, and when it stalls, the cause is rarely the model or the maths. It’s usually one of three things:

• Data quality. AI can’t improve a forecast built on data nobody trusts.
• Governance. AI speeds up decisions, and the wrong governance speeds up the wrong decisions.
• Operating model. If nobody owns the output, nobody acts on it.

None of these is a failure of ambition or talent. They’re the foundations most planning teams were never asked to build, because until now they didn’t need to.

Three questions to ask before your next AI investment

Before the next AI investment, it’s worth putting these to your planning and finance leads:

1. Which of the three paths are we actually on? If the answer is “all of them”, it’s probably none of them.
2. Would we trust an AI-generated forecast enough to act on it tomorrow? If not, find out whether the problem is the data, the process or who signs it off.
3. When AI changes a number, who explains it to the board? If nobody can answer that yet, governance comes before the next tool.

Our view: choose the path, then build the foundations

The adoption plateau isn’t a technology problem. It’s what happens when AI is added everywhere and pointed at nothing in particular. The way off it is smaller than most AI strategies suggest. Pick one path. Start with one planning cycle. Fix the data, the governance and the ownership in that area first, then let AI accelerate it.

We’ve seen the same pattern: the biggest risk is never the maths. It’s whether people trust the output enough to act on it. As one of 12 Anaplan preferred AI partners, that’s the gap we help planning teams close.

AI accelerates, foundations decide, and people stay accountable.

What to read and do next

Join the webinar: 24 November

Horváth and Bedford Consulting will cover where AI helps planning today, what agentic AI means for finance, and the steps to get ready. It’s 45 minutes with a short live Anaplan demo.

Read the series:

Would you trust an AI agent to build your next Anaplan model?

Can you explain your AI forecast to the CFO?

If this feels recognisable, we would value the conversation. Reach us at info@bedfordconsulting.com or follow Bedford Consulting on LinkedIn.