Latest news with #Datacentrix
Zawya
4 days ago
- Business
- Zawya
From tractor tweets to crop status updates: How AI is cultivating the future of farming?
AI in agriculture is rapidly shifting from being a distant concept to a practical, transformative tool. But amid all the hype, the real story that has emerged doesn't focus on the technology itself but on the people using it. Source: Supplied Farming data should be like farming land Like any other entrepreneur, farmers operate on instinct, gut feel, knowledge, understanding and experience, innately knowing when to plant, when to irrigate and how to respond to seasonal shifts. And this is something that cannot be replaced with AI. What AI does offer, though, is a way to back up those instincts with data. So, the farmers' judgment isn't being replaced with algorithms, but rather AI is there to support and validate it. We call this intuitive AI, not prescriptive AI. Agricultural tools powered by AI are now capable of analysing everything from historical crop data to satellite imagery and weather trends, but as impressive as these technologies are, they do not offer one-size-fits-all solutions. Local farmers, in particular, are sceptical by nature, and rightly so. It's very rare to find a single AI solution that would be fit for all purposes. That's why trust and relevance are key to adoption. Every farm is different, and so is every farmer's relationship with technology. Thus, Datacentrix's message is simple: 'farming data should be like farming land'. Think of AI as being similar to a mobile phone — it has a thousand features, but we probably use only five of them daily. The same logic can be applied to AI platforms. Boxed, generic AI solutions won't be able to tick every box, and instead, the approach of 'aiming small to miss small' should be taken. This places focus on smaller projects and proof of concepts, where a use case is taken and solved specifically based on required outcomes or benefits. How could AI in agriculture actually work? Data lives in many places: on servers, in the cloud and even in personal storage platforms like Google Drive or OneDrive. For farmers, data is available on livestock feed, weight and growth, silos, crops, water utilisation, pesticides, fertilisers, satellite and more. There's also information available on the use of tractors or combines (for example, use hours and fuel consumption), as well as on electricity usage. All these various datasets can be consolidated in an AI-enabled cloud – a central unified data architecture that acts as the repository, or the 'brain', for farming practices. From there, a discovery tool is used to mine and extrapolate answers to certain questions from the data. This type of exploratory data analysis is based on pattern detection and trend analysis, allowing users to listen to their data and uncover real, beneficial use cases. This approach can be used for any farming practice, be it for the farm worker, service schedules, manufacturing, production, or any type of discipline. It's a case of taking the data, unscrambling and adding logic to it through machine learning, and ultimately helping the farmer to do more than predict yields by understanding the past and optimising the future. For example, a recent project embarked on by Datacentrix involves the mining of many years of estuary data for fruit varietals, looking at the pesticides and fertilisers used and mapping the information back to what should be paid per hectare, varietal and colour of grape. A common misconception is that AI is expensive, complex and requires a large upfront investment. To counter this, Datacentrix has developed AI-as-a-Service, a cloud-based offering designed to reduce risk and make adoption easy and accessible. Clients consume AI solutions on a subscription basis, enabling them to start small and thereafter scale based on real value delivered, without the need for a hefty upfront investment. This model works across industries, including finance, retail, manufacturing and, of course, agriculture. Whether it's yield prediction, soil health analysis or pest and fertiliser mapping, solutions are tailored to the client's specific data and needs and provide real-world insight. AI also shouldn't be difficult to use. It should be simple to deploy, low touch and low cost, with minimal impact on workflow or a workforce's day-to-day operations, while still having a high impact on business. We believe the best solutions come from standing alongside our clients — literally. At Datacentrix, we call it 'walking the line'. We visit farms, walk the fields, talk through operations and then co-create use cases with our customers based on what we have learned. It's a farmer-first, use-case-first approach that ensures the technology is never disconnected from the context that it serves. The road ahead The use of AI in agriculture will never replace human intelligence. What it can do, however, is enable farmers, traders and agribusinesses to work smarter, not harder. With the right approach, the technology is less focused on algorithms and emphasises the outcomes. It's not just the tractors that are getting smarter; the decisions are too.
The Herald
23-04-2025
- Business
- The Herald
Court sets aside R1bn ICT contract that SANParks had awarded to Gijima
The Pretoria high court has reviewed and set aside a decision by SA National Parks (SANParks) to award a tender worth close to R1bn to Gijima Holdings for the provision of information and communication technology services for 10 years. The court ordered that Datacentrix continue to deliver the services to SANParks for 12 months from the date of judgment on the same terms and conditions as set out in the previous agreement concluded in February 2019. The court said it was just and equitable to remit the matter to SANParks for reconsideration of the tender. Datacentrix was the incumbent after bits successful bid in the previous tender GNP-008-18, in February 2019. That tender was to last until April 2024. Datacentrix brought the review application before the court in April last year after SANParks had awarded the 10-year tender to Gijima. The tender was for the outsourcing of SANParks ICT services, including its network services, telephony services, internet, cloud and cyber security services and the underlying infrastructure that is integral to SANParks business operations within the 22 national parks under its management and control. The tender was advertised in September 2023 and was awarded to Gijima in February 2024. Judge Mandlekosi Motha, in a judgment passed on April 11, described the case as a titanic battle for the SANParks tender, worth just about R1bn. 'Blissfully unaware of this existential threat are the lions, leopards, elephants, rhinoceros and cheetahs, to name but a few affected animals. I am convinced that if these animals had a say in the matter, no-one would be before this court,' Motha said. In September last year, the Pretoria high court handed down an order dismissing the Datacentrix application for condonation concerning a prayer in its notice of motion, which sought to review and set aside the request for bids relating to the tender. The court said last year the review application before it had not been brought within the 180 days regarding the review on the request for bids. In the judgment passed this month, Motha said the preliminary point about the application not being brought within 180 days was not dispositive of the entire case because the prayer which sought to review and set aside the decision to award Gijima the tender was made well within 180 days. The ground of review was that Gijima failed to submit annual financial statements for the June 2022 financial year and submitted only audited financial statements for the two preceding financial years (June 2021 and June 2020). Datacentrix said Gijima ought therefore to have been disqualified from the tender evaluation process for failure to satisfy the minimum mandatory conditions of tender. Motha said the common denominator in all the four disqualifications was the entities' failure to comply with the requirement of three years audited or independently reviewed annual financial statements. 'What is good for the goose is good for the gander. How is it that Gijima was held to a different standard? This is nothing short of disgraceful conduct,' Motha said. Motha said the fact that SANParks came to court and insisted that only one audited financial statement was required 'stinks to high heavens'. He said this conduct was totally at variance with the constitutional precepts of fairness and equity. He said the decision to award the tender to Gijima was not in keeping with the dictates of the constitution, and it would not be far off the mark to characterise the tender process as being shambolic. 'Ultimately, this entire tender process was not in accordance with a tender process that is fair, equitable, transparent and competitive.' TimesLIVE
TimesLIVE
22-04-2025
- Business
- TimesLIVE
Court sets aside R1bn ICT contract that SANParks had awarded to Gijima
The Pretoria high court has reviewed and set aside a decision by SA National Parks (SANParks) to award a tender worth close to R1bn to Gijima Holdings for the provision of information and communication technology services for 10 years. The court ordered that Datacentrix continue to deliver the services to SANParks for 12 months from the date of judgment on the same terms and conditions as set out in the previous agreement concluded in February 2019. The court said it was just and equitable to remit the matter to SANParks for reconsideration of the tender. Datacentrix was the incumbent after bits successful bid in the previous tender GNP-008-18, in February 2019. That tender was to last until April 2024. Datacentrix brought the review application before the court in April last year after SANParks had awarded the 10-year tender to Gijima. The tender was for the outsourcing of SANParks ICT services, including its network services, telephony services, internet, cloud and cyber security services and the underlying infrastructure that is integral to SANParks business operations within the 22 national parks under its management and control. The tender was advertised in September 2023 and was awarded to Gijima in February 2024. Judge Mandlekosi Motha, in a judgment passed on April 11, described the case as a titanic battle for the SANParks tender, worth just about R1bn. 'Blissfully unaware of this existential threat are the lions, leopards, elephants, rhinoceros and cheetahs, to name but a few affected animals. I am convinced that if these animals had a say in the matter, no-one would be before this court,' Motha said. In September last year, the Pretoria high court handed down an order dismissing the Datacentrix application for condonation concerning a prayer in its notice of motion, which sought to review and set aside the request for bids relating to the tender. The court said last year the review application before it had not been brought within the 180 days regarding the review on the request for bids. In the judgment passed this month, Motha said the preliminary point about the application not being brought within 180 days was not dispositive of the entire case because the prayer which sought to review and set aside the decision to award Gijima the tender was made well within 180 days. The ground of review was that Gijima failed to submit annual financial statements for the June 2022 financial year and submitted only audited financial statements for the two preceding financial years (June 2021 and June 2020). Datacentrix said Gijima ought therefore to have been disqualified from the tender evaluation process for failure to satisfy the minimum mandatory conditions of tender. Motha said the common denominator in all the four disqualifications was the entities' failure to comply with the requirement of three years audited or independently reviewed annual financial statements. 'What is good for the goose is good for the gander. How is it that Gijima was held to a different standard? This is nothing short of disgraceful conduct,' Motha said. Motha said the fact that SANParks came to court and insisted that only one audited financial statement was required 'stinks to high heavens'. He said this conduct was totally at variance with the constitutional precepts of fairness and equity. He said the decision to award the tender to Gijima was not in keeping with the dictates of the constitution, and it would not be far off the mark to characterise the tender process as being shambolic. 'Ultimately, this entire tender process was not in accordance with a tender process that is fair, equitable, transparent and competitive.'
