A new open-source platform is using artificial intelligence to spot potential red flags in government procurement before questionable deals are finalised.
Artificial intelligence could give South African citizens a powerful new way to keep an eye on government spending, with civil society organisation OUTA demonstrating a system designed to detect suspicious tender activity in real time.
OUTA CEO Wayne Duvenage unveiled the technology at the 9th BizNews Conference earlier this month, arguing that independent oversight is increasingly important where government procurement lacks transparency.
The platform is designed to bring government tenders, contracts, variations and payments into one searchable public database, making it easier for citizens, journalists and watchdog organisations to scrutinise public spending as it happens.
Turning Procurement Data Into Warning Signals
According to Duvenage, the system was developed by technical specialists who previously worked on evidence submitted to the Zondo Commission of Inquiry into State Capture.
The technology collects procurement information from government e-tender portals and municipal websites before using AI to analyse the data.
Among the issues it can potentially identify are unusual supplier behaviour, inflated prices, conflicts of interest and suspicious procurement patterns.
One of its more notable features is a “spiderweb” tool, which allows users to investigate connections between companies and suppliers.
In one example presented by Duvenage, a single phone number was connected to 15 companies that collectively held 48 tenders worth R3.6 billion.
The system can also compare tender specifications and pricing against market averages. A road construction tender, for example, could be assessed against typical costs per square metre to identify potential overpricing.
Putting Citizens On The Ground
The idea goes beyond simply creating a searchable database.
OUTA envisions a network of community-based volunteers who could monitor projects in their own areas and flag problems with the delivery of infrastructure such as roads, clinics and sewage systems.
Duvenage described the concept as putting “hundreds of local eyes” on government projects rather than relying solely on a limited number of investigators and municipal inspectors.
The goal is to identify problems while public money can still be protected, rather than attempting to recover funds after questionable contracts have already been paid.
Government Uptake Remains A Challenge
Despite demonstrating the technology to National Treasury, government departments and municipalities, OUTA says there has been little interest in adopting the system.
The organisation also claims some public entities continue to fall short of requirements around updating the official e-tender portal.
OUTA believes the platform should remain independent of government to prevent interference and ensure that oversight remains citizen-led.
While the technical team has offered the technology for free, running the infrastructure and coordinating the proposed network would cost an estimated R1.5 million to R2 million per month.
AI Also Changing How Businesses Find Tenders
The technology could have implications beyond corruption monitoring.
South African AI startup Axel Technologies has developed Tender Finder, a platform designed to make the country’s complicated government tender environment easier for businesses to navigate.
Founder Lars Gumede has criticised the fragmented and inconsistent way tender information is presented online, saying businesses can spend hours searching for relevant opportunities.
Tender Finder uses AI to sift through tender invitations and surface opportunities based on a company’s products or services.
A civil engineering company specialising in roads, for example, could use the system to identify relevant infrastructure tenders without manually searching through multiple government platforms.
Axel Technologies has also explored expanding the technology to help businesses complete tender documentation.
A New Era Of Procurement Transparency?
South Africa’s public procurement system has long faced criticism over transparency, inefficiency and allegations of corruption.
AI will not solve those problems on its own, but technologies capable of analysing vast amounts of procurement data could make it considerably harder for suspicious patterns to remain hidden.
The bigger question may now be whether government, civil society and businesses are prepared to embrace these tools — and whether greater transparency can become a permanent feature of how public money is spent.
Main Image: BizFunding
