By Jonathan Allmayer, Senior Account Executive, Dell Technologies South Africa

 AI is not a future concept; it’s a powerful tool shaping how businesses operate today. The key to unlocking its potential lies in a thoughtful, strategic approach that moves beyond hype and focuses on tangible outcomes. By applying AI to their most impactful processes, organisations can drive meaningful improvements across the enterprise.

This is particularly relevant in South Africa, where new research from Dell Technologies shows that South African businesses are increasingly viewing AI as a strategic priority. The global study, which surveyed 2,850 business and IT decision-makers, of which 50 were from South Africa, found that 92% of South African companies now view it as a ‘key part’ of their business strategy. Additionally, 32% of South African organisations report seeing tangible productivity and financial returns from initial AI investments.

At Dell Technologies, we see AI as a vehicle for human progress. Our journey has provided a clear blueprint for turning AI vision into reality. This involves asking the right questions, establishing a solid data foundation, and implementing solutions that deliver real business value. Let’s explore how you can chart your own course for successful AI transformation.

Charting your course: A top-down and bottom-up approach

A successful AI strategy starts with introspection. Before diving into technology, you must understand what makes your organisation unique.

·       What makes us special? Identify your core differentiators.

·       What problem are we trying to solve? Define the specific challenges you want to address.

·       What specific process are we changing with AI? Pinpoint the exact operational areas for improvement.

This top-down approach ensures your AI initiatives are directly tied to your business goals. It’s about being actionable and strategic from the very beginning.

Simultaneously, a bottom-up approach is necessary to identify common patterns across your organisation. Not every use case requires a unique tool; by identifying these patterns, South African companies can develop versatile capabilities that serve multiple functions, creating efficiency and preventing the need to create a bespoke snowflake for every new project.

Prioritising AI projects for maximum impact

With countless potential AI applications, prioritisation is critical. We recommend evaluating use cases against two core criteria: business value and feasibility.

Business value measures how a project will contribute to your bottom line. Will it make your team, products or processes:

·       More effective?

·       Better in quality?

·       Faster to execute?

·       Cheaper to operate?

Feasibility assesses the practical and technical viability of a project. This involves evaluating your readiness across key areas:

·       Data: Is the necessary data available, accessible, and of high quality?

·       AI model: Do you have the right model, or can you build or acquire it?

·       Process: Can the existing process be adapted for AI integration?

·       People: Do your teams have the skills to implement and manage the solution?

·       Platform: Is your underlying technology infrastructure ready to support AI?

By plotting potential projects on a matrix of high/low business value and high/low feasibility, you can quickly identify the initiatives that promise the greatest return on investment and are most likely to succeed. High-value, high-feasibility projects are your clear starting point.

The foundation of AI: Modern data management

Data is the lifeblood of any AI system. Without a solid data foundation, even the most advanced models will falter. Establishing this foundation is a multi-step process.

·       Data discovery: Begin by uncovering all relevant data sources within your organisation. The goal is to identify datasets that can be paired together to provide rich context for your AI models. Think about broad reuse, treating your data as a product that can serve many applications.

·       Data preparation: This crucial phase involves cleaning and sanitising data to ensure its quality. It also includes managing data access, classification, and security to maintain governance and compliance, no matter where the data resides.

·       Implementation: With clean, well-organised data, you can move to implementation. This can range from using pre-trained models for inference to more complex tasks like model augmentation with Retrieval Augmented Generation (RAG), fine-tuning, or training a new model from scratch.

Dell’s AI journey: From process to platform

Our own AI transformation followed this same path. We started by identifying what makes Dell special; in our case, that’s: our go-to-market engine, our world-class services, our market-leading products, and our supply chain.

From there, we determined which processes would benefit most from AI as well as what capabilities were needed.

·       Go to market: To improve sales meeting preparation, we developed a RAG-based chatbot to provide product and solution content, and advanced search/research.

·       Services: To deliver more efficient and effective support, we built a hybrid AI and RAG-based chatbot that supports our service professionals with next-best action recommendations.

·       Product development: To accelerate innovation, we implemented coding assistance tools to enable our software developers to go faster.

·       Supply chain: To create more predictive systems, we leveraged a hybrid AI and agentic system for enhanced supply chain intelligence.

This structured approach, moving from differentiators to processes and finally to specific projects, ensures that every AI initiative is purposeful and aligned with our core mission.

Lessons learned on the path to AI transformation

Our journey has been one of continuous learning and adaptation. We’ve discovered several key lessons that can help guide any organisation embarking on AI transformation.

1.     Evaluate projects holistically: An AI project’s success depends on more than just the technology. Evaluate it against your broader business strategy and its technical viability, not just its fit within an isolated AI strategy.

2.     Focus on core capabilities: Projects will evolve, but the underlying patterns and required capabilities often remain the same. Understand the core functions you need, and you will probably be able to support many outcomes with a few AI platforms.

3.     Embrace ecosystem partners: Enterprise AI projects can easily become complex, unique ‘snowflakes’ that are difficult to scale. Lean on ecosystem partners which bring expertise and help you standardise and scale your solutions.

4.     Accept platform evolution: Technology platforms have short shelf lives. Embrace a ‘two-year rule,’ knowing that you will need to re-evaluate and adapt your platforms to keep pace with innovation.

By turning vision into a structured, actionable plan, you can harness the power of AI to not only solve problems but also to drive your organisation forward. The AI era is here, and with the right strategy, it offers endless opportunities for progress.