Interview with Andrea Thomaz (co-founder of Diligent Robotics): socially clever automation options for hospitals

By Sonia Roberts, with further enhancing by Dharini Dutia

Diligent Robotics, based by Andrea Thomaz and Vivian Chu, develops socially clever automation options for hospitals. Moxi, their flagship robotic, delivers gadgets like drugs and wound dressings between departments to avoid wasting the medical employees’s time. Diligent has simply closed their Collection B funding spherical with $30 million. 

We sat down with Dr. Thomaz to speak about Moxi, methods to handle individuals’s expectations about robots, and recommendation for younger individuals and girls in robotics. This interview has been calmly edited for readability. 

Andrea Thomaz with Diligent’s flagship robotic Moxi.

What sorts of issues are you attempting to unravel with Moxi?

We’re constructing Moxi to assist hospitals with the huge workforce scarcity that they’re seeing now greater than ever. We truly began the corporate with the identical intention a number of years in the past earlier than there was a worldwide pandemic, and it actually has simply gotten to be a fair greater drawback for hospitals. I really feel actually strongly that robots have a spot to play in teamwork environments, and hospitals are an awesome instance of that. There’s nobody individual’s job in a hospital that you’d truly need to give over to automation or robots, however there are tiny little bits of lots of people’s jobs which are completely in a position to be automated and that we may give over to supply robots specifically like Moxi. [The main problem we’re trying to solve with Moxi is] level to level supply, the place we’re fetching and gathering issues and taking them from one space of the hospital to a different. 

Hospitals have numerous stuff that’s shifting round day-after-day. Each individual within the hospital goes to have sure drugs that must be delivered to them, sure lab samples that must be taken and delivered to the central lab, sure provides that want to come back as much as them, meals and diet day-after-day. You could have numerous stuff that’s coming and going between affected person models and all these completely different help departments. 

Each considered one of these help departments has a course of in place for getting the stuff moved round, however it doesn’t matter what, there’s stuff that occurs each single day that requires ad-hoc [deliveries] to occur between all of those departments and completely different nursing models. So typically that’s going to be a nurse that simply must get one thing for his or her affected person and so they need that to occur as quickly as doable. They’re attempting to discharge their affected person, they want a specific wound dressing package, they’re going to run down and get it as a result of they need to assist their affected person get out. Or if there’s one thing that must be hand carried as a result of the common rounding of medicines has already occurred, numerous instances you’ll have a pharmacy technician cease what they’re doing and go and run some infusion meds for a most cancers affected person, for instance. It form of falls between these departments. There’s completely different individuals that will be concerned however numerous instances it does fall on the nursing models themselves. A nurse defined to us one time that nurses are the final line of protection in affected person care.

Moxi performing a supply for a medical employees member.

What’s altering with this most up-to-date spherical of funding?

Over the past 6-12 months, the demand has actually skyrocketed such that we’re barely maintaining with the demand for individuals eager to implement robots of their hospitals. That’s the rationale why we’re elevating this spherical of funding, increasing the crew, and increasing our means to capitalize on that demand. A few years in the past, if we had been working with a hospital it was as a result of that they had some particular funds put aside for innovation or that they had a CTO or a CIO that had a background in robotics, however it actually wasn’t the very first thing that each hospital CIO was fascinated with. Now that has utterly modified. We’re getting chilly outreach on our web site from CIOs of hospitals saying “I must develop a robotic technique for our hospital and I need to study your resolution.” By the pandemic, I feel everybody has seen that the workforce scarcity in hospitals is barely getting worse within the close to time period. Everyone needs to plan for the longer term and do the whole lot they’ll to take small duties off of the plates of their medical groups. It’s been actually thrilling to be a part of that market change and see that shift to the place all people is actually actually open to automation. Earlier than we needed to say “No no no, this isn’t the longer term, I promise it’s not scifi, I promise these actually work.” Now [the climate has] actually shifted to individuals understanding “That is truly one thing that may impression my groups.”

[Two of our investors are hospitals, and] that’s been considered one of our most enjoyable components of this spherical. It’s at all times nice to have a profitable funding spherical, however to have strategic companions like Cedars-Sinai and Shannon Healthcare coming in and saying “Yeah, we truly need to construct this alongside you” — it’s fairly thrilling to have clients like that. 

What sorts of technical issues did you run into once you had been both constructing Moxi or deploying it in a hospital surroundings? How did you resolve these issues? 

One which was nearly stunning in how typically it got here up, and actually impacted our means [to run Moxi in the hospital environment] as a result of now we have a software-based robotic resolution that’s connecting at a daily foundation to cloud providers, [was that] we had no clue how horrible hospital WiFi was going to be. We truly spent fairly some time constructing in backup techniques to have the ability to use WiFi, backup to LTE if now we have to, however be good about that so we’re not spending a complete bunch of cash on LTE information. That was an issue that appeared very particular to hospitals specifically.

One other one was safety and compliance. We simply didn’t know what among the completely different necessities had been for hospitals till we truly acquired into the environments and began interacting with clients and understanding what they wished to make use of Moxi for. Once we had been first doing analysis trials in 2018 or 2019, we had a model of the robotic that was somewhat bit completely different than the one now we have at present. It had numerous open containers so you would simply put no matter you wished to on the robotic and ship it over to a different location. We rapidly discovered that that restricted what the robotic was allowed to hold, as a result of a lot of what [the customers] wished was to know who pulled one thing out of the robotic. So now now we have an RF badge reader on the robotic that’s related to locking storage containers which are solely going to open in the event you’re the form of individual that’s allowed to open the robotic. That was an fascinating technical problem that we didn’t learn about till after we acquired on the market. 

Moxi’s locking storage containers.

How did you’re employed with nurses and the opposite healthcare professionals you had been working with to determine what could be essentially the most useful robotic for them? 

My background, and my co-founder Vivian Chu’s background, is in human-robot interplay so we knew that we didn’t know sufficient about nursing or the hospital surroundings. We spent the primary 9 months of the corporate in 2018 constructing out our analysis prototype. It appeared rather a lot like what Moxi appears to be like like at present. Below the hood it was utterly completely different than what now we have at present when it comes to the reliability and robustness of the {hardware} and software program, however it was sufficient to get that platform out and have it deployed with nursing models. We embedded ourselves with 4 completely different nursing models throughout Texas over a year-long interval. We’d spend about 6-8 weeks with a nursing division, and we had been simply there — engineers, product individuals, and all people within the firm was biking out and in per week or two at a time. 

We’d ask these nurses: “What would you truly need a robotic like this to do?” A part of this that was actually vital was they didn’t have good concepts about what they’d need the robotic to do till they noticed the robotic. It was a really participatory design, the place they needed to see and get a way for the robotic earlier than they’d have good concepts of what they’d need the robotic to do. Then we might take these concepts [to the company] and are available again and say “Sure we will do this,” or “No we will’t do this.” We got here out of that complete course of with a extremely nice concept. We prefer to say that’s the place we discovered our product market match — that’s the place we actually understood that what was going to be most useful for the robotic to do was connecting the nursing models to those different departments. We may also help a nurse with provide administration and getting issues from place to put inside their division, or we may also help them with issues which are coming from actually far-off. [The second one] was truly impacting their time method far more.

As a result of the capabilities of robotic techniques are often misinterpreted, it may be actually exhausting to handle the connection with stakeholders and clients and set applicable expectations. How did you handle that relationship?

We do numerous demonstrations, however nonetheless with nearly each single implementation you get questions on some robotic in Hollywood, [and you have to say] “No, that’s the flicks” and clarify precisely what Moxi does. 

From a design perspective, we additionally restrict the English phrases that come out of Moxi’s mouth simply because we don’t need to talk a extremely excessive stage of intelligence. There are many canned phrases and interactions on the iPad as a substitute of through voice, and numerous instances the robotic will simply make meeps and beeps and flash lights and issues like that. 

Earlier than beginning the corporate, I had a lab, and one of many large analysis subjects that we had for numerous years was embodied dialogue — how robots might have an actual dialog with individuals. I had an excellent appreciation for the way exhausting that drawback is, and likewise for simply how a lot individuals need it. Folks come as much as a robotic, and so they need it to have the ability to discuss to them. How one can [set expectations] with the design and conduct of the robotic has been a spotlight of mine since earlier than we began the corporate. We purposefully don’t make the robotic look very human-like as a result of we don’t need there to be android human-level expectations, however [the robot does have a face and eyes so it can] talk “I’m that factor” and “I’m about to control that factor,” which we expect is vital. It’s actually about placing that stability. 

What would you say is one lesson that you simply’ve discovered out of your work at Diligent to date and the way are you seeking to apply this lesson shifting ahead?

The distinction between analysis and observe. On the one hand, the motivation and purpose for beginning an organization is that you simply need to see the sorts of issues that you simply’ve accomplished within the analysis lab actually make it out into the world and begin to impression actual individuals and their work. That’s been some of the fascinating, impactful, and provoking issues about beginning Diligent: With the ability to go and see nurses when Moxi is doing work for them. They’re so grateful! If you happen to simply dangle again and watch Moxi come and do a supply, nearly at all times individuals are tremendous excited to see the robotic. They get their supply and so they’re like, “Oh, thanks Moxi!” That looks like we’re actually making a distinction in a method that you simply simply don’t get with simply analysis contributions that don’t make all of it the best way out into the world. 

That being stated although, there’s a lengthy tail of issues that you need to resolve from an engineering perspective past [developing a feature]. My VP of engineering Starr Corbin has this good way of placing it: The analysis crew will get a sure factor on the product to be function full, the place we’ve demonstrated that this function works and it’s an excellent resolution, however then there’s this complete part that has to occur after that to get the function to be manufacturing prepared. I’d say my greatest lesson might be the whole lot that it takes, and the complete crew of individuals it takes, to get one thing from being function full to manufacturing prepared. I’ve a deep appreciation for that. How briskly we will transfer issues out into the world is actually dictated by a few of that. 

Andrea Thomaz (left) and Vivian Chu with Moxi.

What recommendation would you give younger ladies in robotics? 

If I put my professor hat on, I at all times had recommendation that I preferred to provide ladies in robotics, in academia, and simply form of pursuing issues on the whole. Imposter syndrome is actual, and all people feels it. All you are able to do to fight it’s not underestimate your self. Converse up and know that you simply deserve a seat on the desk. It’s all about exhausting work, but additionally ensuring that your voice is heard. Among the mentorship that I gave to numerous my ladies grad college students after I was a professor was round talking engagements, talking types, and communication. It may be actually uncomfortable once you’re the one something within the room to face up and really feel such as you should be the one talking, and so the extra that you simply observe doing that, the extra snug it will probably really feel, the extra assured you’ll really feel in your self and your voice. I feel discovering that assured voice is a extremely vital ability that you need to develop early on in your profession. 

What’s one piece of recommendation you’ve obtained that you simply at all times flip to when issues are powerful? 

There are two mentors that I’ve had who’re ladies in AI and robotics. [In my] first yr as a college member [the first mentor] got here and gave a analysis seminar discuss. I for some purpose acquired to take her out to lunch on my own, so we had this wonderful one-on-one. We talked somewhat bit about her discuss, most likely half of the lunch we talked about technical issues, after which she simply form of turned the dialog [around] and stated “Andrea, don’t neglect to have a household.” Like, don’t neglect to deal with that a part of your life — it’s an important factor. She acquired on a soapbox and stated “You need to have a piece life stability it’s so vital. Don’t neglect to deal with constructing a household for your self, no matter that appears like.” That basically caught with me, particularly as [when you’re] early in your profession you’re frightened about nothing however success. It was actually highly effective to have anyone robust and influential like that telling you “No, no, that is vital and it’s essential to deal with this.” 

The opposite individual that’s at all times been an inspiration and mentor for me that I’ll spotlight [was the professor teaching a class I TA’d for at MIT]. I had discovered a bug in considered one of her homework issues, and he or she was like, “Oh, fascinating.” She was so excited that I had discovered a query that she didn’t know the reply to. She [just said], “Oh my gosh I don’t know, let’s go discover out!” I keep in mind her being this nice professor at MIT, and he or she was excited to search out one thing that she didn’t know and go and study it collectively versus being embarrassed that she didn’t know one thing. I discovered rather a lot from that interplay: That it’s enjoyable to not know one thing as a result of you then get to go and discover the reply, and regardless of who you might be, you’re by no means anticipated to know the whole lot.

tags: Enterprise, c-Well being-Medication, humanoid, interview, Individual, Service Skilled Medical Different, startup