Why AI Isn't Reducing Costs & Ways to Get Results

The recent wave of executive comments about AI not paying off isn’t really surprising. Uber's Andrew Macdonald recently pointed out that the company is spending more and more on AI without seeing a matching improvement in what customers get. Other companies are noticing the same thing. Using more AI doesn't automatically lower costs or grow revenue.

Imagine a construction crew that's always built houses with hammers. Management buys everyone nail guns (faster, more powerful, a real upgrade in capability). But the schedule still assumes hammer-speed. The crew size stays the same and the sequence of work never changes. What happens? The nail guns get used, but the house doesn't go up any faster. The crew finishes their daily quota early and waits. Or they keep the old pace because that's what the schedule calls for. The tool is faster, but the output isn't because nobody redesigned the job around what the new tool can do.

Companies are having the same experience with AI. Leadership buys the capability, hands it to the team, and expects the P&L to improve. But if the workflow, approvals, roles, and expectations aren't adjusted, the new tool just sits inside an old system.

AI is helping people work faster, but many companies haven’t changed the work itself. If a business wants cost reduction from AI, it’s going to require leadership to make tangible operating decisions.

Where to Start

To reduce costs with AI, start with these operational changes:

Meetings. Calendar and transcript data can show who is spending time where, who is contributing, and who’s time can be better spent elsewhere. Rethink meeting attendance and use AI agents to push summaries and action items to the right people afterward.

Organizational layers. Look at how many layers of management are really needed. AI gives managers better information and faster answers, which means one manager can effectively handle more people and doesn’t need as many middle layers to pass information up the chain.

Approval cycles, handoffs, and internal reporting. For many companies, wasted time comes from waiting for approvals, reformatting updates, and moving information from team to team. Bolt recently folded its HR department into a smaller "People Ops" team as part of a broader push to operate "leaner and more AI-centric." Whatever you think of the move, it's an example of a company rethinking whether a long-standing function still fits the business vs. just bolting AI onto the existing org chart.

With each of these changes, AI creates the opportunity to simplify the organization, but leadership has to actually do the simplifying.

How to Identify Where to Cut Costs

How should a company identify where cuts need to be made?

1. Map where labor hours go by function/team. This includes meetings, reports, approvals, coordination, fixing mistakes, etc.
2. Identify the jobs where most of the time is spent shuffling information instead of making decisions or helping customers.
3. Count the handoffs. How many people touch a decision, document, or request before it's done?
4. Run small pilots where AI plus a redesigned process removes steps, cuts down meetings, or lets one manager handle a bigger team. Then compare results before and after.

The biggest opportunities for greater efficiency lie in administrative coordination, extra layers, duplicate reporting, and approval traffic. Companies can see real productivity gains by honestly looking at where time goes, redesigning the workflow around what AI can do now, and then reshaping the organization to match how work gets done.

Remember, using more AI doesn’t mean a company will produce more value.

At Trenegy, we help organizations turn AI ambition into measurable results. To chat more about this, reach out to us atinfo@trenegy.com.