If you own a high-end robot vacuum or mop from iRobot, it’s about to get easier to use. The iRobot Genius Home Intelligence Platform launched today, adding a range of software features to iRobot’s Wi-Fi connected Roomba vacuums and Braava jet mops.

Using the new iRobot Home App, the Genius platform offers personalized control over how, when and where the robots clean. And iRobot claims the system gets smarter over time, similar to the Nest thermostat. iRobot Genius is currently available on the Roomba i7/i7+ and s9/s9+ robot vacuums and Braava jet m6 robot mops.

Theoretically, iRobot Genius could also work on iRobot’s robot lawn mower, if it ever sees the light of day. During its Q1 2020 earnings calls in late April, iRobot said it “suspended our go-to-market plans associated with our Terra robot mower” due to COVID-19.

Machine learning enables precision cleaning

One of the major features is precision clean zones. Using machine learning, the robots can automatically detect and suggest areas that are messy most often. In my home, for example, the floor under the kitchen table is always a disaster as I have three kids under seven years old. The robot would recommend, via the app, that this area be designated a precision clean zone. Users can manually set clean zones, too.

I could then use the app to clean the specific area or use a smart speaker, such as Amazon Alexa or Google Home, and say, “Roomba, clean around the kitchen table.” I could also tell the Roomba to clean around the couch, for example, and the robot will target just that area.

“Because we have enough machine learning on the platform, a robot can cruise around the house and say, ‘hey, this looks like a table. Would you like to place this on your map?'” said iRobot CPO Keith Hartsfield.

iRobot’s VSLAM technology gathers 230,400 data points per millisecond, the company claims, and it has sold 30 million robots worldwide. iRobot has seen its fair share of tables, chairs, couches, kitchen islands, and dog bowls over the years. It’s learned what those, and many other household objects, look like. And its Imprint Smart Mapping helps it remember different rooms.

“We’ll recognize more objects as we go, but that’s the easy thing at this point,” Michelle Gattuso, iRobot’s VP of Digital, told The Robot Report. “The point of view is the tricky thing as we’re taking it from the robot’s level. So building a box around the objects can be tricky.”

Simplifying when to clean

iRobot Genius also hopes to make it easier for the robots to get to work when you want them to. Here are the ways iRobot is doing this at the start:

Event-based automations: These let the robot know when to start or stop cleaning based on user prompts. This will work well for users that have other smart home devices, such as a smart door lock. When the smart door lock is triggered as a user leaves, for example, the Roomba could start cleaning the house.

“We’re trying to eliminate the idea of having to think about cleaning for users,” said Gattuso. “They just want it clean.”

Recommended cleaning schedules: If the robot is typically cleaning first thing in the morning, for example, it’ll recognize this and offer it as a recommended cleaning schedule. The robot can also recommend room-specific cleaning schedules.

“We want the recommendations to be personalized, but we also want to give the users control,” said Gattuso. “The robot will start by making a single recommendation once in a while. If the user likes the recommendations, we can offer them more. If they decline the recommendations, we can pull back on them. We can learn as we go with the users.”

There’s also seasonal recommendations for when your home may need more frequent cleaning, like pet-shedding or allergy seasons.

Favorites: These are essentially pre-set cleaning routines set by users. With a favorite called “After Dinner,” for example, the robot could clean the dining room and in front of the kitchen counter. Or a “Bedtime” favorite could have the robot clean the living room and kitchen, for example.

Another feature of iRobot Genius includes recommended keep out zones. Yes, Roombas still get stuck. But if one routinely gets stuck on a rat’s nest of wires near the TV stand, for example, it’ll learn over time that this area gives it trouble. The robot will recommend that it should stay away from the area.

“We realized in the early days people were skeptical about the autonomy of the Roomba,” said Hartsfield. “But in the last year or so, we’ve seen a significant shift. They want us to clean on their behalf, not giving them a helping hand. They don’t want a helper, they want us to do it for them.”

“Autonomy is not the end game. Users want collaborative intelligence,” said Hartsfield. “A partner that can listen and adapt to your needs, and that’s where we’re headed. That’s a new angle on the problem we’re looking to solve.”

Smart speakers can be used for some of these features.

iRobot Genius just getting started

This is Version 1.0 of iRobot Genius, and Gattuso said a new iteration will be released every six months. It’s available to both Android and iOS users, but it’ll take two weeks to be rolled out to all users. Gattuso said integrating the new iRobot Home App with the apps for the smart speakers was particularly challenging.

“Most users interact with the Amazon app, not necessarily our app,” she said. “The Alexa app didn’t understand what a sofa is, for example, but we do. The tighter integration and set-up was the hard part.”

iRobot Genius is free for users. “We’re starting where users eat and collect food the most – in the kitchen around a table or in a living area, which is typically a sofa or a chair,” said Gattuso. “As we get higher confidence, we’ll continue to roll out new features from there.”

Data privacy concerns

Of course, understanding all this context about users and their homes raises data privacy concerns.

“We will never sell anyone’s data,” Hartsfield said. “We’ve spent a lot of time and effort to make sure the data is secure. We encrypt the data on the device, we encrypt it when we send it to the cloud. We use the data to allow for better outcomes for users, if they allow us to.”