Common questions

We are a behavioural AI consultancy and product partner for digital health providers, including those delivering GLP-1 and obesity care. We help providers design, evaluate, and safely deploy the behavioural layer that makes patient-facing AI actually change behaviour, so they improve adherence, retention, and outcomes.

It is an intelligent layer between your platform and the patient. Grounded in the science of how AI can support human cognition and behaviour, it works out what type of support each patient needs and provides it when they are most likely to be receptive, so patients trust the AI, use it, and stick with it, which is what drives engagement, behaviour change, and outcomes.

A chatbot answers questions. Behavioural AI is the layer around and beyond it: the design and logic that decides what support a patient needs, when they need it, and how it should be delivered so it actually changes behaviour, safely. Most of our work is that layer, not the chatbot itself.

Both. We work from strategy through to hands-on design, delivery, and evaluation of AI systems, often as an extension of your team.

Digital health providers building or deploying patient-facing AI, especially in GLP-1, obesity, and behaviour change. All of our current client work is in GLP-1 and weight management.

Much of the benefit of GLP-1 medicines depends on continued, well-supported use, and in real-world programmes most patients stop within the first year for a mix of reasons, many of them behavioural. We design the behavioural layer that adapts across the whole journey: from treatment initiation and onboarding, through the first weeks and the management of side effects, through the pauses and dips when life gets in the way, on to reaching a goal and into long-term maintenance, whether a patient stays on a maintenance dose, steps down, or stops. At every stage it tailors support to the individual, a true n-of-one experience, and we evaluate what actually moves adherence, persistence, and long-term maintenance.

Yes. We run independent behavioural and clinical evaluation, including PromptSafe simulation testing, to find safety, quality, and behavioural gaps before or after launch.

PromptSafe is our AI safety and quality testing tool. It runs offline simulation testing of conversational AI agents across hundreds of realistic and adversarial conversations to find where they fail users before real users do.

FAST is our published framework (Frontiers in Digital Health, 2025) for evaluating conversational AI health coaches across four domains: fidelity, accuracy, safety, and tone.

Yes. We design and run clinical and behavioural research, from study design to peer-reviewed publication, to build credible evidence of effectiveness, safety, and impact.

That is what the Before You Build method is for. Five uncertainties decide whether a health technology creates value in practice: whether the people it is built for will use it, whether it fits real work and who acts on its output, what should happen once it knows something, who benefits and whether they are the ones who pay, and what regulation and governance require of it. Model performance is rarely the one that settles it. We resolve those questions through an Evidence Phase, before the budget is committed and the answer gets expensive.

Most engagements start with an Evidence Phase: a fixed fee piece of work, eight to twelve weeks, that resolves whether the thing you are planning is worth building. It applies the Before You Build method and ends with a recommendation and the evidence behind it. From there, engagements are scoped to your goals, from delivery alongside your product team to evaluation or research programmes. The best next step is usually a short discovery call.

An unusual combination in one place: decades of clinical and behavioural science credibility, peer-reviewed research, hands-on AI delivery and safety, and genuine independence.

Yes. We own no platform or app, so our advice is based on what works, not on what we have to sell.

We are a UK-based consultancy (Capital Consult Ltd, trading as Sacher AI), working with digital health teams internationally.

Book a short discovery call. We will work out whether there is a fit and point you to a sensible first step. If you would rather read first, the Before You Build page explains the method, and the playbook goes further in twenty-three pages.