Data is the raw material of the generative AI age. At Dell Technologies, we understand that any successful GenAI model needs accurate data in its foundation to produce relevant output. Just as chemists work to turn basic elements into something new, organisations are eager to transform raw data into insights using AI.
By Doug Woolley, vice-president: sales southern Africa and director at Dell Technologies South Africa
At the backend, AI models work like chemistry. They can extract value from data, but only if we provide the right ingredients. We now generate 0.33 zettabytes of data per day – and the number keeps growing.
McKinsey estimates that GenAI could add the equivalent of $2,6-trillion to $4,4-trillion in annual economic benefits. This is a formula with an impact comparable to innovations like the periodic table, and discoveries like the DNA molecule.
In South Africa, the integration of AI into the economy can contribute to a stronger economy by increasing productivity and improving competitiveness on a global scale.
But just like any chemical discovery, you must use the right method to get valid results. Here’s a breakdown of the steps to transform your organisation.
Data is the substrate used to train AI systems and is what AIs ultimately act on. Because clean and reliable data is the key to drive insights and actions, data quality is paramount. Like chemists purifying substances, organisations must cleanse and refine their data.
According to Dell Technologies’ 2024 Innovation Catalyst, only one in three organisations, both globally and across the EMEA region, report that they can currently turn data into real-time insights. As the adage says, garbage in, garbage out. AI/GenAI outcomes are only as powerful as the data that the model is running on.
But for most organisations data is distributed across different locations. Most of it resides on-premises, while more than 50% of enterprise data is generated at the edge. It’s difficult and expensive to move data from one location to another. It’s more efficient to bring AI to the data.
Training and running AI models on-premises can bring benefits to processing, analysis, compliance and intellectual property management.
Dell recently validated with ESG that LLM inferencing on-premises is 75% more cost-effective than in the public cloud. Organisations win when they bring the right GenAI model to their prepared data.
Combining elements
Chemists prepare compounds of different elements, mixing them to create new substances. In the world of GenAI, you can do the same by working with open ecosystems – operating models that share data and services to create value.
AI/GenAI workloads require flexibility in infrastructure and software that can adapt as fast as the models evolve. Open LLMs create equal opportunity across the ecosystem, which in turn allows organisations to accelerate progress and solve problems. From startups to public sector and enterprise organisations, all corners of industry have a role to play.
The mixing of different elements – in other words, collaboration – fosters new opportunities and can reduce the cost of AI development. Openness ensures healthy competition, choice and knowledge sharing. And we must not forget the ethical component. Open models are under public scrutiny, which pushes research labs to reduce bias and secure the data. It’s like combining and distilling elements with an ethical lens.
The formula of insights
Once you purify your base material – your data – and combine the right elements in an open ecosystem, you can achieve the breakthrough, the formula for insights. AI algorithms predict trends, customer behaviour and market dynamics. These insights work like a formula to help sustain organisations and guide strategic decisions.
AI is not magic – it’s a disciplined practice. Data scientists and engineers follow precise methodologies to unleash innovation. Instead of flasks and beakers, their labs are workstations, data, compute and storage. These are valuable tools to extract wisdom from data. What do you currently have in your chemist toolkit?
The chemical imperative
AI and data management are deeply intertwined. You need a rigorous data strategy to reap the benefits of generative AI models. We recommend treating your data as a raw element. It requires refinement and a detailed process to turn into a valuable substance. The call is for your organisation to evoke the chemist’s fundamentals – stay curious, be persistent and committed to turning your data into insights. Only then will you unlock the transformative value of AI.
In South Africa, the first priority is to accelerate customers’ journey to AI. To do so, we are offering our customers a free test environment where we will work with you to build a proof of concept around your specific AI requirements.