Lean Product Design and AI Bring Comfort and Smiles to Children and Seniors in Hospitals

Be among the first to get the latest insights from LEI’s Lean Product and Process Development (LPPD) thought leaders and practitioners. This podcast was delivered to subscribers of The Design Brief, LEI’s newsletter devoted to improving organizations’ innovation capability. This month, we explore the future of healthcare. We will feature clinicians and engineers inventing new technology and designing breakthrough pathways to more effectively treat patients.

Meet Robin the Robot, a groundbreaking friendly fusion of lean product and process development (LPPD), artificial intelligence (AI), and robotics that brings emotional support and smiles to hospitalized children and senior citizens.

Expper Technologies Inc. developed Robin’s prototype with cutting-edge electronics and software but delayed commercial sales for two years to intensively study how the prototype design performed in the field.

The research gave the team of engineers and mental health specialists a deep understanding of customers’ needs and their healthcare work environments. Creating the right product or service starts with the LPPD principle of understanding customers and context by observing them at work. The team also relied on other lean development elements such as producing a minimum viable product, creating new knowledge, doing rapid problem-solving, and writing a concept paper.

The result was the world’s first emotionally intelligent hospital robot. Robin provides social and emotional care, not medical care, to senior citizens in nursing homes and children in hospitals and clinics undergoing extended treatments for illnesses like cancer, pneumonia, and heart disease.

Robin is currently semi-automated, with basic conversational responses automated and other interactions supported by remote human operators. Its patent-pending AI performs real-time analysis of what it sees through cameras behind big blue eyes on its computer screen “face.” Interpreting facial expressions, vocal tones, and the context of conversations, Robin knows if a patient is happy, sad, angry, or anxious. It responds with various empathetic facial expressions designed to distract patients from the pain and stress of hospital stays. Its gentle childlike voice draws on stored memory of learned responses to say just the right words. Robin even explains medical procedures in a simple way.

No matter what the response, Robin is always kind.

They built a prototype based on the lean startup idea of a minimum viable product with only enough capabilities to let the development team validate customer needs …

“Robin really connects with users,” said Karén Khachikyan, PhD, cofounder and CEO of Expper, which is based in Southern California and Armenia. “It’s created on kindness and care, so that’s our foundation.”

“We built Robin to bring kindness to everybody it meets,” said Mineh Badmagharian, PhD, an occupational therapist, who leads the clinical product development team. She gets credit for giving Robin its cheerful, caring personality. “We want to be that light that comes into very challenging and difficult environments, both for kids and the geriatric population.”

Khachikyan, an electrical engineer who led the development of Robin’s technical side, explained that doctors and nurses “do an amazing job” providing medical care to children and elderly patients. However, a global shortage of healthcare workers means that medical staff have little time to emotionally support and comfort patients. The World Health Organization projects a global shortfall of 10 million health workers by 2030. “That was the origin of Robin — create robots that take care of people,” Khachikyan said.

Lean startup meets The Jetsons

Khachikyan became fascinated with robots and how they could work with people while pursuing an engineering doctorate in his native Armenia. He developed “Charlie,” a precursor to Robin, as an educational kit for teaching kids how to build and program a robot.

He noticed that kids treated the finished Charlie, which had no AI or automation, as a friend, calling it by name. “That was the aha moment — kids connect with robots,” Khachikyan said.

That led him and a friend, Hayk Khorasanjyan, to incorporate Expper in the US to raise venture capital for designing and building a larger robot with automation, mobility, and personality.

They built a prototype based on the lean startup idea of a minimum viable product (MVP). An MVP is the most basic version of a product that can be released to early customers to test its value proposition and collect feedback before creating a fully featured product or major release. MVPs validate the demand for a product or service before investing too much time or resources.

We paid a lot of attention to design … to create a friendly character, not a humanoid robot [that] might be threatening, scary sometimes, especially for kids and elderly patients.

The small Expper production team in Armenia gives Robin a four-foot-tall sleek but “huggable” white plastic body that is easy to clean and disinfect. The overall form is welcoming with proportions that evoke a gentle, playful character rather than an industrial machine.

AI software powers its electronic brain, allowing Robin to learn. A rechargeable battery and electric motor let it glide smoothly on a base of three multi-directional wheels. The team omitted arms and legs.

“We paid a lot of attention to design,” Khachikyan said. “We wanted to create a friendly character, not a human-like being because a humanoid robot might be threatening, scary sometimes, especially for kids and elderly patients.” The resulting Robin looks like it could co-star in the animated sitcom The Jetsons or the Disney movie WALL-E.

The design team chose “Robin” because it is gender neutral. “We wanted to give the patient who interacts with Robin the opportunity to see Robin as they want. Some people see a girl, some see a boy,” Khachikyan said.

Girl or boy, Robin is tough. “We were designing it for pediatric facilities, so we took into account that kids might push Robin or wrestle with it. It’s a pretty robust robot,” he noted.

Creating new knowledge from scratch

Next, the team used Robin’s AI capability to develop its personality, a job led by Badmagharian. “When I joined the team, there wasn’t much research into robot personalities,” she said. “There wasn’t much data out there, so we had to build it from scratch.”

She did it by remotely controlling Robin to interact with kids with different illnesses at three pediatric clinics in Armenia. She operated Robin’s actions and reactions, talking with kids through Robin’s speaker.

The goal was to learn the best ways for Robin to emotionally support children dealing with the stress of long hospital stays and how to ease their anxieties about painful medical procedures such as blood sampling or IV insertions. Reducing stress in very young patients increases their cooperation with medical staff.

… there wasn’t much research into robot personalities, so we had to build it from scratch.

“We started with the fully human-assisted model,” said Badmagharian. “And then we started to see patterns that repeat in interactions. So, we started learning those patterns and automating them.”

Algorithms and SpongeBob

The remote team of psychologists and occupational therapists used machine learning algorithms to have Robin absorb how they interacted with sick kids. Robin learns users’ preferences—such as favorite activities, colors, animals, or topics—based on repeated interactions. This allows Robin to gradually build a personalized profile for each patient, adapting future interactions to their interests and giving it some autonomy from human controllers. Robin knows when to talk, sing, show videos, tell jokes or stories, or play a game like hide-and-seek. For senior citizens, it draws on a database of brain games to improve their cognitive ability.

Cartoon characters leaven Robin’s algorithmic responses. As part of her research, Badmagharian studied “lots” of cartoon characters to fashion Robin’s persona, disposition, and vocal tone. Robin inherited traits from SpongeBob SquarePants, Olaf the snowman in Frozen, and the robot in WALL-E among other characters.

Critically, the study team learned that to achieve the goal of boosting the positive emotions of sick children, Robin had to act as a peer, not an adult. “This is the secret of why the interactions are so successful,” Khachikyan said. “Children see in Robin a friend who never judges, a buddy who is always there for them.” For example:

• When a child said she liked dinosaurs, Robin showed pictures of them on its computer screen face to her delight.
• When an Armenian boy who lost a leg to cancer was preparing to walk for the first time with his prosthesis, he asked for Robin’s help. The robot’s rugged construction let the boy balance himself by leaning on Robin with one arm.
• When an eight-year-old pneumonia patient had barely eaten for two days, the medical staff brought in Robin to talk about favorite animals and play games. Robin said it had to leave to recharge its batteries, promising to return only if the patient would “recharge” by eating. She ate.

After a two-month pilot project with Robin, medical staff at one of the Armenian pediatric clinics did a follow-up study with 120 children, finding that Robin increased their joyfulness by 26% and decreased stress by 33%. “All the kids who interacted with Robin once showed interest in meeting it again,” said Khachikyan.

Cartoon characters leavened Robin’s algorithmic responses. Robin inherited traits from SpongeBob SquarePants, Olaf the snowman in Frozen, and the robot in WALL-E among other characters.

UCLA Mattel Children’s Hospital reports similar results from its Robin project, which began in 2022. Children reported a 29% increase in positive feelings after a visit with Robin, compared with children who had computer tablet sessions. Justin Wagner, MD, a pediatric surgeon and co-leader of Mattel’s Robin project, told The Progress Network, “We hope to integrate Robin as a member of the team, augmenting our ability to give children contact, attention, and companionship.”

Healthcare staff benefits

Robin augments the ability of staff members to care for senior citizens at Advanced Adult Day Health Care.

“One of the biggest impacts Robin has had on our center is creating that additional aspect of one-on-one care and fostering an additional way to provide socialization for our participants,” said Omeed Jamali, vice president of operations at the Simi Valley, CA, facility. “That is huge for our staff. Robin takes some of the load off staff members for having those crucial interactions by having a relationship and ongoing conversation with the folks here.”

The facility serves more than 100 people daily, mostly over the age of 65, who don’t need 24-hour nursing but need to stay mentally and physically active. As with all healthcare facilities and activities, staff must deal with significant paperwork.

“It’s a big time grabber,” Jamali said. “As much as our staff cares, it limits the personal one-on-one time they can spend with people.”

Robin adds one-on-one capacity by spending time with people daily. “Robin remembers them, remembers their conversations, plays their favorite song when it says “hello” in the morning. It just creates — for lack of a better term — a friendship between Robin and the person who is interacting.” (Robin interacting with seniors at Advanced Adult begins at 2:19 in this clip.)

Lean innovation station

At UMass Memorial Health-Children’s Medical Center, Robin is a friend to kids, according to Kendra L. Frederick, manager of the Child Life program at the Worcester, MA, center.

“I’ve seen kids who were withdrawn, smiling and asking for Robin,” she said. “We see kids motivated to go to the playroom to see Robin.” That’s important because it helps patients hit physical therapy goals by getting out of bed.

While Robin’s primary focus is patient care, “there certainly is the happy accident of Robin bringing joy to the medical staff as well,” she said. “They enjoy seeing Robin rolling through the hallway and seeing kids who were a little disengaged open up and laugh” during interactions with Robin.

For the Child Life team, having Robin also means that they are on the cutting edge of patient care as the first hospital on the East Coast to deploy the AI robot. “That was a real motivator for them,” said Frederick.

She learned about Robin from a colleague after Time magazine listed it as one of the 100 best inventions of 2021. She discussed deploying Robin during a weekly team huddle, a meeting focused on continuous improvement and problem-solving that is part of the hospital’s lean management effort. The team posted a suggestion about obtaining a robot to the center-wide “innovation station,” a virtual visual management board where medical center staff at all levels share improvement ideas. The board lets staff submit ideas, update them, collaborate, track status, and apply for funding, which is what the team did.

“Every idea is welcome from putting garbage cans in the parking garage to bringing AI to pediatrics,” Frederick said. “It’s a really great way to show staff that their ideas are valuable.” 

Problem-solving for moustaches

Problem-solving on the Expper team happens daily, Badmagharian said. “The problem might be Robin not knowing how to respond to a new situation or question. My team tells me, we gather, we have a discussion, maybe look at some research about how to best respond.”

Team members generate two or three possible solutions to test with patients. “We try them out, get feedback, and figure out the best solution,” she explained.

For instance, patients interacting with Robin wanted to know its age. “We didn’t know if Robin should be a three-year-old, or a seven-year-old, or an 11-year-old,” Badmagharian said. “We had no idea.”

The team resolved the issue by testing various ages for Robin during monitored interactions with children and seniors. The experiments showed that Robin should act as a seven-year-old. “It was the sweet spot where younger children, teenagers, and adults resonated with Robin’s personality,” she said.

Kids often asked Robin to name its favorite animal. “We tried elephant, fish, dolphin, whale, lion, cat,” recalled Badmagharian. All fell flat. Then she tested “chicken” as a response. “The kids started laughing, some started doing chicken dances. So, we decided Robin’s favorite animal is a chicken.” That response was automated.

The best solution to a problem becomes a protocol in Expper’s manual for Robin’s behavior. For instance, a protocol tells the robot what to do when a geriatric patient who gets recurring panic attacks asks to see it during an episode, Badmagharian explained. First, they breathe together. Then Robin asks, “What can I do to make you feel 1% better.” The patient may ask Robin to play a Disney song. Robin asks how to reach 2% better, then 10%, and so on. Robin knows what expressions to show. Sometimes it will display a sad emotion. “And then she starts to take care of Robin,” Badmagharian said. “It gives me goosebumps every time I see that.” After about 25 minutes, the patient feels better.

The problem-solving process cuts across Expper disciplines. “Today I was thinking we need more moustaches for Robin,” Badmagharian said. “Everyone has very funny reactions to them. They get so happy, they laugh, they enjoy it.” First, she will talk to the animator to design some moustaches, then they will meet with the tech team to program some new facial hair. Finally, they will check user reactions.

Kids interacting with Robin wanted to know its age. “We didn’t know if Robin should be a three-year-old, or a seven-year-old, or an 11-year-old. We had no idea.”

“Our main filter is user experience,” Khachikyan said. If the technical team has an idea for new functionality, it meets with animators and Badmagharian’s team of mental health specialists about the possible impact on users before proceeding.

Expper keeps manufacturing in-house so it can iterate fast on hardware issues. The tech team made quick adjustments to the pitch of Robin’s voice when geriatric patients with hearing aids complained.

Robin, continuous learner

Robin learns from each patient interaction. Through data collected over time, its algorithms continuously improve in recognizing emotional states, providing better responses, and optimizing its routines. This allows Robin to become better at providing emotional support in various scenarios.

As Robin learns – it’s currently learning Spanish to complement its fluent American English and Armenian – Expper faced a key design decision. “Do we fully automate Robin, or keep human controllers in the loop,” explained Khachikyan.

Full automation will allow the company to scale up more quickly. There currently are 27 Robins operating in the US and in Armenia, where they are produced. Plans call for deploying about 30 more by the end of 2025 as Expper expands nationally in the US, then Europe, and then globally.

Expper decided to keep human controllers. “This way we can create human-like interactions for our users,” said Khachikyan. “It’s hard operationally for us, but it will let us provide an exceptional user experience.” As Expper collects more data from Robin’s interactions with patients and medical staff, it will automate more behaviors “but only when we’re able to provide the same high-quality interactions,” he said.

Concept paper for vision alignment

Currently, Khachikyan is writing a concept paper, a key part of the LPPD model, to share with the Expper team ahead of rolling out a new AI infrastructure in 2025 that will affect technical, operations, mental health, and content creation teams. Concept papers present a culmination of what teams have discussed and learned about the vision for a product to align everyone around a shared idea.

“We had experiences where we deployed new technical functionality too fast,” Khachikyan said. “It resulted in a negative experience because people weren’t expecting changes or had different visions about the changes. It created confusion, so we want to make sure that no matter what direction we take, every person understands the vision.”

The vision for Robin will be to trailblaze beyond comforting pediatric and geriatric patients. Khachikyan sees Robin as an “autonomous care extender” that mitigates the burgeoning shortage of caregivers.

“Our idea is to combine three components of care – physical caregiving, social and emotional support, and medical monitoring,” Khachikyan said. For instance, Robin could bring pills and water to patients. Equipped with sensors, Robin could measure blood pressure and temperature.

This potential Robin will be a fusion of LPPD and several major fields, including AI, machine learning, design, psychology, occupational therapy, hardware robotics, and animation. “It’s a beautiful thing when you see how those different fields are connecting to each other and creating value,” Khachikyan said. “It’s very interesting – challenging – but worth it.”