Work — Riff Apps

Live on iOS and Android

Riff

Most social apps ask you to judge a photograph. Riff asks twenty‑five questions instead — about values, goals, perspective and how you communicate — then matches people on how they think rather than how they look. Conversation starts in text and voice; faces appear only when both people tap ready, at the same moment, so nobody holds the power in the exchange.

Two modes run in parallel: a one‑to‑one Deep Connection aimed at mentors and collaborators, and a four‑person Friend Circle built around shared ambition. AI companions keep new users company while real matches are found, and they are labelled as AI throughout.

What made it hard

Trust. The product only works if everyone is real, so it carries government ID and live selfie verification, liveness checks with blink and head‑turn capture, a visible trust score, encryption of voice and media in transit and at rest, and a layered detection system for harmful behaviour. Subscriptions, App Store and Play billing, and the full public policy set shipped with it.

React Native Compatibility modelling ID verification Voice Subscriptions

Visit riff-app.co.uk

Questions answered 25 of 25

Compatibility 4‑layer score

Voice note 0:42

Verification Green

AI companion Labelled

Reveal Both ready

Live on the web

TeachWise AI

A curriculum for whatever you need to learn next. Instead of a catalogue of fixed courses, TeachWise assembles a structured path around the thing a learner is actually trying to do, then adapts it as they work through it.

Generating a single good lesson is easy now. Generating forty that build on each other — without repeating themselves, contradicting each other or assuming knowledge that hasn’t been taught yet — is the real problem, and it is where most of the engineering went: prerequisite graphs, sequencing rules, consistency checks between modules, and evaluation of generated material before a learner ever sees it.

Next.js Curriculum generation Adaptive sequencing Content evaluation

Visit teachwise-ai.com

Learning goal Set

Curriculum built 9 modules

Prerequisites Checked

Progress

Next lesson Ready

In build

Project controls for the energy sector

Capital projects in energy live or die on cost, schedule, change and risk information — which is usually spread across a dozen spreadsheets, three reporting tools and the memory of one very experienced planner. Our current build brings that into one application: a baseline that holds, progress tracked against it, and variance explained in language a steering committee can act on.

AI does the reading and the drafting. It parses contractor progress reports, reconciles them against the plan, flags drift early, and writes the first version of the monthly narrative. What it does not do is decide: every figure stays traceable to its source document, and every forecast is reviewed and signed off by a named person before it goes anywhere.

Where it goes next

The core — baselines, earned value, change control, risk and reporting — is common to any capital programme. Energy is first because the reporting burden and the consequences of drift are highest there. Construction, utilities, infrastructure and manufacturing follow, and we are speaking with teams in each about pilots.

Energy Cost and schedule Earned value Change control Document parsing

Ask about an early pilot

Cost performance CPI 0.97

Schedule performance SPI 1.02

Float remaining 4 weeks

Open changes 12

Risk exposure

Monthly narrative Awaiting sign‑off

Your project could be next

We take on a small number of builds at a time so the same senior team stays on each one from scope to launch.

Start a project