Digitisation Doesn’t Automatically Equal Transformation”: A HealthTech Builder on Why African Pilots Keep Failing to Scale

Dr. Agbor Ashu has spent over a decade watching health technology meet the reality of African health systems, first as the builder of GiftedMom, an automated SMS and voice messaging platform that reached over 600,000 pregnant women and new mothers, and later as an advisor examining why so many promising HealthTech pilots never survive past the pilot stage. We spoke with him about the specific failure patterns he’s identified, and what he now tells investors and governments to test before believing a pilot will scale.

You’ve spent over a decade watching HealthTech meet the reality of African health systems. What first made you start noticing the gap between what technology can do and the assumptions it depends on?

I first began noticing it while building GiftedMom. The initial idea seemed simple: use mobile technology to provide pregnant women and new mothers with timely health information and reminders. But once we started implementing it, we realised that sending the message was the easy part. Did the woman own the phone, or was it shared with her husband? Could she read the message? Would she trust it? If she was advised to visit a facility, could she afford the transport? And when she arrived, would the service or vaccine actually be available? The technology might perform exactly as intended and still fail to produce the expected health outcome. That experience changed how I looked at HealthTech. I stopped seeing the technology as the whole solution and started paying more attention to the chain of assumptions between the technology and the outcome. Usually, that chain is where the real risk sits.

Of the failure patterns you’ve outlined, which do you see most often?

My main base is Cameroon, so I’d be careful not to speak as though every part of, say, the Nigerian health system behaves the same way. But the pattern I see most often is this: each tool may solve a legitimate problem for the organisation that introduced it, but nobody has removed anything from the health worker’s workload. The result is not always digital transformation. Sometimes it is simply digital accumulation.

You’ve described nurses entering the same immunisation data three times, across a paper register, a tablet, and a WhatsApp group. What’s the actual fix, and why doesn’t it happen more often?

The fix is not just at the level of the healthcare worker. It’s largely a political, governance, and coordination issue. The paper register may be required by the national programme. A donor-funded project introduces a tablet with its own indicators. The district requests updates through WhatsApp. The nurse is simply at the point where all these requirements meet. I would not prescribe another application. The starting point is to agree on a phased digital transformation plan, not another pilot that runs for a year and leaves the existing reporting system unchanged. From the beginning, the digital platform should be designed to integrate with the government’s reporting system, like DHIS2, so aggregated data flows through without being entered again. That should be a core requirement from day one, not something considered after the pilot succeeds. On our Wellness Pass project in Mauritania, we worked within the existing government reporting structure rather than creating a separate system, and it improved reporting timeliness and reduced duplicate data entry by about 75%. Why doesn’t this happen more often? Because every programme has its own reporting priorities, timelines, and funders. It’s easier to launch a standalone pilot than to do the slower work of aligning indicators, systems, and government requirements.

When a HealthTech company comes to you with a government MoU or a successful pilot, what questions do you ask to test whether it will actually translate into sustainable financing?

I usually begin with one simple question: who will pay when the pilot ends? Then I go deeper: is there an actual budget line, or only expressed interest? Who owns and controls that budget? What procurement route would let the government purchase the solution? What existing cost or process will it replace? Who pays for implementation, devices, connectivity, training, and maintenance? Are people still using it when the pilot team is no longer physically present? I learnt this firsthand with GiftedMom. We had an MoU with the government, but in practice it mainly gave us authorisation to work within public health facilities. We still had to raise the funding ourselves, implement the programme, and report results back to government. There was no clear transition from that support to government financing, and eventually the model was no longer sustainable. We had to shut it down.

If you were advising an investor evaluating an African HealthTech company, what’s the first assumption you’d tell them to stress-test?

Two connected ones: is the problem painful enough, and who will pay to solve it? Is this addressing a real painkiller or offering a vitamin? How are people managing the problem today, and what is that workaround costing them, in time, money, or patient outcomes? Then the payment assumption: who pays, why will they pay, and what evidence shows they’ll keep paying? In healthcare, the beneficiary, user, buyer, and gatekeeper are often different people entirely. A company that treats all of them as one “customer” can easily mistake usage for willingness to pay.

Where have you seen this done right, something built for the health system that exists, that then helped change it?

Our maternal and child health work in Mauritania. We didn’t assume every facility was ready for a completely new digital ecosystem. We worked through existing facilities, government structures, and DHIS2, focused on a defined set of indicators, and designed the intervention around how care and reporting were already organised. The technology didn’t sit outside the system and ask it to adapt. It connected to structures people already recognised, while reducing duplication within them. Reporting timeliness improved significantly, and duplicate data entry dropped by about 75%.

What’s one HealthTech assumption you think costs Africa the most right now?

That digitisation automatically equals transformation. It does not. If we digitise a fragmented workflow without redesigning it, we create digital fragmentation. If every organisation builds its own closed platform, we end up with more systems, not a stronger health system. There’s a direct financial cost, money spent on platforms that duplicate one another or can’t exchange data. But there’s also a human cost: health workers grow frustrated because technology adds administrative work without improving care, governments become cautious after pilots that never scale, and patients keep repeating their histories because their information can’t move with them. Perhaps the greatest cost is the loss of trust. Once health workers experience digital health as another burden, even genuinely useful solutions become harder to introduce.