Measuring AI ROI honestly: baselines, owners and the numbers that matter
Return on AI is not a promise made in a slide. It is a comparison between a measured starting point and what changed in a real workflow.
By the Halden editorial team
The trap
Usage is not value.
Dashboards often report how many people logged in or how many prompts were run. Those numbers show activity. They do not show whether the work became faster, better or cheaper to deliver.
An honest ROI conversation starts with a different question: what was this piece of work like before, and what is it like now?
Start with a baseline
Measure the current work before changing it.
Pick one workflow and record how long it takes, how often it is redone, how long customers or colleagues wait, and what quality problems occur. A few weeks of rough, honest data is enough.
Without a baseline, every later number is an opinion. With one, even a modest improvement becomes a credible, shareable result.
- Cycle time: elapsed time from start to finished.
- Effort: hours of people's time per item.
- Quality: errors, rework and escalations.
- Experience: customer or employee feedback.
Count the whole cost
Include the work that change creates.
New tools bring costs that rarely appear in a business case: licences, integration, training, time spent reviewing output, and the effort of maintaining prompts and processes.
Include them. A result that survives an honest cost count is far more persuasive than one that ignores the review time it quietly created.
Decide in advance
Set the scale, redesign or stop rule before you begin.
Agree the thresholds with the workflow owner and a sponsor up front. For example: if cycle time falls meaningfully and quality holds, scale; if time falls but rework rises, redesign; if neither moves, stop and record what you learned.
Stopping a pilot with clear evidence is also a good outcome. It saves the organization from scaling something that does not work.
A next step
Turn this into a decision for your organization.
Explore how AI Strategy works with leaders and teams, or start with a short reflection on where you are today.