Samsung Bespoke Jet Bot AI+ robovac VR50T95 series (cleaning review)

The Samsung Bespoke Jet Bot AI+ robovac VR50T95 series and clean station uses AI object recognition, Lidar navigation, IR sensors and a camera for live home monitoring. It also has quite an edgy sci-fi feel that would not be out of place in a Star Wars movie.

Before the review, read Five tips for choosing a robovac/mop (guide) to see where we are coming from.

A few minor caveats.

  • This is a robovac – it does not mop like many competitors.
  • The clean station empties the inbuilt dustbin and charges the battery – it has no additional features.
  • It is both big and tall, 305 x 136.5 x 320 mm (WxHxD) x 4.4kg plus the station. Most robovacs are <105mm to get under overhanging chairs, tables etc.
  • It is also a D-shape design with a 230mm roller brush width. But the machine is 305mm wide, leaving a 37mm gap on either side. Without rotating whiskers and with its height, it does not do edges well.
  • Maximum carpet pile length is 25mm, so you need to carefully set no-go zones if you have feature rugs and long pile.
  • The App is pretty limited and does not include things like saving different maps, cleaning repetition by room/surface, etc. A software upgrade can fix most of these things, and we hope Samsung does so soon.

Samsung Bespoke Jet Bot AI+ robovac Model VR50T95735W/SA

Note: Samsung Australia uses the term Bespoke but judging from the model number, it is different to the Jet Bot 90 AI+ sold elsewhere. If you are searching, look for VR50T95735.

WebsiteProduct page
ManualManual PDF
RRP$1,899
FromSamsung Online. As it is new, there is not a full retail presence yet
Warranty12-months ACL
Made inVietnam
CompanySamsung is a South Korean multinational manufacturing conglomerate headquartered in Samsung Town, Seoul, South Korea. Samsung Electronics (the world’s largest information technology company, consumer electronics maker and chipmaker.
MoreCyberShack Samsung News and Reviews

We use Fail (below expectations), Passable (meets low expectations), Pass (meets expectations), Pass+ (near Exceed but not class-leading) and Exceed (surpasses expectations or is the class leader) against many of the items below.  You can click on most images for an enlargement.

First Impression – Pass

We mentioned Star Wars, and it could easily double for a Death Star cleaning robot. And that is what it is – a dedicated robovac – no mopping.

That is not a bad thing, as most combos have static mop platens that drag dirty water around. You will pay a lot more to get a decent robomop that does as well as a hand mop.

Navigation-wise, it uses 2D LiDAR, IR collision detection, and has a camera for 3D object recognition, so it is among the brainer robovacs.

It appears well-made. Our only complaint is that Samsung’s one-year warranty is way too short. Never mind, you have Australian Consumer Law on your side if it breaks down within a reasonable time – we expect this to last at least five years before you may need a battery change.

Setup – Pass

Setup is via the SmartThing App. It was relatively straightforward, requiring the setting up of a Samsung account and privacy conditions signoff. We won’t explore Samsung Privacy because you will likely own other Samsung SmartThings compatible products. We will say that SmartThings does its best to get you to use Samsung products and services.

Your privacy is also subject to Alexa or Google Assistant voice assistance policies if you use these.

The App – basic – Passable

SmartThings is Samsung’s future ecosystem of interconnected smart things – washers, fridges, cooking, cleaning, and so much more. It will be Matter compatible, but that is not an issue for existing devices. Think of SmartThings like Google Home.

The App is basic missing many features that we have come to expect. Things like multiple maps, programmable cleaning by floor type, resume cleaning after charge, and more are identified in the review.

Cleaning efficiency – depends on what you expect

The test sample is 100g, including sand, sugar, hair, rolled oats, and Nutrigrain spread over 1 m2.

  • Hard floor – 95% leaving the larger debris (Nutrigrain), unable to fit under the rotating brush guard
  • Sisal carpet – 84% leaving lint, hair, fluff and larger particles.

The App has a ‘Avoid particle size’ setting from a Lego block to a large rug – we left it on default at ‘Large’.

Pet Hair – depends on the surface – Pass to Fail

We are dog-sitting a 6-month-old Corgi, and it is shedding 20-30mm hair. It removed pet hair quite well from hard surfaces but left quite a lot (we estimate 50+%) on the sisal carpet. Again, this is not unusual, and you would benefit from a dual cleaning of the carpet areas, but the App does not appear to support that. Alternatively, buy a more powerful vacuum with significantly increased air wattage.

The rotating brush was easy to remove and clean with a scrubbing brush.

The ultimate test – Passable to Fail

The ultimate test is to run the robovac, then follow with a Dyson V15 Detect stick Laser head vac that counts and measures different particles. It found about 20g more debris on the hard floor and 130g more on the carpet. Note that the home is thoroughly vacuumed one week before the four-week test.

Most of the extra hard floor detritus was from the edges with a reasonable clean of open areas.

The carpet test was an apple-for-apples test immediately after the robovac. We put the poor carpet result down to the rotating brush material – it is a soft microfibre with a recessed rubber ridge and simply does not deep clean. The analogy is that Dyson provides a ‘Mr Fluffy’ head for hard floors (like this is) and a bristle and rubber, ribbed head for carpets. If you use Mr Fluffy on carpets, you get a poor result.

It earns a Passable if you expect a moderate clean sans edges. It earns a Fail if you have lots of carpet, want better edge cleaning, and expect more from a $1900 robovac. It is the worst result so far for mid-and-premium range robovacs tested.

The cure: There is an inelegant App setting to keep cleaning until the battery runs out. Most competitors allow you to select the number of repeats by floor type, room, etc.

Edge Clean – Passable

The App has a setting called Safe Driving mode (default off) that sets a 20mm distance between it and the edges to avoid hitting furniture and walls.

During the initial vacuum, it leaves about 25mm edge clearance, plus the rotating brush has a 35mm gap between it and the edge of the robot.

The Apps allow the selection of edge cleaning at the beginning/end of a cleaning run (better cleaning but extends the cleaning time) or a zig-zag cleaning pattern (shorter cleaning time, but the edges are not specifically cleaned).  

We tried all settings, and regardless, you still must use a vacuum cleaner to do edges. Like all robovacs, it trades off good edge cleaning to reduce the potential for wall damage. However, we noticed that robovacs with dual front whiskers do better.

Cleaning power – Pass

It claims 30 Air Watt (about 1500Pa) suction power using an eight-cyclone system. Better ones like the Deebot Omni X1 have 100 Air Watts, and the Ultenic D10 has 66 Air Watts. There is roughly 50 Pa (Pascals) per Air Watt if you see vacuum power as Pa.

It has Normal, Smart and Max settings that impact battery life accordingly. It auto-ranges to suit the surface when set to Smart.

Cleaning time/speed – Passable

It seems slower than our benchmark of about 1m2 per minute. This was closer to .8m2 per minute (49m2 in 61 minutes). It also coincides with the battery being exhausted.

If you have a larger home, you may need to do it over several days between charges, as there appears to be no option to resume if cleaning is not complete after charging (we assume it defaults to that).

Sill negotiation – Pass+

Samsung claims 50mm, and it reliably negotiates sills up to 25mm (tested and better than most competitors). However, it needs enough room to approach this forward-on as it cannot reverse over sills. See reliability below.

Cliff Detection – Pass

The Cliff detectors include one under the front, two immediately forward of the drive wheels and two behind the drive wheels. It is a little scary to see the robovac move at full speed over the top of stairs and then judderingly reverse. Still, they are effective.

But we noticed that it did not vacuum some darker mats. It seems the sensors interpret black or dark mats as cliffs.

House Prep – still necessary despite good obstacle avoidance

If you prepare the house for the robovac, you enhance reliability and reduce the need for obstacle detection. See our 5-tip guide above.

We also tested with various obstacles, including Lego bricks, sports shoes with long laces, and a child’s stuffed toy. It managed to avoid the larger obstacles, albeit the lighter Lego bricks were pushed around at the front of the robovac. It will get caught in long shoelaces and cables – these must be out of harm’s way.

Magnetic Boundary markers – and No-go zones – Pass

It comes with a 1m sample of a self-adhesive wood-grained magnetic strip. You can buy more to set up permanent no-go areas that the App may not be able to define. For example, under the edges of a feature rug, and around areas it will get stuck, etc.

We cannot find this on any Australian website or retailer. Its code is VCA-RMT10 (2m). We understand that it should cost about $40 here. This technology is used by several robovac makers, including Neato (Amazon AU $65 for 6m) and Ecovacs U2 (JB Hi-Fi $24 for 3m), and these will work too.

Obstacle avoidance – Pass

Its 3D camera will avoid obstacles as low as 10mm high (think Lego brick). It also uses AI Object recognition (from machine learning) to avoid (there is no exhaustive list, but we have gathered the following):

Refrigerator, TV, washing machine/dryer, air purifier, AirDresser, sofa, bed, bookshelf, table, Bespoke (Refrigerator, Dishwasher), excrement, electrical cable, towel/sock, cup, Glass cup, bottles, bowl, and vase.

Our dog-sitting dog is well trained, so we could not test on dog poo, but as Samsung has stated this as a fact on its website, we look forward to hearing about user experiences.

It marks obstacles in the app for further investigation. Unfortunately, in every case, marked obstacles were not there and incorrectly identified. It did not identify tables, chairs etc., that other obstacle avoidance models do.

Reliability – Pass

Over four weeks of testing, it did not always complete the cleaning run and return to base. In one case, it edged in behind a half-open door and closed it – blocking its escape until the battery ran out. It has powerful wheel motors that can push doors closed and light rubbish bins aside.

It gets stuck in a special set of circumstances involving door sills and limited manoeuvrability space. The answer is to set no-go zones for problem areas.

Make sure you have strong whole-of-home Wi-Fi 5, 6 or 6E. Our test home has the Netgear Orbi RBKE963 Quad-band Wi-Fi 6E AX 11000 mesh with a special 2.4GHz IoT band without WPA3 encryption. We experienced no Wi-Fi dropouts or loss of location or base. The App has a calibration setting to reorient the robovac with the base station if you have issues.

Cleaning/charging bay – Pass

As the robovac only has a 200ml dustbin, it must return when full. The Station dustbin is 2.5L. These capacities are a little small for a big home.

The App has a setting to empty the robovac dustbin when full or when charging, but it did not return once on any run, yet we know it was full.

We found that the cleaning station did not fully empty the dustbin, more often leaving the pet hair. A repeat empty cycle fixed that.

The cleaning station is 280 x 525 x 400 mm x 5.1kg and requires a power point and clear space around it.

Battey – Passable

Samsung claims 90 minutes in Normal Mode. But on Smart mode with a mix of hard and soft floors, it is closer to 50m2/60 minutes. Recharge time is around three and a half hours.

This creates issues for a larger home where you may need to break cleaning up into 50m2 zones and schedule this over several days.

It is a 21.6V/3.8A/80W battery. Power use is a maximum of 170W while emptying the dustbin and 100-135W when charging.

Noise – Pass

There are three power levels – Normal (45dB), Smart (varies between 45 and 65dB) and Max (62dB). When self-emptying, it hits 75dB for about 30 seconds. This is average for premium robovacs.

Maintenance – Pass

You must regularly clean

  • Rotating brush
  • Front and rear wheels
  • Sensors
  • Fine dust filter in the vacuum (replace annually)
  • Fine dust filter in the clean station (replace annually)
  • Replace the Clean Station dust bag when full (we assume there is a notification for this)

Unlike most robovac apps, Samsung SmartThings does not appear to monitor total usage time, and provide notifications for recommended cleaning.

Clean station dust bags: We could not find the model number of the replacement dust bags. Nor can we find the bags anywhere at Samsung online or Australian retailers. Users have commented that bags are not available, so check this out before you buy.

Voice Assistance – Pass

SmartThings supports Samsung Bixby, Amazon Alexa, or Google Assistant (Siri is not mentioned or tested). Commands are limited (as in all robovacs) to start/stop cleaning, return home or clean a nominated area.

Missing

  • Side whiskers (not expected on a D-shape robovac)
  • Remote Control (use the App)
  • Some fundamental App features

Not Tested

The robovac has a forward-facing camera. Smart Things has a Pet Care Mode and Home Monitor Mode (each requires a separate SmartThings download). We could not get Pet Care to work (camera issues – video recording requires a paid plan), and Monitor Mode assumes using Samsung SmartThing’s sensors. We also got an error message about SmartThings not supporting Android 13.

In the box

CyberShack’s view – Samsung Bespoke Jet Bot AI+ robovac VR50T95 series is as good as it gets in the right home

We review a lot of robovacs and robomops using the same test home and test parameters. It allows us to rate these devices according to their success.

It is pretty good as a dedicated robovac on hard floors – less so on carpet. It is by no means the class-leader, a position Samsung usually occupies. That is a hotly contested position occupied by Ecovacs Deebot Omni X1 and Roborock S7 MaxV Ultra, which add mopping and a complete suite of App features for not much more cost.

Who is the Samsung Bespoke Jet Bot AI+ robovac for?

I can’t help but feel that its designers live in smaller, single-level, hard-floor apartments (a reasonable assumption as they live in South Korea) because this is where it is best used. Give it an 8/10 for that scenario.

But this is Australia, and Aussie homes tend to be larger, multi-level, and with various soft and hard floor coverings. Its battery life on the Smart setting with a 70/30 mix of hard and carpeted floors is about 60 minutes. This means about a 50m2 cleaning area – think of that as 2 x 5 x 5m bedrooms. The App does not appear to support multi-level/multi-maps, and the 200ml internal dustbin and 2.5L cleaning station dustbin are too small for larger homes. Give it a 6/10 for that scenario because it is not the one for you.

Would I buy it?

If I had hard floors, lived in a smaller apartment, and wanted to stay in the SmartThings ecosystem – then yes. But you are paying a premium price.

Competition

Rating Explanation

We will not formally rate this as it is barely a pass on our test home. But in a suitable small hard-floor apartment, it is pretty good.

  • Features: Robovac only. The app needs more reasonably expected features
  • Value: While Samsung gear is always well-made, at $1899, it is out-classed in every category by the Ultenic T10 with a cleaning station (and mop) at $690.
  • Performance: On hard floors, it is great, except for edge cleaning. On carpet, it lacks the power to do a deeper clean despite its Smart setting.
  • Ease of Use: SmartThings is relatively easy to set up, but it lacks too many expected features
  • Design: While I like the Misty White Star War Troopers look, its added height makes it less useful in cleaning under lower furniture.

Samsung Bespoke Jet Bot AI+ robovac VR50T95 series – summary

ItemExplanationGrade
ShapeD-shapeIt has a 230mm brush but no front whiskersPass
Lidar2DStandard LDS 360° single LiDAR builds a 2D map – Gen 3Pass
SensorIRForward and side IR sensor helps to avoid obstaclesPass+
CameraYesFor pet care and helps object recognitionPass+
AppSmartThingsIt only supports one map (an issue for multi-level homes). Meets all typical needs. Modes include vacuum, suction (three levels), battery level and cleaning diary, area vacuum, navigation patterns and schedules. It is less comprehensive than competitors.Passable
Map editYesAll the usual, including no-zonesPass
Magnetic BoundaryYesOnly a 1m sample of the tape – you need more to Samsung-proof your home.Pass
Edge D-shapePoor edge cleaning, even with the App set to do a separate edge clean runFail
Height136mmToo high for cupboard overhangs and under beds, couches etcFail
Carpet Max 25mmCarpet cleaning needs at least two runs to reach average standards, but the App does not allow for thatFail
Hard floors 95% on hard floorsPass+
Pet Hair Good on hard floor but fail on carpetPass/Fail
Sills25mmWorks to 25mmPass+
Other SensorsCollision and cliffA typical forward bumper sensor and cliff detector stop it from going down the stairs. But it also detects black (very dark) carpet as a cliff and will not enter that area.Pass
Suction1500Pa (30AW)On Maximum – other robovacs are 2500-5000PaPassable
Dustbin200mlSmall, but there is a self-empty functionPass
Battery life Claimed 90 minutes on low setting, 60 minutes on Smart and 20 on Maximum. Note that the cleanable area in m2 is roughly equivalent to battery life.Pass
Battery2800mAhIt is 21.6V/3.8A/82W  Pass
Clean Speed.8m2/hr.8m2/min is below the average of 1m2/hrPass
Wi-Fi2.4GHz Same as most robot vacuum/mopsPass
Size 305×136.5×320 mm x 4.4kgPass
Dock2.5 dust bagOn the smallish side – many have 4L or greaterPass

Jet Bot AI+, Jet Bot AI+, Jet Bot AI+, Jet Bot AI+, Jet Bot AI+, Jet Bot AI+

Samsung Bespoke Jet Bot AI+ robovac VR50T95 series (NOT RATED)

$1899
0

Pros

  • Removed 95% of test debris on hard floors
  • Interesting look and design
  • Easy maintenance but replaceable dustbin bag and filter costs are unknown.

Cons

  • Removed 84% of test detritus on the carpet but failed deep clean tests
  • App needs work to add expected competitor’s features
  • 60 minutes on Smart Model means 50m2 of cleaning before charging
  • Not for large, multi-surface Aussie homes
  • Edge cleaning is below average