Digital Identity Demo
I turned a biometric system custom-built for a single client into a generic demo: a consultancy uses it to show, on a phone, how a person is enrolled and recognized.
- Client
- Technology consultancy (under NDA)
- Type
- Commercial proof of concept
- Period
- 2025 – 2026
- Status
- Completed
Project figures. The client is protected by a confidentiality agreement.
What wasn't working?
The technology already existed and worked, but it was built around one client's rules. To sell it to others, it had to be shown, and as it stood, it couldn't be.
- The whole system was tied to a single company's transactions and rules
- A demo with fake data doesn't convince anyone: evaluators need to see a real biometric match, including a rejection
- It had to run on an ordinary Android phone, in an in-person meeting
- It couldn't depend on the original client's systems or send them any data
What I built
I separated the biometric capability from everything specific to the original client and packaged it into two walkthroughs that can be demonstrated in minutes.
Complete enrollment
Registering a person from scratch: face, ID document, face-to-document matching, signature, fingerprint, and a summary of the captured data.
Face login
A Face ID-style login: a reference face is enrolled, then a live capture is compared against it to grant or deny access.
Liveness detection
The face capture asks for a head turn, to rule out someone holding up a photo instead of being present.
Automatic ID reading
Detects the document type and extracts the number, full name, and nationality, each with a confidence score.
Fingerprint two ways
With the phone's camera or with an external Bluetooth fingerprint reader, depending on what the prospect wants to evaluate.
The rejection demo
If a different person tries to authenticate, the system detects it and shows it. That's the proof the matching is real and not scripted.
What changed for the client?
A build made for one client became a reusable commercial asset, without losing a single biometric capability along the way.
- The transaction engine went from 741 to 42 lines (−94%) and about 2,800 lines of custom logic were retired
- Two complete scenarios, demonstrable end to end in a single meeting
- Sends no data to outside systems: it can be shown to any prospect without risk
- Adaptable identity: name, colors, and branding were fully swapped for the consultancy
Technology used
Have a similar project in mind?
Let's start with a free Discovery Session. In 5–8 minutes I'll pin down exactly what your project needs and give you a straight proposal.
Start a Discovery SessionFree · No strings attached · I reply personally