Phased Pilots for Multisite Operators: Start at One Location

Why start with one site instead of the whole portfolio?

A phased pilot at one representative location is the cautious way for a multisite operator to learn what a robot does on its real floors. The sequence is short: pick one site that stands in for the rest of the portfolio, agree on what success looks like, run the robot through a full operating cycle, and only then expand. Every later decision inherits the quality of that first site's record.

The market context explains why operators feel pressure to move faster. According to Automate, which summarized IFR's World Robotics 2026 service robot report, professional service robot shipments rose 24% to almost 250,000 units in 2025. Transportation and logistics was the largest professional application, at 117,500 units and a 47% share, according to Supply Chain 247.

Read those figures as a sign of direction, not a forecast for your sites. The same Automate coverage notes that the data come from a supplier survey of 238 producers out of 1,001 manufacturers, that totals are not projected to the whole market, and that IFR warns against comparing figures across reports because the sample changes.

How do you pick the pilot site?

Choose the site that looks most like the portfolio average. A flagship with unusual traffic teaches little about the other locations, and a struggling store can bury robot results under problems that have nothing to do with robots. You want a site whose ordinary week says something true about the rest. Screen candidates against four tests:

Large sites can still serve as the first phase. According to Becker's Hospital Review, Mercy sent robots to its largest hospital first and then to other Mercy hospitals through the year. That sequence is defensible only when the team writes down why the site was chosen and what it is meant to prove.

• Foot traffic and shift patterns close to the portfolio median
• Floor types, aisle widths and layouts shared by most sites
• Door, elevator and dock systems that other locations also use
• A site manager who will log every exception, every shift

How long should the pilot run?

No single source sets a standard pilot length, and the published examples vary widely. IU Health piloted delivery robots at two hospitals for four months, according to Inside Indiana Business. A ChristianaCare document on issuu, which we have not independently checked, describes a three year cobot pilot launched in 2022 that aimed to learn whether cobots could cut nonclinical workload.

The better yardstick is the full operating cycle, not the calendar. For a multisite operator that means the busiest and quietest weeks, at least one staffing change, any seasonal swing, and recurring events such as month end or delivery days. A pilot that stops before the slow stretch shows only the easy version of the job.

What zones and measures belong in place before go live?

Divide the pilot site into zones before the robot runs. Each zone needs a defined task, a time window, a route, and a named person who handles exceptions. Without that person, every failed run becomes a debate about whose job it was, and the debate buries the data.

Capture a baseline before go live, then set the measures. Useful ones include the share of scheduled runs completed, exceptions per shift with their causes, staff minutes spent on robot touchpoints, downtime and recovery time, and any incident or near miss. Keep the list short. Five measures the whole team reads will beat twenty that nobody checks.

What on-site friction should you expect?

Predeployment work usually moves faster than the site itself. Toyota's hospital robot team, as described on its research page as of November 2026, said the biggest challenge was coordinating with other companies' automatic doors and elevators, and that issues visible only on site still appeared after predeployment work was complete. IU Health likewise had to coordinate with its elevator company so the robot could call elevators.

Track these coordination tasks as their own workstream, with a named contact at each third party, and log them apart from robot faults. A door that will not open for a robot is not a navigation failure, and mixing the two sends the expansion decision in the wrong direction.

This is the kind of work Service Robot Co. handles as a full service commercial robot integrator for US businesses. We choose robots across manufacturers, and deployment, integration and go live support run through one partner and one number, backed by a nationwide US engineer network that can dispatch on site.

What evidence should a regional or corporate team require?

A regional or corporate team should not approve expansion on the site manager's enthusiasm. The approval packet should show pilot data across a full operating cycle, measured against the baseline captured before go live, along with exception logs that name causes.

It should also include a separate integration log that records each door, elevator or dock issue and who resolved it, plus a staff time comparison that finance can rebuild from raw figures. If finance cannot reproduce the result from the logs, the result is not ready for a decision. Test the cost side against the same data by comparing observed staff time with the fixed lease payment on a three year term.

Health care research points the same way. An ACHE research poster describes a sequence of proof of concept, then pilot testing, then phased deployment and optimization. ACHE narrative lessons add that teams should set a common vision and engage affected stakeholders early, since frontline staff work beside the robot every day.

How should expansion proceed after approval?

Expand in waves, not in one jump. The first wave should add two or three sites that resemble the pilot site, so the lessons transfer. The second wave covers the rest of the region once the first wave moves the same measures in the same direction. Each wave gets its own go live support and exception review, and no site joins until its doors, elevators and charging points have been checked.

Our customers get Service Robot Co. robots through a lease. We are OEM neutral, so the robot for each zone is chosen on its task, and we handle financing, deployment, integration, training and service through our nationwide network. Our standard offer is a three year lease with one fixed payment quoted after a site assessment, so the next site can be assessed and quoted before the expansion decision is due.