A favorite vacation spot. A childhood neighborhood. A beloved pet. A lifelong hobby. For someone living with dementia, as memory and communication become more difficult, these details can turn into powerful prompts, sparking memories, stories, and joyful moments of connection.
Memory Lane Games has long used frustration-free, quiz-style games to create these connection moments for people living with Alzheimer’s and dementia. Now, with help from SAP and EY, the organization is exploring how AI can personalize its games at scale.
“What we saw, as AI was coming in, was that we could take two or three more steps to really personalize the experience and trigger positive memories for each individual,” Bruce Elliott, CEO and cofounder of Memory Lane Games, says. “But, as a six-person startup on the Isle of Man, a little island in the middle of the Irish Sea, it was a daunting task. We had a brilliant vision, but to bring it together we needed support.”
Collaboration for good
Through the MovingWorlds’ platform, Memory Lane Games collaborated with SAP and EY to move beyond generic reminiscence content, creating an AI prototype that is designed to help transform personal details, family photos, and life experiences into memory games.
This project was part of Scaling AI for Good, a MovingWorlds cooperation between SAP, EY, and Microsoft that provides social enterprises—that often have limited resources, complex technical needs, and a lack of access to partners—with strategic coaching, pro-bono consulting, expert-led workshops, and access to a global partner ecosystem at no cost.
For employees at companies like SAP and EY, these engagements are equally transformative. Real-world problem-solving with social enterprises is among the most effective forms of experiential learning, helping to develop skills that benefit both the individual and the organization.
Through this program, Memory Lane Games gained access to the technical support needed to explore how generative AI could deepen personalization in its games. To start, the project team—Wade Tsai, global client technology architecture leader at EY, and SAP developers Robin Baeurle and Michael Zadikowitsch—looked at what the organization had done in the past and how it could be transformed by generative AI. The intersection of personalization and images was where the team landed.
Pro-bono consulting: good for the world and for the people doing the work
SAP’s investment in programs like Scaling AI for Good is rooted in a simple belief: we want to bring out the best in our people and we want to bring our best to the world. Over the last decade, thousands of SAP employees have generated more than €30 million in in-kind social investment, partnering with social enterprises across more than 60 countries. Eighty-five percent of those social enterprise partners report an increased ability to serve their beneficiaries, and 88% of participating SAP employees say the experience sparked new ideas they brought back to SAP. Pro-bono consulting isn’t a side program; it is experiential learning at scale, and a core part of how SAP is building a skills-led organization.
Turning memories into personalized games
The foundation of the project team’s prototype is a persona profile that the AI can reference. A caregiver or loved one creates a profile for the person living with dementia with basic information like age bracket, gender, first language, places they’ve lived, what they did for work, and cultural background as well as other details like past vacations, favorite foods, hobbies, pets, and more. The goal is to capture the culture of the individual, not personal data. Using that persona profile, the AI suggests topics and then generates games by pulling in open-source images and writing multiple-choice questions.
For example, a demo of the Memory Lane Games prototype showed the creation of a persona profile of a 75-to-80-year-old woman who lived in Savannah, Georgia, was an elementary school teacher for 35 years, vacationed on Hilton Head Island and in the Blue Ridge Mountains, had a tabby cat named Magnolia, and enjoys Motown and soul music, gardening, birdwatching, and cooking Southern classics like pecan pie and shrimp and grits. The AI prototype generated several games for her: Savannah’s Historic Squares, Blue Ridge Mountain Getaways, Savannah’s Southern Kitchen Favorites, and Magnolia the Cat and Backyard Birds.
“At the start, we didn’t know how to solve this problem of extracting metadata from images, or even how to collect images, where to find them, and how to create these games,” Zadikowitsch says. “And then we started experimenting and trying things out—generating questions and then finding images, which didn’t work well, then finding images and then generating questions, which worked better.”
The prototype is currently in the testing stages. Once it is deployed, there is the possibility to expand game personalization with photos submitted by a caregiver or loved one.
Tsai explains that current AI models can identify objects in images, but directing them to extract metadata and EXIF data, which contains the exact date, timestamp, and GPS coordinates indicating when and where the photo was taken, can take it a step further.
The layers of information stored within a photo file can be a treasure trove for Memory Lane Games’ AI prototype. “We can use that information as a hint to extrapolate what else we can pull from the surroundings that could help reshape that memory experience,” Tsai says. “It’s going beyond where that photo is taken.”
AI with tangible human impact, not just productivity gains
While the AI prototype helps Memory Lane Games create personalized games more efficiently, its greater promise is human: helping people living with dementia and loved ones connect through joyful memories.
“Often in the context of AI, the typical audience is people working at other companies, not, for example, people with dementia,” Baeurle says. “There’s a lot of potential that goes to waste in not realizing that these audiences can also benefit from digital solutions and, specifically, AI-powered solutions.”
For the project team, that human outcome was the guiding principle. “Whatever we build, it should spark joy,” Baeurle adds.
“Caregivers are really busy. Family members are really busy. Care staff in care homes are very busy. If we can take one photo from a family and a couple lines of text and create six to 10 questions that can really pull out all of those memories, trigger a positive memory, and start those wonderful stories—that’s what [this collaboration] has been able to help us deliver,” Elliott says.
AI makes it possible to create memory games not only about a city, but about a neighborhood, a local landmark, or another detail closely tied to someone’s life. The more personal the prompt, the more likely it is to encourage conversation and connection.
“We know that social isolation is one of the top modifiable risks for dementia in older age, so the more we can get people talking and those neurons firing, just by letting them talk about something they want to talk about, is powerful,” Elliott says.
The future of AI-enabled memory care
What this collaboration between Memory Lane Games, SAP, EY, and MovingWorlds showed is that AI has a place in memory care. “The team demonstrated that AI can help create engaging [memory] games. I think this was the goal,” Zadikowitsch says.
This achievement was recognized at the United Nation’s International Telecommunication Union (ITU) AI for Good Global Summit this summer, after Tsai won the “AI for Good for Entrepreneurship” category in EY’s inaugural AI for Impact Challenge with this Memory Lane Games project.
The project team is confident that this is just the beginning for AI-enhanced memory care and has already come up with more use cases: an AI companion optimized for dementia that runs on speech-based interfaces and emotional and voice intelligence, adding more languages since dementia patients often revert to their first language, and using generative AI to create images of places for which no photos exist.
“We’ve always taken a very simple yet scalable approach. And this [project] has taken both of those: kept it simple for the user but made it infinitely scalable. And I think that’s where the real magic of AI is,” Elliott says. “AI is the key to us scaling, and we needed this collective help, this collaboration, in order to test those interesting hypotheses. And we’ve seen positive early results.”
For Memory Lane Games, that magic is not technology for technology’s sake. It is the possibility of helping more people reconnect with the stories, places, and people that make life feel familiar and spark joy.
MovingWorlds has supported more than 3,000 successful projects across 110 countries, unlocking more than $50 million in pro-bono consulting expertise for social enterprises. If you are a social enterprise in need of skilled support in AI or another line of business, apply for support on the MovingWorlds’ platform.
