How AI can help fix SA’s crippled municipalities

You can also listen to this podcast on iono.fm here.

Most South African municipalities are in serious trouble. Infrastructure is failing, finances are under strain, and only a small fraction receive clean audits from the Auditor-General.

Yet councils must still process mountains of paperwork to award tenders, collect revenue and account for how public money is spent.

Read:

Glen Jordan, CEO of housing finance company Empowa, saw the potential for AI to address many of these municipal challenges after coming up with an AI-powered system for mortgage credit approvals.

What often took days or weeks can now be done in minutes. It made sense to start applying this to the municipal sector.

The starting point is a familiar municipal problem: tender submissions arrive in different formats, often with hundreds of pages of supporting documents. Someone must then extract the relevant information, check it against the tender requirements and explain why one bidder won.

AI could read those submissions, pull out the information needed to assess them and show exactly where each figure came from, Jordan tells the Moneyweb Crypto Pod.

This would give officials more time to verify whether certificates are genuine, claims are accurate and bidders have the capacity to deliver.

“The AI is not designed to make decisions, but it’s designed to assist the workload,” he says.

The idea grew out of Empowa’s work with information that does not fit neatly into conventional forms.

ADVERTISEMENT

CONTINUE READING BELOW

Jordan says the company has also applied its technology to assessing water compensation projects in the US, where very different proposals must be measured against complex requirements. He sees a similar opportunity in municipal procurement.

Read: Just 16% of public infrastructure tenders concluded in the past year

A tender would still be scored against a defined set of rules, with a person validating the result.

AI’s job would be to find and organise the evidence. An official could trace a figure back to the relevant passage in the original submission, while any decision to override a score could be recorded with an explanation.

“Literally, you can link a figure back to the originating document and highlight exactly where that came from,” says Jordan.

That safety valve is important because AI can make mistakes. A system that misreads a certificate or accepts a false claim could simply speed up a bad decision.

Jordan argues that making the source documents, scoring rules and human checks visible would make errors easier to identify and decisions easier to challenge.

The potential uses extend beyond tenders.

AI could help municipalities analyse who is paying their bills and who is not, identify transactions that warrant investigation, and compare maintenance spending with the performance of the equipment being maintained.

Rather than waiting for a year-end audit, officials and auditors could receive alerts when something falls outside expected patterns.

“We should be managing by exception, not by sampling,” says Jordan.

ADVERTISEMENT:

CONTINUE READING BELOW

Where blockchain comes in

Blockchain could preserve a record of the process: the information considered, the rules in force at the time, and the decisions made along the way.

Jordan says Empowa uses blockchain in housing finance to record a buyer’s contributions towards a home, creating a verifiable payment history.

In a municipality, the aim would be to make the record of a procurement or spending decision harder to alter after the fact.

Blockchain cannot establish whether a submitted document is genuine or whether an official made a sound judgement. Those checks would still require human oversight. Its value, Jordan argues, lies in maintaining a record that others can inspect.

Read/listen:

Providing housing to the informal sector using the blockchain
Bring on blockchain: Only 3% of Africans qualify for mortgage financing

The biggest obstacle may be political. A system that exposes inconsistent scoring, unexplained decisions or questionable spending is unlikely to appeal to those who benefit from keeping such decisions obscure.

Jordan acknowledges that implementation will require political will.

For councils willing to try it, the immediate benefit would be practical: less time spent copying information between documents and spreadsheets, and more time checking bids, investigating anomalies and holding decision-makers to account.

For previous Moneyweb Crypto Pod episodes, click here.

You can also sign up for our crypto newsletter.

#fix #SAs #crippled #municipalities