Municipal Innovation and AI: Leveraging Data and Technology for Inclusive Governance

Topics: Artificial Intelligence - Digital Transformation - Data Governance - Municipal Innovation - AI Maturity - Inclusive Governance

This webinar brought together public sector thought leaders, municipal practitioners, technology partners and sector representatives to take stock of where South African municipalities currently stand on AI and digital transformation - and what it will take to move from informal experimentation to deliberate, data-driven governance.

The state of play

A live poll of attendees revealed that 39% of municipalities have basic digital systems in place but they remain fragmented, 30% have integrated systems and are beginning to use data for decision-making, and only 7% are actively using AI, automation and data analytics across municipal functions. The message from every panelist was consistent: AI without a clean data foundation will not move the dial.

Speaker contributions

Msizi Gwala - Lead: Public Sector Thought Leadership, South African Institute of Chartered Accountants (SAICA)

Msizi presented the AI Maturity Assessment Framework - currently the only national AI readiness instrument in South Africa, developed in partnership with GIZ, the University of the Western Cape and the Department of Communications and Technology. His contribution focused on how the framework, originally designed at national level, can be adapted for the municipal environment.

• The framework measures AI readiness across eight domains: national or international AI leadership profile; enabling regulatory environment; citizen engagement and empowerment through AI; education and workforce enablement; innovation and research; ecosystem growth; data and digital infrastructure; and security and privacy.
• Five maturity levels are used, from zero (no evidence) to five (optimised and embedded), adapted from ISO standards.
• For municipalities, the framework must be anchored to IDP priorities and applied with a clear question in mind: are you measuring internal municipal operations, or AI activity across the entire jurisdiction?
• South Africa currently scores around 1.4 on international AI leadership - something exists, but it is limited. Security and privacy scored the highest at level three.
• The framework is not a tick-box exercise. It exists to generate insights that inform investment decisions, improvement roadmaps and advocacy for resources.
• SAICA is open to supporting municipalities in applying the framework and is in conversation with SALGA about incorporating it into their existing digital maturity model.

"You should have reasons why you are doing it. If you do it and it is not useful to you, it is still okay to abort mission. It is something you voluntarily use as a foundational tool to measure your maturity."

Johnet van Asperen - Divisional Manager: Business Architecture and Customer Relations Management, Overstrand Municipality

Johnet shared Overstrand's practical journey in building a business case for AI - using the decoding and plain-language translation of municipal bylaws as their first use case. Her presentation answered directly the question of whether municipalities are using AI: yes, and here is how one started.

• Rather than adopting technology for its own sake, Overstrand identified a specific service problem: citizens struggle to find and understand bylaws. The strategy is to enable self-service through AI-powered plain language translation of all municipal bylaws.
• Key principles guiding implementation include: AI assists but does not replace human legal expertise; enterprise-level tools must be used that do not train on input data; outputs must be rigorously verified against original text; and clear accountability must be assigned for maintenance and checking.
• The practical workflow involves converting bylaws to machine-readable format, uploading to a secure platform, using specific prompts, verifying outputs, and piloting with internal staff and ward committees before public rollout.
• The solution will be launched on Overstrand's website and citizen app, with multilingual support as a key accessibility requirement.
• Metrics will include call volume reduction, citizen satisfaction and return on investment. The underlying commitment is simple: if you do not try, you will never know.

"Instead of adopting technology for its own sake, we decided to identify a specific business problem or workflow bottleneck that we can resolve by implementing AI."

Leanne Botha - District Manager, Business Engineering

Leanne offered a behind-the-scenes perspective on what makes AI innovation in municipalities sustainable, drawing on Business Engineering's decade-long partnership with Overstrand through its Collaborator records management and workflow platform.

• AI cannot simplify bylaws unless those bylaws are first captured, structured and managed correctly. Clean, well-indexed records are the unglamorous but essential foundation.
• Behind every citizen-facing AI application is a chain of workflows: uploading documents to secure repositories, applying metadata and version control, routing through approval workflows to legal experts. These checkpoints make accountability possible.
• Reporting infrastructure closes the governance loop: tracking which bylaws are accessed most frequently, where citizens struggle with compliance, and whether call volumes are decreasing - these are the KPIs that prove whether AI is delivering.
• AI is the visible tip of the iceberg. Records management and workflow systems are the submerged foundation. Long-term partnerships between municipalities and technology partners are what make innovation sustainable rather than once-off.

"When systems help people, and AI helps systems, government becomes more than just administration - it becomes transformation."

Abdul M Gabier - Deputy Director: ICT, Cape Winelands District Municipality

Abdul shared Cape Winelands' experience of rolling out AI across the municipality - starting with the concept of the "cyborg": a human employee augmented by an AI assistant. His contribution was one of the most candid of the session, covering both the gains and the governance challenges that have emerged.

• AI adoption at Cape Winelands has been remarkably smooth from a change management perspective - unlike most IT implementations, staff took to it immediately. Usage is monitored and uptake is high across the organisation.
• Practical gains include significantly improved policy and SOP generation, better HR and disciplinary documentation, enhanced report quality, stronger data analysis capability, and higher volumes of internal audit work being produced in less time.
• However, time saved has largely been reinvested into the same tasks rather than redirected to new priorities. Output quality has improved; output volume has not necessarily increased. The municipality is not yet building more roads or delivering better healthcare - employees are simply more confident in their work.
• New governance challenges have emerged: ownership and capability are inconsistent because AI is a non-deterministic tool and different users get different outputs; expectation management for task timelines has become more complex; and management oversight requirements have increased as reports become longer and more detailed.
• On ethics, Abdul argues that ethical AI is not about removing bias - it is about ensuring that any prioritisation is fair, justified and aligned to policy objectives.
• The roles needed to take AI to the next phase - data architects, AI solutions architects, cybersecurity specialists - do not yet exist formally in the municipality's structure. Hardware and licences have been procured but the human capability to fully exploit them is not yet in place.

"Are we managing the AI, or are we letting the AI manage us? Ownership and capability are going to be the big issues in the next year."

Moshe Mabanna - Managing Director: Business Consulting Division, Ntiyiso

Moshe reframed the AI conversation around its most overlooked prerequisite: data. His presentation, titled "Truth First, Then Intelligence", argued that most South African municipalities are not yet ready to derive value from AI because their data foundations are not in place.

• Most South African municipalities sit between data maturity levels one and two: decisions are based on gut feel or manual reports, data is seen as IT's problem, and there is no master data management in place.
• Before investing in AI, municipalities must work through eight data pillars: data strategy; data governance; technology and systems; reporting and business intelligence; master data management; analytics; AI and machine learning; and agentic AI. AI sits at pillar seven - it can only work when pillars one to six are functional.
• The practical starting point is data cleansing - reconciling billing, valuations and meter records; resolving duplicates; and standardising reference data. One customer, one record, one version of truth.
• Where the data foundation is sound, AI earns its place quickly. Ntiyiso used an AI model to scan two years of prepaid token sales at a Johannesburg utility and flagged approximately R1.5 billion in anomalies for review.
• Municipalities that will win the AI race will not be those that invest most in AI - they will be those that invest in data, people and governance that make AI deployment possible.

"You put AI on bad data and it is not going to move the dial for you. Fix the data foundation first, then choose the AI use cases worthy of spending on."

Kutlwano Chaba - Chief Digital Officer, South African Local Government Association (SALGA)

Kutlwano brought a sector-wide perspective, drawing on SALGA's work across all 257 municipalities and sharing several commitments SALGA is making to accelerate AI adoption in local government.

• At minimum, most municipalities are already using AI informally through Microsoft Copilot embedded in their existing Microsoft 365 stack. The ChatGPT phenomenon of four years ago normalised personal AI use, and that behaviour has carried into the workplace. Shadow AI - staff using personal AI accounts for work tasks - is a growing governance risk.
• Globally, over 60% of AI projects have failed or been cancelled, largely due to the absence of a business case and roadmap. Local government cannot afford to experiment without structure. Every AI initiative must be anchored to a business case with clear use cases, a roadmap, and measurable returns.
• The most popular early use cases - document generation, report writing, basic data analysis - are valuable but represent only the surface of what AI can deliver. The bigger opportunity lies in backend AI agents running revenue, fleet, billing and compliance workloads, freeing up professional time currently consumed by manual tasks.
• SALGA is developing a multilingual large language model for local government, currently being trained in isiZulu, Xhosa, Setswana, Sesotho and Sepedi, with Venda and Xitsonga planned for the next phase. This model will be available as a shared platform for all municipalities to build on.
• A digital transformation grant for local government is needed. Money and skills remain the two biggest barriers to meaningful digital transformation in the sector.
• Vendors and service providers must be held accountable through procurement and contract management to embed AI capabilities into their existing products as part of their normal product lifecycle - not as optional extras.

"AI is just another puzzle piece in the digital transformation journey. Let us not forget that. And if we have this conversation again next year, local government will be far ahead - because the energy is right."

Key takeaways for practitioners

1. Start with a specific business problem, not with technology. Define the workflow bottleneck, establish measurable metrics, and build a proper business case before procuring anything.
2. Data first. AI on a weak data foundation will not deliver results. Conduct a data maturity assessment, fix master data, and work through the foundational pillars before investing in AI use cases.
3. Use the AI maturity framework as a diagnostic, not a compliance exercise. It exists to generate insights and investment priorities, not to produce a score.
4. Govern shadow AI actively. Staff are already using personal AI accounts for work purposes. Policies, training and governance frameworks must address this explicitly.
5. Demand AI capability from existing vendors. Contracts and procurement reviews should require service providers to embed AI features into their products as part of standard product development.
6. Invest in the invisible infrastructure. Records management, workflow systems, data governance and metadata are the foundation that makes AI trustworthy. They are unglamorous but non-negotiable.
7. Plan for new roles. AI creates demand for data architects, AI solutions architects and prompt engineering capability that most municipal organograms do not yet accommodate. Start planning now.
8. AI is the great equalizer. Access is cheap, change management has been easier than expected, and the skills gap between municipalities can narrow quickly for those that act.

Moderated by Mr. Zolani Zonyane, The Municipal Edge
Hosted by Mr. Dan Claassen, Future Cities Africa.

CPD points available- contact cpd@cigfaro.co.za