1,000 Pages, One Disease, and the Fight Most People Never See
How a personal story turned into a Digital Twin for people living with Multiple Sclerosis and why it's the project I'm proudest of.
My mother was diagnosed with Multiple Sclerosis when I was young. Back then, there was nothing I could do about it. I watched her live with a disease most people don't understand: unpredictable, invisible on the good days, devastating on the bad ones. And I watched her navigate a system that simply wasn't built for her.
That memory is a big part of why Mnemonic AI exists. So when I tell you the project we just finished is personal, I mean it literally.
The fight most people never see
Here's something that doesn't make headlines: MS doesn't have a powerful lobby. The people living with it are often too exhausted by the disease itself to take on the bureaucratic battles stacked on top of it.
And there are a lot of those battles. In Switzerland alone, thousands of MS patients struggle to get the disability classification they are legally entitled to. The medical files behind a single claim can run past 1,000 pages. Years of diagnoses, therapy reports, and insurance correspondence layered on top of each other. Insurers deny claims or undervalue them. And patients, already depleted, are expected to find the energy to push back.
Most don't. Not because they don't have a case, but because the case is buried in a thousand pages nobody has time to read.
Why we didn't think twice
We have people touched by MS in our closest Mnemonic family, so this isn't an abstract problem to us. When the Swiss Multiple Sclerosis Society (SMSG) asked whether we could help their caseworkers, we didn't have to deliberate. Honestly, we barely let them finish the question.
SMSG does the unglamorous, essential work of standing up for patients, independent counseling, medical guidance, and legal advocacy to help people secure fair treatment under the social insurance system. The bottleneck wasn't their expertise. It was time. Their caseworkers were spending hours, sometimes weeks, manually combing through enormous dossiers just to find the points that could turn a case around.
That's exactly the kind of problem our technology is built for.
What we built
Together with SMSG, we developed a version of our Digital Twin designed specifically for their caseworkers.
It reads an entire case file, and the latest letters from the insurer,and surfaces what matters most:
- Contradictions between the medical evidence and the insurer's decision
- Symptoms or administrative details that were overlooked
- The specific medical and legal grounds that can support a dispute
Work that used to take weeks now happens in minutes. And because every answer is grounded directly in the source documents and engineered to avoid the confident-sounding guesses that plague general-purpose AI, caseworkers can trace each flag straight back to the page it came from. When a wrong reading could affect someone's claim, that traceability isn't a nice-to-have. It's the whole point.
Two things mattered enormously to us here. First, the tool doesn't replace caseworkers. It clears the heavy lifting off their desks so they can do the part only humans can: fight for people, with empathy and expertise.
Second, privacy was non-negotiable. The Digital Twin runs entirely on-premises, on SMSG's own infrastructure. Patient data never leaves the building, never touches an external model, and never lands on a third-party server.
The metric that actually matters
We talk endlessly in this industry about KPIs. ROI. Efficiency gains. Cost savings.
Here's a different metric, and the only one I cared about on this project:
how many people finally got the support they deserved.
I didn't get into AI to optimize ad spend. I got into it for moments like this, where the technology meets a real human need and quietly changes someone's life.
What's next
The tool is in testing now with SMSG's advisory team, who are validating its findings against real cases. The goal is to have it in full operation by Q3 2026, screening every incoming case file so that more patients get faster, better-founded guidance on their claims.
If you know someone living with MS, share this with them. They deserve to know that people are building for them.
- Eliot Knepper, CEO, Mnemonic AI
