Separate requirement ownership
Two kinds of requirements will land on your program, and conflating them is how banks end up building things nobody asked for. External requirements come from the card scheme and the CBN, as well as from your processor and the certification body. Internal requirements are everything you decide about architecture and configuration, along with controls and staffing for service management.
External requirements are verifiable. The Guidelines on Operations of Electronic Payment Channels set out stakeholder obligations in writing. The ₦100 million minimum share capital for a PTSP license is published. Where a requirement is external, cite the document and the clause. Where it's internal, say so and put a name next to it.
Then apply one rule across the whole program. Every readiness claim carries a source and an owner, or it doesn't go in the status report. "The terminals support it" leaves the owner and the source blank. "Build 4.2.1 passed sandbox certification on 14 March, evidence in ticket PAY-2291, owned by the terminal engineering lead" is a claim. The second kind survives an audit, and the first kind collapses the week before go-live.
Build the workstream matrix
Nine workstreams cover a card scheme rollout, and each one needs a lead and a deliverable, with a dependency list attached. Participation and approvals. Acceptance channels. Terminals and estate management. Host and processing. Security and keys. Testing and certification. Operations and disputes. Settlement and reconciliation. Merchant and cardholder support.
Run them in parallel, but connect them explicitly. Testing can't start before keys are exchanged. Settlement design can't close before the processor confirms file formats. Support can't be staffed before scope is frozen. Draw those arrows on one page and keep that page in front of the steering committee every week.
Owners and evidence
For each workstream, the matrix holds an accountable owner and required documentation. Architecture evidence and vendor confirmation sit with configuration proof, and open questions fill the last column. The last column is the one that matters. A workstream with no open questions three weeks before a pilot is either finished or not being examined honestly.
Evidence has to be artifacts. A signed scope letter from the processor. A screenshot of the routing table. A certification pass report with a reference number. Interface specifications with version numbers. Anything a new engineer could read in an hour and understand what state the program is in.
Pilot acceptance criteria
Define entry and exit criteria as numbers before the pilot starts, because defining them afterward turns into negotiation. Entry means approvals granted and connectivity proven on both aggregator paths, with certification passed and keys loaded, then monitoring live and the support escalation path tested with a real call.
Exit criteria are measured:
-
Authorization success rate at or above your existing scheme baseline on the same merchant set
-
Every reversal and timeout completed within the agreed window, with zero unexplained exceptions
-
Settlement files reconciled to the last naira across the full pilot period
-
Dispute cases resolved inside the 72-hour regulatory window
-
Production volume held at the agreed cap until all four above hold for a defined number of consecutive days
Hit them and you expand. Miss one and you fix it before adding merchants.
Scope engineering support
The engineering work described above is specialized and time-boxed, which makes it badly suited to being squeezed around a core banking upgrade. Energize Global Services (EGS) has built banking and payment systems since 2007 from delivery centers in Yerevan and Sofia as well as Tbilisi, and the work covers core banking platforms and payment infrastructure along with fraud and anti-money-laundering systems for financial institutions.
What that means for a card program is concrete. Architecture assessment against your intended scope. Card processing and switch-side development. POS application and estate changes. Security and key management design. Test planning and execution support through certification. Production monitoring and hypercare once the pilot opens.
Worth stating plainly: EGS is not a scheme member and does not sell a ready-made certified connector. Scheme participation and BIN allocation come from the scheme and the CBN, as does certification. The engineering that gets your systems to the point where certification is winnable is where an external team adds capacity.
Plan the next step
Four answers determine the shape of your program. Which role you're taking. Which acceptance channels are in the first release. What your current processor and switch environment actually looks like. And where you're drawing the pilot boundary. Write them down in a page, and the engineering scope for your card rollout becomes a planning conversation instead of a discovery exercise. Discuss your AfriGO rollout with EGS and map the engineering work from current systems to pilot readiness.