Piloting new technology on a job without disrupting the work

Most construction technology is bought after a good demo and abandoned after a hard month on site. The tools are often fine. The pilot was not set up to show whether they help. A pilot is an experiment: one problem, one job, a clear measure, and a decision date. This checklist works for sensors, apps, drones or anything else.

1. Start with a problem you can name

Write the problem in one sentence, in your own numbers if you have them. "We ran out of diesel twice last quarter and lost half a day each time." "Daily logs reach the office three days late." "We find out about short deliveries at the end of the month."

If you cannot name the problem, you will not be able to tell whether the pilot solved it.

2. Pick one job and one owner

Choose a job with enough time left to learn something, a superintendent or foreman who is willing, and a problem the tool addresses. Name one person who owns the pilot day to day. Without an owner, setup questions sit unanswered and the crew drifts back to the old way.

3. Decide what success looks like before you start

Pick one or two measures you can check without much effort, and the number that would make it worth continuing:

• run-outs or emergency deliveries per month;
• variances found and explained, and how quickly;
• days from site to office for a daily log or a timesheet;
• time spent on a weekly count or report.

Set a decision date, usually four to eight weeks out. Agree in advance what you will do if the answer is no.

4. Run it alongside, then instead

For the first weeks, keep the old process running beside the new one. That gives you a comparison and a fallback. Once the new figures agree with the old, or prove better, stop the duplicate work. Running both for too long wears people out.

5. Bring the crew along

• Show what is in it for them. Fewer phone calls about fuel, no end-of-week timesheet chase, less climbing up to dip tanks.
• Keep the first step tiny. One scan, one tap. Add more once that is a habit.
• Train on site, on the real thing. Ten minutes at the tank beats an hour in the office.
• Report back. When the new tool catches something, say so at the next toolbox talk.

6. Check the practical details early

• Connectivity. Does the site have signal where the work happens? Does the tool work offline?
• Hardware suitability. Anything fitted near fuel or chemicals must suit the hazardous location. Have a competent person assess it.
• Who can see what. Subcontractors, owners and responders need different views.
• Your data. Can you export everything in a usable format, and keep it if you stop paying?
• Running costs. Data plans, batteries, replacement labels, subscriptions per user or per job.

7. Decide, and write down why

On the decision date, look at your measures and the crew's view. Expand, change something and run again, or stop. Write down what you learned either way. A pilot that ends in "no" but tells you why has done its job.