Book a call

Microsoft Solutions Partner · Johannesburg

Enterprise AI you can trust to act.

AFRIBIT builds AI agents, Microsoft Fabric data platforms and secure Microsoft 365 workplaces for South African enterprises and the public sector, and governs them so people stay in charge.

  1. 01Data lands in Microsoft FabricSAP, Dynamics 365, Sage, Salesforce: one governed copy.
  2. 02Intelligence is governedSemantic models and machine learning.
  3. 03Agents do the workGrounded answers and drafted actions.
  4. 04People decideTwo AI keys check it, then a person approves.
The AFRIBIT platformFour layers: data on Microsoft Fabric, intelligence in semantic models and machine learning, AI agents on Azure AI Foundry and Copilot, and people. Data rises through the layers, two AI keys check the agent's action in Biclave, and a person approves it.SalesforceSAPDynamicsSageApprovedBiclaveCheckingPassedKey oneKey two01DataFabric · OneLake02IntelligenceSemantic model · ML03AgentsAI Foundry · Copilot04PeopleApprovals · audit trail

Microsoft partnership

Recognised by Microsoft. Built on Microsoft.

AFRIBIT is a Microsoft Solutions Partner for Modern Work, a designation Microsoft awards for proven capability and customer success, and part of Microsoft South Africa's Emerging Partner programme for 2025 and 2026. Everything we build runs on Microsoft's cloud, so it fits the platform you already trust.

Read the announcement
  • MicrosoftSolutions PartnerModern Work
  • MicrosoftEmerging Partner2025 and 2026
  • Our platformFabric · Azure AI Foundry · Microsoft 365 · Entra ID · Purview · Defender · Marketplace

Flagship work

Procurement intelligence, from ERP to decision.

Built for Bidvest Plumblink on Microsoft Fabric and Azure AI Foundry: a platform that tells buyers what to reorder, what is slowing down and why. It advises and drafts; people decide.

Client Bidvest Plumblink

A Fabric data foundation, a governed model and an AI assistant that briefs buyers every day.

  • Microsoft Fabric
  • Azure AI Foundry
  • Entra ID
  • Azure Managed Application
  • South Africa North
Read the case study
  1. ERPSource dataDaily data from the ERP01
  2. BRZBronzeRaw data lands in Fabric02
  3. SLVSilverCleaned and conformed03
  4. GLDGoldModelled for the business04
  5. AISemantic model and MLGoverned definitions and predictions05
  6. AGTDashboard and assistantBriefing, ranked actions, answers06
  7. YOUThe buyer decidesDraft order exported, person approves07

Public sector

Trusted with government's digital workplace.

We moved SASSA to SharePoint Online, gave SASSA and DFFE modern intranets with their own Copilot agents, and built a modern intranet for KZN DARD.

SASSA South African Social Security Agency

13TB

Migrated from on-premises SharePoint to SharePoint Online, with a modern intranet and a custom Copilot agent.

Read the SASSA story
Based on our experience, we confidently recommend Afribit Software (Pty) Ltd for Microsoft 365, SharePoint Online, Microsoft Teams, Viva, digital workplace transformation, change management, migration, and related consulting services.
Elizabeth JoalaSASSA Project Manager: Infrastructure Management
Afribit Software was engaged to design, develop, and implement a comprehensive digital transformation solution on Microsoft 365, which included a Modern SharePoint Intranet, a custom Copilot Agent, and Power Apps solutions.
Nomcebo DlaminiDFFE Director IT Business Solutions

Our people

25+ specialists. One team.

AI engineers, data engineers, Microsoft 365 specialists and security specialists, led by our core team in Rosebank, Johannesburg, and trusted by government and enterprise.

Meet the team
  • Sylvester Meyer
    Sylvester MeyerCEO
  • AI EngineeringAgents, Copilot and Azure AI Foundry
  • Kananelo Tshabalala
    Kananelo TshabalalaPartnerships Lead
  • M. Yaseen KhaderSecurity Specialist
  • Marshall Dube
    Marshall DubeData and AI Team Lead
  • 25+specialistsAcross AI, data, Microsoft 365 and security, led by our core team in Rosebank.
  • Mbuyiseni Mbatha
    Mbuyiseni MbathaTeam Lead
  • Thato MetsingAI Engineer
  • Change ManagementAdoption, training and support
  • Nthabeleng Makae
    Nthabeleng MakaeAI Engineer
  • Data and FabricLakehouses, models and Power BI
  • Ofentse Magoro
    Ofentse MagoroAI Engineer
  • Nonhlanhla MbuliData Engineer
  • Kemelo Malau
    Kemelo MalauAI Engineer
  • Modern WorkSharePoint, Teams and migrations
  • Akhona Sitshaluza
    Akhona SitshaluzaModern Work Expert
  • Obey MusimboData Engineer
  • SecurityEntra ID, Purview, Defender

How we engineer

Every change checked twice.

We use AI to build faster, and we never let one AI check its own work. Every change is challenged by a second AI from a different company, passes automated security gates and is verified running by an engineer.

  1. 01

    Built with AI, in small steps

    Our engineers work with AI coding agents to plan and build in small, reviewable changes, so every step can be checked properly.

  2. 02

    Challenged by a second AI

    An AI from a different company reviews every change adversarially. Different model families catch different bugs, so one AI never marks its own work.

  3. 03

    Through security gates

    Automated checks for leaked secrets, insecure infrastructure, risky code and failing tests must all pass. They are not allowed to fail quietly.

  4. 04

    Verified running, then merged

    An engineer verifies the change against the running application. Nothing merges until the reviewers and the gates agree.

Our product

biclaveEarly access

Two keys before any agent acts.

An AI from a different company checks every consequential action. Your people approve it. Then it runs.

Insights

What we are reading, building and learning.

AI4 min read

Agentic AI and the Future of Creative Collaboration in Business

Artificial intelligence is advancing at an unprecedented pace and the next major shift is not simply more powerful chatbots, but the rise of agentic AI systems that can plan, decide, and act independently toward defined goals.

Book a discovery call

Let's build what's next.

Tell us about one decision you want AI and data to improve. We will come back with the smallest useful first step.