Our Work
What this looks like in practice
Two projects, both built by The Good Labs. Different shapes — same methodology. Real numbers, written by the people who wrote the project plan.
ComeThru
A mental wellness companion that meets people where they already are — on WhatsApp.
The Good LabsThe problem
Mental health systems in East Africa are clinic-bound, English-first, and gated by stigma. Most people who need support never make a first appointment. Those who do face waitlists measured in months. Telehealth solutions assume a smartphone, an app store, an appointment, and the language of the platform — assumptions that exclude the majority of the population that mental-health frameworks claim to serve.
What we did
We built where people already are. WhatsApp is the de-facto operating system for communication across the continent — it works on entry-level phones, it doesn't require an install, and people are already comfortable inside it. ComeThru is a companion that lives inside that channel, designed in close collaboration with therapists, peer counsellors, and people with lived experience of the conditions it supports.
The product isn't a chatbot trying to replace a therapist. It is a triage layer, a check-in companion, and a structured set of self-guided exercises grounded in protocols clinicians already use. When someone needs a human, ComeThru hands them off — to a peer counsellor for some classes of need, to a referral pathway for clinical cases, to crisis resources when safety is at stake.
Privacy was a first-order design constraint. Conversations are end-to-end encrypted by the channel itself; the companion stores only what is necessary to provide continuity, and discloses that storage in plain language. Language and tone were calibrated by a working group that included Sheng-speakers and rural users — not just a Nairobi product team.
We're rolling ComeThru out in phases against a research protocol. Early indicators we're tracking: completion rates of self-guided modules; time-to-first-handoff for cases that warrant human escalation; user-reported usefulness across age, language and rural/urban segments. Outputs we are explicitly not optimising for: total messages exchanged, daily active users, or any other engagement metric that risks becoming a proxy for product success.
What we learned
- WhatsApp is not a workaround. For most of our users it is the canonical app for communication; designing for it primarily — not as a fallback — surfaced design decisions a smartphone-first product would have missed.
- Peer counsellors are an underused resource. Triage that routes appropriate cases to trained peers (rather than queuing everything for clinicians) shortens time-to-first-help by an order of magnitude.
Maji Maisha
Solar-powered water systems owned and run by the community.
The Good LabsThe problem
Mbeere North is semi-arid. The closest reliable water for many families was a four-hour round trip, almost always walked by women and girls. Boreholes did exist, but they ran on diesel, and diesel prices in this part of Kenya are volatile. When fuel costs spiked, pumps stalled, communities reverted to walking, and the previous decade of donor investment was effectively erased.
What we did
We co-designed three solar-powered hubs with the community. Solar removes the diesel dependency — once the panels are up, the marginal cost of pumping water approaches zero. But hardware was the smaller half of the project. The harder half was governance.
From day one we negotiated a community-ownership transfer schedule with a stated end date. Eighteen months from commissioning, AnduBer would step out and the community board would step fully in. That schedule wasn't aspirational — it was the contract. We've held to it.
The board itself was structured deliberately: 50% of seats reserved for women's representatives. Women carry the water in this region; ownership of the water system without women in governance was a structural absurdity we refused to entrench. The seats are not advisory; they are voting.
Maintenance was localised. We trained a cohort of solar technicians from within the community, sourced spares from a regional supplier with a maintenance contract, and documented every failure mode we'd seen during the prototype phase in a maintenance playbook the board owns.
Three years on: pumps still run, costs are down 75% versus the diesel baseline, women's leadership representation has held above 50%, and the next two communities in the county are scoping replication using the same governance template.
What we learned
- A community-ownership transfer schedule with a date on it is more credible to funders, easier to manage, and structurally fairer to communities than vague "capacity-building" framing.
- Quotas for women's seats are not a side benefit. They were the difference between a board that knew when the pumps had to run and a board that learned about it from a complaint.
- Infrastructure projects are governance projects with hardware attached. We sequence the social architecture before the engineering and the cost curves bend in the right direction.
What’s next
More to come
We’re working on projects across mental health, climate adaptation, community-owned infrastructure and informal-economy livelihoods.
Case studies land here when the work is mature enough to write about honestly — including the parts that didn’t work the first time.
Bring us a projectLet’s talk
Let’s build new worlds
Tell us what you’re working on.
info@anduberinnovate.space
Phone
+254107025817
Headquarters
Nairobi, Kenya
Global reach
/company/anduber
Or write to us directly. We answer everything we receive.