By Christopher Saul, Territory Sales Lead for East Africa at Red Hat
The release of Kenya’s draft National Artificial Intelligence (AI) Strategy is a signal that the nation, and indeed the whole East African region, is heading in the right direction when it comes to building and fostering a technology-driven and innovation-friendly business environment. Kenya aims to leverage AI technologies to address national challenges, develop infrastructure, and establish itself as a regional hub for expertise, all of which help businesses plan for an AI-powered future.
A big part of businesses adopting AI is ensuring that their data aligns with their planned use cases and that their data ecosystems, regardless of their size or scope, are optimised to meet the demands of AI-enabled applications. Strong data ecosystems are essential for both public and private organisations, and they become even more essential if and when they become the foundation of any AI initiative.
More than that, to effectively manage the large amounts of data they handle, businesses need to leverage the full resources of cloud computing. A hybrid approach is the way forward, and with that approach, as well as strategic planning, businesses can adopt AI technologies that best fit their vision of the future.
What does it mean to be AI-ready?
Enterprise IT leaders must understand there are stark differences between traditional data management and AI-ready data requirements. These differences help determine whether their data can be used for any initiative. There are many requirements to consider, including data quantity, quality, lineage, trust and diversity. There are also several parameters that help organisations ensure that their data is ready for AI use. These include validation and verification, data versioning (to avoid AI model drift), and general observability metrics that determine the overall health of the data, as well as delivery times and accuracy levels.
Getting your data AI-ready also means organisations need to establish and meet government requirements. They need to apply policies throughout the data life cycle, comply with all relevant legislation (an important example is the EU AI Act, one of the most comprehensive legal frameworks to date), and monitor for data bias and sharing by users. A big part of being AI-ready is governance and proving that your organisation is ready to innovate with AI responsibly.
It all comes down to use case and quality
Though data quality is a non-negotiable trait of data used to train AI models, the readiness of said data is down the use case. In other words, what are your business goals, pain points, and project objectives? Whether a business is looking to build a chatbot that interacts with customers via social media channels or a model that helps detect fraud by collating financial transactions, each use case requires organisations to deploy different sets of techniques, such as generative AI, machine learning (ML) or predictive modelling.
Once they know what they want their application to do, organisations can focus on data quality. A good place to start is with automation. Automating data capturing and validation can remove the potential for manual input errors. Organisations should also utilise an always-on monitoring platform that creates feedback loops and flags any anomalies that may impact upstream data quality.
Innovation happens in the clouds
Cloud computing has been central to businesses’ adoption of AI worldwide, and the same goes for East Africa. That said, businesses can’t embrace everything by dumping all of these applications and systems into the public cloud and calling it a day. Hybrid cloud models combine the necessary control and security with the innovation capabilities that East African enterprises need to realise their AI-powered ambitions.
Most importantly, the hybrid approach means enterprises can distribute workloads between cloud environments where they need to be to minimise latency and maximise efficiency. They can also access more specialised hardware, scale their AI systems horizontally when needed and, importantly, store and process data based on their lifecycle and regulatory requirements.
All this leads to organisations building and refining data ecosystems that best help them build and deploy the AI-powered applications of tomorrow. And it all starts with data, the lifeblood of any modern enterprise in East Africa.
