Early detection · Primary care

Built for the appointments where memory loss is never mentioned — because nobody thought to ask.

The earliest signs of cognitive decline show up years before a diagnosis. We're building clinical AI that helps primary care see them sooner — while there's still time to act.

Primary care Software as a Medical Device Responsible AI UK-built
Cognitive function over time — illustrative
signal begins usually identified the window we open time / age cognitive function
Founder-designed and built
Building toward the UK medical-device pathway
Grounded in lived experience
The problem

Dementia is usually identified late — long after the first detectable changes.

By the time symptoms are unmistakable, the window for early planning, intervention, and support has already narrowed.

Caught too late

Dementia is often recognised only once symptoms are clear, when earlier, subtler signs were already there to be seen.

Primary care is under-equipped

GPs are the first point of contact, but lack practical, consistent tools to assess cognitive risk early in a routine appointment.

Time is the intervention

Earlier identification opens the door to planning, lifestyle change, and treatment — and the chance to live well for longer.

The evidence

The scale of the problem, and the size of the opportunity.

Our clinical rationale rests on published, peer-reviewed evidence and national data — not assumption.

982,000
People living with dementia in the UK, projected to reach 1.4 million by 2040.
Alzheimer's Society, 2024
1 in 3
People living with dementia have no formal diagnosis — the England diagnosis rate sits at 66.3%.
NHS England, Dec 2025
3.5 yrs
Average time from first symptoms to diagnosis — 4.1 years in early-onset dementia.
UCL meta-analysis, 2025
£42bn
Annual cost of dementia to the UK, forecast to reach £90bn by 2040.
Alzheimer's Society / Carnall Farrar
The Lancet Commission, 2024

Dementia risk is, in large part, modifiable.

The standing Lancet Commission on Dementia Prevention, Intervention and Care identifies 14 modifiable risk factors across the life course — including education, hearing loss, hypertension, depression, traumatic brain injury and, newly added in 2024, high LDL cholesterol and untreated vision loss.

Together, the Commission estimates these account for around 45% of dementia cases worldwide — the theoretical proportion that could be prevented if every one of them were eliminated.

Livingston et al., The Lancet, 2024
45%
of dementia worldwide is potentially preventable by addressing modifiable risk — which is why identifying risk early matters.

Figures are population-level estimates drawn from published sources and national statistics; the Lancet Commission's 45% is a population attributable fraction, not an individual risk estimate. We do not claim to prevent dementia. Our purpose is to help clinicians identify risk earlier, within routine primary care.

Our approach

The method is what we're protecting. What it's built to do isn't a secret.

Our underlying approach is proprietary and not public yet, for IP and regulatory reasons. Here's how to understand what it's designed for — and how it's meant to fit a clinician's day.

In the flow

Built for the appointment, not a referral

Designed to support assessment inside routine primary care, rather than adding another specialist hand-off.

More than one signal

A fuller picture of risk

Looks wider than any single test, to surface risk earlier and more reliably than one measure can alone.

Decision support

Supports judgement, built to standards

Built to inform a clinician's decision — not replace it — and developed toward formal medical-device standards.

Why it holds up

Founder-built, responsibly designed, and building toward regulation.

01

Conceived and built by our founder

Our platform was conceived, designed, and developed by our founder. The idea and the intellectual property are ours.

02

Grounded in the evidence

Our clinical rationale is built on the peer-reviewed literature, including leading journals like The Lancet.

03

Responsible AI by design

Transparency, clinical safety, and keeping a clinician firmly in the loop are built in from the start — not bolted on later.

04

Building toward regulation

The platform is designed to meet UK medical-device (UKCA/MHRA) standards, and we're seeking innovation support to take it through validation and the formal pathway.

05

Honest about the stage

Our platform is pre-market and research-stage. We're clear about what's proven, what's in progress, and what comes next.

Responsible AI

Built on seven ethical pillars.

Everything we build rests on the BEBA Ai Responsible AI Framework™ — seven pillars that turn responsible AI from principle into daily practice, across the whole AI lifecycle.

Governance & accountability

Clarity

Clear ownership of every AI decision, with traceable accountability for errors, bias, or misuse.

Data ethics & privacy

Purity

GDPR-compliant, consent-driven data use, with anonymisation by design and regular ethical audits.

Bias detection & inclusion

Fairness

Proactive bias testing across gender, race, and socio-economic groups, with inclusive design from the start.

Explainability & documentation

Transparency

Models that are explainable to users and regulators, with documented data, methods, and decisions.

Security & robustness

Resilience

Built-in security, stress-tested against adversarial and edge cases, with fail-safe recovery.

Compliance & standards

Integrity

Designed to align with the EU AI Act, UK guidance, and medical regulation, with independent review as a core commitment.

Human, social & environmental

Impact

Human-centred design with social value and sustainability built in, and real-world consequences monitored.

The framework

BEBA Ai Responsible AI Framework™

Our proprietary model — aligned with ISO/IEC 42001, IEEE 7000, and the EU AI Act.

What's next

Phase one finds the signal. Phase two becomes the platform.

Today, our focus is dementia and Alzheimer's. The same architecture is built to extend — across other chronic, long-term conditions, with a companion portal and support in any language. More to come.

Phase 2 · multi-modal

Sees, hears, learns

New input channels let the platform draw on more of each consultation — and extend beyond dementia to other chronic, long-term conditions. Sensitive capabilities are consent-first and private by design.

Phase 2 · companion portal

Clarity™

One secure place where people and their families find their information, guidance, and support together — continuity that lasts well beyond the appointment.

Global by design

Any language, anywhere

Built to translate across languages worldwide, so earlier insight isn't limited by language or geography.

Founder & team

Founder-led, engineered to be trusted.

Elizabeth founded Beba Ai and built its core platform. She brought in Dan as CTO to lead engineering — chosen for his credentials and a vision he shares completely. Together they're taking AI in digital health further: responsible, safe, ethical, and built around the person.

Elizabeth Shapton
Elizabeth Shapton
Founder & CEO · Beba Ai
30 years in engineering Safety-critical systems MSc AI · Distinction 15+ years in dementia research

For Elizabeth, dementia isn't an abstract problem to solve — it's personal. She has spent more than fifteen years immersed in dementia care, drawing on lived experience of what the condition means for the people living with it and the families beside them.

That experience shaped a single conviction: dementia should be met earlier, and always with dignity. This is a person living with dementia, never a diagnosis that defines them — and the sooner the signs are seen, the more life there is to protect.

It's also why the rigour matters. Thirty years engineering real-time, safety-critical systems — where failure isn't an option — and a Master's in AI, completed with Distinction, are what let her turn that conviction into clinical software built to be trusted.

Dan
Dan
Chief Technology Officer · Beba Ai
MSc AI & Data Science Systems & data architecture Architect of Clarity Responsible AI by design

Dan is chief technology officer of Beba Ai — an engineer with a Master's in AI and data science. Clarity™, the companion portal, is his: he leads its design and build, and engineers the data and systems foundations the platform runs on.

He builds the way Elizabeth does — meticulous, evidence-led, and exacting about the details that matter, with integrity in every decision. For Dan, innovation and responsibility aren't a trade-off: the most advanced system is the one people can trust. He shares the conviction completely — AI in digital health must be responsible, safe, and ethical by design.

Systems that matter — for people that matter.

Get involved

We're looking for the right partners early.

If any of this is close to your world, we'd like to hear from you — early collaborators shape what we build.

  • NHS teams interested in piloting
  • Investors and grant funders
  • Clinicians and academic advisors

We'll only use your details to get in touch about our work.

Beba Ai logo
Who we are

Responsible, safe, ethical AI.

AI for Humanity · Innovation with Integrity · Culture for Life.

Beba Ai builds responsible, applied AI for digital health. Our first platform helps primary care identify cognitive decline and dementia risk earlier — with ethics, transparency and safety at the heart of everything we build.

More about Beba Ai →