The traders are not hiding. They have phones.
Every African revenue authority has its plans for closing the informal sector gap. Most of those plans have failed for the same reason: they were designed for an economy of storefronts and accountants. The economy now runs on a phone. The mobile money rail is the route to taxpayer reach. eTIMS-M-Pesa is the demonstration.
The Kenya Revenue Authority's Commissioner for Micro and Small Taxpayers, George Omondi Obell, has become one of the recognisable architects of the authority's informal-sector strategy. The strategy is structurally different from what came before. Older approaches tried to extend the formal-sector compliance model — registered businesses, accountants, periodic filing — to the informal sector by simplifying the forms and reducing the rates. The strategy did not work because it was solving the wrong problem. The informal trader was not failing to comply with a slightly-too-complex form. The informal trader was operating outside the system the form belonged to.
The mobile money rail changed this. As of 2026, M-Pesa, MTN MoMo, Airtel Money and the other major mobile money networks settle a meaningful share of all retail transactions in Kenya, Tanzania, Uganda, Rwanda, Ghana, Nigeria and the Côte d'Ivoire–Senegal corridor. The trader who refused a paper invoice will, when prompted by the customer's phone, accept a Lipa Na M-Pesa till payment. The transaction is recorded — by the mobile money operator, by the customer's phone, by the till the trader uses to receive funds.
KRA's response — the integration of eTIMS with M-Pesa — is the structural move that follows. When a customer pays via Lipa Na M-Pesa to a registered trader, the transaction can be captured into eTIMS in real time. The trader's tax position becomes a function of the till's traffic. The trader does not need to file an invoice. The customer does not need to ask for a receipt. The Authority's compliance system catches the transaction at the moment of settlement. From the KRA Commissioner Obell's public statements, this is the architecture KRA is building toward — 'real-time monitoring and advanced analytics' that 'flag inconsistencies as they occur, rather than months or years later.'
The same logic is unfolding across the continent at different speeds. Rwanda's RRA has run the Electronic Billing Machine programme for years; the EBM-MoMo integration is the equivalent move. The RRA's '10 percent of the VAT you paid' reward for EBM-receipted purchases — accessible through *800# on any mobile phone — is the demand-side mechanism that pulls traders into the system through their own customers. FIRS in Nigeria has the FIRSMBS e-invoicing platform that, integrated with NIBSS, will reach a similar point. The Tanzania Revenue Authority's electronic fiscal device integration with mobile money is following the same shape.
The compliance unit a trader could ignore was the inspector who arrived once a year. The compliance unit the trader cannot ignore is the till that pays them.
The technology challenge is not the integration. The integration is, in the main, solved. The challenge is what the authority does with the data once it has it. A real-time view of millions of micro-transactions across hundreds of thousands of traders produces a data problem that is structurally different from the income tax return review the Domestic Taxes Department was built around. Each trader's data is small. The volume of traders is enormous. The compliance question for each is the same — does this trader's mobile money flow, properly interpreted, support the declared position; if not, what is the additional assessment; and is the additional assessment defensible at the level of a single trader who may not have professional representation.
This is where the project either works or produces the same backlash the Intelligence Analysis Tool reporting generated in March 2026. A real-time compliance system that produces well-reasoned assessments at trader-level scale is the system that closes the gap. A real-time compliance system that produces poorly-reasoned assessments at trader-level scale is the system that destroys the political consent the project rested on.
Where each sits.
Akki is the substrate the integration runs on. The eTIMS-M-Pesa interface, the FIRSMBS-NIBSS link, the RRA EBM-MoMo channel, all sit on Akki. Mobile money transaction data enters the substrate structured. The trader's iTax or equivalent profile is the unit of analysis. Akki produces the data view that the compliance officer or the model reasons over.
Solva does the reasoning at trader level. The compliance question for a single trader is small but real. The trader's M-Pesa till settled KES 2.4 million last quarter. The trader's most recent return declared KES 380,000. The eTIMS invoice issuance was sporadic. Solva structures the case the way a competent compliance officer would. It restates the question — is the trader's declared position inconsistent with the mobile money flow. It surfaces what the data supports — the till receipts, the invoice issuance pattern, the prior-period filings, the trader's category and sector benchmarks. It triangulates between the trader's data and what the system knows about similar traders. It produces a recommended action — accept the return, issue a soft compliance prompt, raise a default assessment, or refer for investigation — with the basis explicit. When the data does not support an assessment, Solva refuses to recommend one. The refusal behaviour does specific political work at this scale. An authority that issues hundreds of thousands of automated assessments based on mobile money flow without trader-level reasoning will produce, on a non-trivial fraction of files, assessments that are wrong. Those wrong assessments will reach the political offices of MPs whose constituents are affected, and the Twitter accounts that produced the IAT backlash in the first place. The authority that issues fewer, better assessments — and lets the others go to soft compliance prompts that the trader can act on — protects both its yield and its consent.
SyniSense holds the data sovereignty position for the mobile money interface. M-Pesa transaction data is held by Safaricom under the Data Protection Act and the Communications Authority of Kenya's regulatory framework. The data the authority receives is governed by cooperation arrangements and is subject to the ODPC's oversight. SyniSense ensures that the AI reasoning happens against anonymised flows, with re-identification only inside the authority's environment for the cases that are escalated.
The informal sector compliance programme changes shape. It is not an annual census of traders. It is a continuous compliance signal on the traders the system has visibility on, with assessments raised only where the case is defensible. The yield rises faster than the political cost.
The trader's experience changes. The compliant trader receives no inspector. The non-compliant trader receives a prompt that says, in concrete terms, what their data shows and what they need to do. The prompt either produces a corrected return or escalates to an assessment that, when it comes, is reasoned.
The Commissioner for Micro and Small Taxpayers' presentations to the National Treasury change. The story is not 'we plan to bring the informal sector into the tax net.' The story is 'we are bringing in the informal sector, here is what we have collected from it this quarter, and here is the trader-level evidence behind it.'
And the conversation with civil society changes. The ODPC, the consumer protection organisations, the trader associations — all are looking for evidence the authority is doing this with discipline. The reasoning trails are the evidence.