Global legal publishers spent 2026 wiring their libraries into the AI assistants lawyers now use every day. African law was largely missing from that wiring. Afriwise has started to close the gap, and says the hard part is the data underneath.
Ask a general-purpose AI assistant how to end a fixed-term employment contract early in the Democratic Republic of the Congo and you will get a fluent, confident answer. Whether it is correct is harder to say. On its own, the assistant often answers from what has been written about the law (commentary, summaries, articles of uncertain date) rather than from the law itself.
For lawyers in London or New York, that problem is fading. Over the course of 2026 the large international legal publishers have connected their research libraries directly to both general AI assistants such as Claude and ChatGPT as well as legal and corporate AI platforms such as Microsoft Copilot, Legora and Harvey, using the Model Context Protocol (MCP), an open standard that lets an assistant call on outside sources and cite them. For most African jurisdictions, there was nothing equivalent to connect.
But this has now changed, with pan-African legal intelligence company Afriwise launching its own MCP connector. Organisations that subscribe to Afriwise can now connect it to Claude, ChatGPT or Copilot and receive answers drawn from the consolidated text of the law, with the article cited and a link to the source.
At launch the connector covers 21 countries: South Africa, Namibia and 19 French-speaking jurisdictions, including the DRC, Senegal, Côte d’Ivoire, Cameroon, Gabon and Madagascar. Questions can be asked in English or French; for French-speaking countries, Afriwise holds both the original French text and a full English translation. More countries are added as Afriwise’s AI-assisted research goes live for them.
The division of labour is deliberate. The assistant reasons about the user’s situation and writes the response; Afriwise supplies the law. A user can ask for the rules on transferring personal data abroad in Senegal, Côte d’Ivoire and South Africa and receive a side-by-side comparison with each provision cited. Where a country is not yet covered, the answer says so instead of filling the gap from general knowledge. Client documents stay inside the client’s own AI environment: Afriwise only deals with the legal question, and nothing is used to train any model.
“Your AI now gets the law from us,” said Steven De Backer, Afriwise’s founder and chief executive. “That sounds simple. It isn’t, because in much of Africa the law itself is hard to get at.”
The data problem
De Backer, a lawyer who has worked in African legal markets for over 20 years, describes three separate problems: finding the law, knowing whether it is still in force, and reading it as it stands today after years of amendments. In many jurisdictions amendments are published separately in the gazette and never folded back into the principal act, leaving practitioners to piece together the current text themselves.
“Global AI is being sold into Africa on top of a legal data layer that largely isn’t there,” he said. “The model can be excellent. If the text it is reading is out of date or incomplete, the answer will be too.”
Afriwise’s response has been to build that layer: crawling gazettes, regulator portals and parliamentary records every day, and consolidating the results through an in-house team of 43 legal and regulatory specialists, supported by AI and working with more than 170 in-country law firms and legal data partners.
Why in-house teams care
The pressure is felt most by corporate legal departments covering many countries with small teams. A survey of African general counsel by the African Corporate & Government Counsel Forum, conducted with Afriwise, found that many rely heavily on external counsel or informal networks to keep track of legislative change.
Those same teams already use general AI assistants every day, and have learned where their limits lie. “ChatGPT is useful to guide initial thinking, but Afriwise is where I go for reliable legal references,” said Anastasia Posokhov, Regional Head of Legal AMEA at Aggreko. The connector puts the two in one place.
Not the whole answer
Afriwise is careful about what the connector does not do. Afriwise answers from the text of the law and does not interpret it; the professional judgement stays with the lawyer.
De Backer is also wary of treating AI as the destination. “AI alone is not the endgame,” he said. “Clients still need to know what is changing before it lands, how the law works in practice, and who the right lawyer is in a country where they have none. Your AI gets the law from us. You still get us.”
His wider argument is that as more legal work moves into AI assistants, value shifts to whoever holds reliable, current source material, and that for Africa that material has to be built before it can be connected. “If the future of legal work is built on data and generative AI, Africa must make sure its legal frameworks are part of the dataset,” he said. “We are still early enough to shape that.”
