Human-centred,
system-aware
Services rarely fail at the centre. They fail at the seams — where a policy meets a form, a form meets a counter clerk, and a counter clerk meets a family that has already travelled two hours.
This page sets out how we work: the five stages of an engagement, the standards we hold to, the methods we use, and what we ask of a partner before we start.
Design as a public capability, not a vendor service.
We draw on service design as practised in public institutions, participatory and co-design traditions, and systems thinking — then adapt them for the conditions we actually work in: thin budgets, high staff turnover, procurement cycles longer than a pilot, and real accountability to the public. Method is negotiable. The commitment to evidence from the people a system serves is not.
Five stages, and what each one produces
Stages overlap in practice, but each has a deliverable a partner can hold, review and act on without us in the room.
Frame the problem with the people who own it
2–3 weeksBefore solutions, we agree on what is actually broken. Framing workshops bring citizens, frontline staff and decision-makers to the same table, and the disagreement in that room is the finding. Most briefs we receive change shape in the first three weeks — a request for an app becomes a question about how a notification reaches someone with no smartphone.
A written problem statement and system map both parties sign off on.
Research in context, in language
3–6 weeksFieldwork where the service is used: homes, clinics, schools, counters, queues. Contextual interviews and observation in the language people speak, with deliberate sampling for the users a system already fails — women, people with disabilities, informal workers, families without documents. Frontline staff are researched as users too; they carry the workarounds that keep a broken service running.
Journey maps, evidence library and a findings session run for the partner's own team.
Design the whole service, not the screen
4–8 weeksAn app cannot fix a process it sits on top of. We blueprint front stage and back stage together: what the citizen sees, what staff must do, what the policy permits, and where the data actually moves. Forms, letters, SMS wording, signage and staff tooling are design work — and usually the highest-leverage part, because they can change without a procurement cycle.
A service blueprint, redesigned artefacts and a prioritised change list.
Prototype small, measure honestly
6–12 weeksWe test at the lowest cost that produces real evidence — paper, a pilot desk, one union council, one district — and we report what failed. Operational measures (time to serve, drop-off, error rate, cost per case) sit next to user measures, so a partner can defend a decision to scale, adapt or stop in front of their own leadership and auditors.
A pilot report with measures, costs and an explicit scale / adapt / stop recommendation.
Leave the capability, not the consultant
Built in from week oneEvery engagement ends with a hand-over designed from the start: research repositories, blueprints and playbooks the partner owns outright, plus named staff trained to run the next round. We work alongside a partner's team rather than in parallel to it. The measure of success is the second project we are not needed for.
Owned artefacts, a trained internal team and a standing offer of advice.
What we hold to
Six commitments that survive contact with a difficult project. If an engagement cannot honour them, we say so before we start.
Nothing about people without people
Affected communities take part in framing and in judging what worked, not only in being interviewed.
Design for the hardest case first
If it works for a woman with no phone, no documents and no literacy, it works for everyone else too.
Respect the frontline
Staff are not the obstacle. Their workarounds are evidence about the system they were handed.
Consent, dignity, data care
Informed consent in plain language, no extractive research, no identifiable data shared without permission.
Publish what we learn
Methods, tools and findings released for others to reuse, with partner agreement and identities protected.
Say when design is the wrong tool
Some problems need funding, staffing or a change in policy. We will tell you that instead of running a workshop.
A first antenatal visit, stage by stage
A pilot with a district health office, currently in week 6 of 12. It shows how a brief moves from a request for an app to a change in a referral slip.
The department asked for a reminder app. Framing with lady health workers showed the drop-off happens before any reminder would land.
Home interviews in two districts, plus shadowing at facility counters. Travel cost, permission from family and unclear timing outweighed forgetting.
A rewritten referral slip that names a date, a person and what the visit costs — plus a hand-off script for health workers.
Running at two facilities against a matched pair, measuring first-visit attendance and time from referral to arrival.
Slip templates, the script and a short research playbook go to the district office in September, with two staff trained to run the next round.
Ways to work with us
Not every problem needs a full engagement. Three shapes, and the one we suggest depends on how settled your problem statement is.
Diagnostic sprint
Rapid framing and light fieldwork to establish what the real problem is and whether design can move it. Ends with a written recommendation you can take to your board.
Best when the brief is still a hunch.Full engagement
All five stages with an embedded team working alongside yours, through research, service design, a measured pilot and hand-over.
Best when a service is failing and you can commit staff time.Capacity building
Training and coaching for your own team in research, service design and facilitation, with our designers on call while they run their first project.
Best when you want the capability in-house.What we ask of a partner
These are the conditions that decide whether a project changes anything. We would rather agree them in the first meeting than discover them in month four.
Start a conversationA named owner with authority. Someone who can change a form, a script or a process without a six-month approval chain.
Access to users and frontline staff. Permission to spend time in facilities and homes, not only in the head office.
Staff time, not just budget. Two or three people from your team in the work, so the capability has somewhere to land.
Room for the answer to be uncomfortable. If research says the plan is wrong, we will report it, and you keep the option to act.
Permission to publish the method. Findings can be anonymised, but the learning should be reusable by others working on the same problem.
Tell us about the service you want to get right.
We will tell you honestly whether design is the right next step, which of the three shapes fits, and what it would take from your side.