Robot Guide Olly’s First MRT Trial: What Went Right and Wrong

Olly the Robot Guide Hits the MRT: How Did the First Trial Actually Go?

If you have ever watched a tourist spin around helplessly at Dhoby Ghaut trying to figure out which exit leads to Plaza Singapura, you will understand exactly why Singapore is experimenting with robot guides on the MRT. Olly, a humanoid service robot, completed its first live trial on the MRT network in late August 2026, and the results were, well, mixed.

We covered Olly’s earlier appearance in our piece on the humanoid robot trial at Little India MRT station, where the robot was tested in a more controlled environment. Moving it into the wider MRT network is a significantly bigger step. More passengers, more exits, more line changes, more questions in Mandarin, Tamil and Malay. The complexity ramps up fast.

What Did Olly Get Right During the MRT Trial?

Olly handled several categories of commuter queries surprisingly well. According to reports from the trial, the robot correctly directed commuters on which MRT line to take for common destinations, identified the right interchange stations, and explained the SimplyGo contactless payment system clearly to confused first-timers.

That last point matters more than it sounds. As of September 2026, the Land Transport Authority (LTA) has fully transitioned the network to SimplyGo, and a steady stream of visitors still approaches station staff asking how to tap in and tap out with their bank cards or mobile wallets. If Olly can field those questions accurately and free up station staff for more complex situations, that is a real operational win.

Olly also reportedly managed to identify the correct platform direction at single-line stations, which is one of the simpler but more frequently asked questions on the MRT. Asking “which side do I board for Jurong East?” at an East-West Line (EWL) station, for instance, is exactly the kind of clear-cut query where Olly performed well.

Where Did Olly Stumble? The Honest Breakdown

The stumbles were real, and worth understanding honestly rather than dismissing or inflating.

Specific exit number queries appear to have tripped Olly up. Telling a commuter to take “Exit A” when Exit B is actually faster for a given destination is the kind of small error that is harmless for a tourist but genuinely frustrating for someone late for a meeting. Singapore’s larger interchange stations, such as Raffles Place, City Hall and Bugis, have multiple exits serving different buildings and street corners. The granularity required there is considerable.

Real-time information is another obvious gap. Olly, at least in its current configuration, does not appear to be pulling live train frequency or delay data. So if a commuter asked “how long until the next Circle Line (CCL) train?” during the trial, the robot could not give a reliable answer. That is not a criticism unique to Olly; it is a hardware and systems integration challenge that requires the robot to be linked into SMRT’s and SBS Transit’s real-time data feeds, which is a non-trivial engineering task.

There were also reported difficulties when commuters phrased questions in informal or colloquial ways. Asking “how do I get to town?” rather than “how do I get to Orchard?” or “how do I get to City Hall?” could leave Olly uncertain, because “town” is perfectly natural Singapore shorthand that a human station staff member understands instantly.

How Does Olly Compare to Existing Station Assistance at MRT Stops?

It is worth putting Olly into context alongside what already exists at MRT stations today.

The comparison makes Olly look like an early prototype, which is exactly what it is. The station staff column still wins on almost every metric right now. But that misses the point. Staff availability has always been variable, especially during off-peak hours and at smaller stations. A robot that can be deployed at 10pm at a quieter station on the Downtown Line (DTL) or the Thomson-East Coast Line (TEL) and handle 70 percent of common queries correctly is still genuinely useful.

Why Is the MRT a Particularly Hard Environment for a Robot Guide?

This is the question that does not get asked enough. The MRT network as of September 2026 spans over 200 stations across six heavy rail lines and three Light Rapid Transit (LRT) lines. Each station has its own exit configuration, escalator layout, and surrounding street-level geography. The three new CCL Stage 6 stations at Keppel, Cantonment and Prince Edward Road, which opened in July 2026, added their own set of exits and surrounding development to that mental map.

You can check out our full breakdown of those stations on the Circle Line page to see just how much spatial detail each station involves. Encoding all of that into a robot’s knowledge base, and keeping it current as stations are upgraded or renumbered, is a significant ongoing task. It is not a one-time programming job.

Then there is the physical environment itself. MRT platforms are loud during peak hour. Acoustic interference will affect voice recognition. Commuters rushing for trains are not always patient enough to wait for a robot to process their query. And the robot needs to physically navigate crowds without becoming an obstacle itself.

What Does This Mean for Singapore’s Broader Push on MRT Technology?

The Olly trial sits within a much wider push to bring technology deeper into the MRT ecosystem. Singapore has recently completed tests on a new MRT backup train control system, with findings currently being evaluated, as reported by The Straits Times in late August 2026. Our own coverage of the MRT backup train control system tests goes into what that could mean for service resilience. And separately, Singapore has been integrating AI into network monitoring and maintenance, something we explored in our piece on Singapore putting AI to work on the metro network.

Olly is the passenger-facing end of that broader technological investment. It is visible, which is partly why it gets attention. But the less glamorous infrastructure work happening in the background, from axle-box monitoring systems rolling out across all lines to improved signalling on newer extensions, is arguably more impactful for most commuters day to day.

There is also an accessibility angle that deserves mention. A well-functioning robot guide could be genuinely transformative for elderly commuters, visually impaired passengers, or tourists unfamiliar with Singapore’s transit conventions. The push to better support commuters in public transport has been a sustained priority for LTA and the operators, and a patient, always-available robot guide aligns well with that goal if the accuracy issues can be resolved.

Apple Pay Express Transit and Other August 2026 MRT News

Olly was not the only MRT-adjacent story in the final week of August 2026. Apple Pay Express Transit Mode was confirmed to be working in Singapore for train rides as of 31 August, meaning iPhone users can now tap in and tap out at MRT fare gates without opening their phone or authenticating with Face ID. That is a meaningful quality-of-life improvement for the roughly one-third of smartphone users in Singapore on iOS, and it complements the existing contactless payment options already available on the MRT.

Meanwhile, a 20-year-old Singapore permanent resident walked the entire 50-kilometre length of the Downtown Line in 16 hours, which is a fun reminder that the DTL’s 34 stations cover a huge swathe of the island from Bukit Panjang all the way to Expo. If you want a more comfortable way to cover that ground, the Downtown Line map is worth bookmarking.

When Will Olly Be Deployed More Widely on the MRT Network?

No confirmed timeline has been announced as of September 2026. The current phase is explicitly a trial, and the honest answer is that a wider rollout depends on how well the trial data is assessed and what technical improvements are made following the August run. Singapore’s approach to public transport innovation has generally been methodical rather than rushed, which is part of why the network’s reliability has reached its best sustained levels in years, as we noted in our coverage of the MRT reliability improving past the two-million train-kilometre mark.

If Olly does expand beyond its current trial stations, the most logical candidates for early deployment would be high-traffic interchange stations where query volume is highest and where staff presence can sometimes be stretched thin during off-peak hours. Stations like Jurong East (NS1/EW24), Bishan (NS17/CC15) and Outram Park (EW16/NE3/TE17) handle enormous passenger volumes and attract a mix of regulars and first-timers.

Should Commuters Be Excited, Sceptical, or Both?

Both, honestly. Healthy scepticism is warranted because a robot guide that gives wrong directions is worse than no guide at all. A confident wrong answer costs a commuter time and trust. At the same time, the fact that Olly got a meaningful number of queries right in its very first live MRT trial is not nothing. These systems improve with data, and every interaction in a real station environment is a training input.

Singapore has a strong track record of iterating on public transport technology carefully before wide deployment, from the phased rollout of platform screen doors decades ago to the gradual expansion of SimplyGo contactless payment. According to LTA’s published approach to transport innovation, passenger safety and service reliability are always the baseline requirements before any technology goes network-wide. Olly will need to clear that bar, and the August trial data is presumably part of building that case.

For now, Olly is an interesting experiment with a mixed but not discouraging first report card. The MRT is a complex, fast-moving environment, and navigating it well enough to help real commuters is genuinely hard. Getting some of it right in the first trial is a decent start. Our full MRT stations directory covers every station on the network if you want to see just how much geography Olly eventually needs to master.

Before You Tap In

Whether Olly is deployed at your regular station or not, knowing the network well is always your best defence against a missed connection or a wrong exit. Bookmark our interactive Singapore MRT map for a quick visual reference before your next journey, and check our MRT operating hours guide so you always know your first and last train options. We will update this page as the Olly trial progresses and more details emerge from the evaluation phase.

FAQ

Keep exploring

• Humanoid Robot Trial at Little India MRT Station
• Singapore Puts AI to Work on the Metro Network
• The Push to Better Support Commuters in Public Transport