A diverse team collaborates in a high-tech control room, analyzing offshore energy operations and AI dashboards.
Seplat Energy's CEO has told African oil and gas operators to move faster on AI. For Microsoft 365 administrators, the more interesting part is how Seplat says it deployed Copilot: training before licences, named business owners for each use case, and a cross-functional governance committee. Every Seplat claim below comes from The Fact Nigeria's report of the conference. It is company-reported and has not been independently verified.

A diverse team collaborates in a high-tech control room, analyzing offshore energy operations and AI dashboards. What was said, and where​

Effiong Okon, CEO of Seplat Energy Plc, called on African energy operators to speed up their adoption of AI and digital technologies. He argued this would improve operational efficiency, strengthen asset performance and speed up decision-making. He delivered the keynote at the inaugural Digitalisation and AI in Energy West Africa conference in Lagos.

The official conference agenda puts the one-day event on Tuesday, 17 November 2026. It lists Okon's keynote and a dialogue called "Operator Address: What It Actually Took to Put AI Into Production." Seplat's General Manager of IT, David Osemwegie, is among the dialogue's speakers. The event organisers frame the conference as "From Pilots to Production." The agenda confirms the event, date, speakers and session topics. It does not confirm what was said or any of Seplat's figures.

The Microsoft 365 Copilot details​

Two statements in the report matter most to IT administrators:

  • Okon said Seplat has rolled out Microsoft 365 Copilot to improve individual and team productivity, with an AI Ethics, Risk and Compliance Committee overseeing responsible use.
  • Osemwegie said employees must complete basic activation training before receiving Copilot licences. He said this has delivered consistent adoption of more than 90 per cent.

The report does not define "adoption." It could mean licence activation, regular use, or use among trained staff only. It gives no denominator, licence count, reporting period or usage telemetry. The 90 per cent figure is therefore Seplat's own measure. It does not show that the training caused the result.

The report also does not say which Copilot features, apps, licence types or tenant data-protection settings Seplat uses. I won't guess at them.

Governance and ownership​

The governance model is the most transferable part. Seplat's AI Ethics, Risk and Compliance Governance Committee brings together legal, human resources, finance and risk representatives. The report gives no charter, approval workflow or risk framework. We can't say how the committee works in practice.

Osemwegie also described a division of responsibility:

  • Asset leaders are accountable for AI-driven outcomes.
  • Every use case gets a business product owner. That person is responsible for its roadmap, performance, development and adoption.
  • Business teams answer for operational outcomes, and IT manages technical performance and development.
  • Vendors are implementation partners.

This is a familiar pattern. Many Copilot rollouts stall when IT owns the licences but nobody in the business owns the results. Seplat's reported model puts accountability for outcomes with the business. That is a sensible approach, though the report offers no measured results to back it up.

The operational example​

Okon's one quantified result is not about Microsoft. He said machine learning improved permeability prediction at Sapele, with accuracy rising from roughly 50 to 60 per cent to 80 per cent or better. The report gives no sample size, validation method, time frame or definition of accuracy, so treat it as a company claim.

He also described seismic reprocessing to improve subsurface imaging. He mentioned planned ocean-bottom-node seismic acquisition as well. That work is a plan, not a completed project, and the report gives no timing, location or cost.

Okon's argument about energy supply is attributed to him and is not a forecast. He said AI needs reliable power, so growing data-centre demand could benefit Nigeria's gas industry.

Data foundations and legacy systems​

Osemwegie drew on Seplat's integration of the former ExxonMobil shallow-water asset. As reported, he said it required:

  • consolidating business data;
  • managing different technical systems;
  • addressing legacy inefficiencies;
  • keeping employees and business owners involved.

He tied this to strong data governance and business continuity. It echoes Okon's three priorities: start with a defined business problem, build data foundations, and bring people along. Okon also said AI should support engineers' judgement and not replace the engineering discipline needed to run assets safely.

What admins can take from it​

  1. Gate licences behind training. Seplat reportedly requires basic activation training first. Whether that explains its adoption figure is unproven.
  2. Define "adoption" before you report it. Say whether you mean assigned, activated or regularly used. Seplat's 90 per cent can't be compared with anyone else's without that definition.
  3. Assign a business owner to every use case. The reported model keeps IT responsible for technical performance and the business responsible for outcomes.
  4. Set up governance early. A committee spanning legal, HR, finance and risk gives Copilot decisions an organisation-wide owner.
  5. Measure the change in decisions, not just usage. Okon said the point is the decision a tool changes. Seplat has not disclosed any Copilot business-impact numbers.

Caveats​

This is a speech report from one company, covered by one outlet. The Fact Nigeria's article is the only source for the detailed claims, and the conference materials corroborate only the event and speakers. The headline speaks for African operators generally, but the examples and figures are Seplat's alone. Other conference speakers, according to the report, stressed linking digital spending to measurable commercial outcomes. Seplat has not published evidence that its Copilot rollout delivers such outcomes.

The Copilot lessons are the transferable part, and they are a useful checklist. Treat the 90 per cent figure as an unaudited claim until Seplat says what it measures.

 

References

  1. Seplat Urges African Energy Operators To Accelerate AI Adoption - https://thefact.ng/ https://thefact.ng/ 2026-11-20T15:00:06+00:00
  2. Agenda: 2026 Digitalisation and AI in Energy West Africa energyconferencenetwork.swoogo.com
  3. Digitalisation & AI in Energy West Africa 2026 – African & Caribbean Energy Network acenergynetwork.org