Five pillars for practical GenAI implementation
These pillars for practical GenAI implementation can help businesses simplify this complex technology.
Is 2025 the year we move in earnest from GenAI hype to GenAI results? Recent research suggests yes, particularly for the UK, which could see almost a doubling in economic growth over the next 15 years thanks to this cutting-edge technology.
However, every tech leader knows they can’t predict every advance on the horizon, even while acknowledging their responsibility to plan for the future as much as possible. Across industries, leaders are faced with taking the plunge and investing in technologies like AI tools to future-proof their businesses. But without the right strategy and plan for adoption, you can end up adrift without a clear idea of where you’re headed next.
Walking this tightrope takes a pragmatic approach, leveraging the best available tools while also maintaining flexibility and control. Practical GenAI implementation isn’t about rigidly committing to one path. Instead, it’s about creating an AI ecosystem that adapts and evolves with your business needs. That could mean choosing platform-agnostic solutions to avoid vendor lock-in, embracing open source to benefit from flexibility and transparency, adopting hybrid and multi-cloud strategies to ensure the best environment for your AI workload or focusing on right-sizing your AI solutions.
Pillars for practical GenAI implementation
Partnering with technology providers can ensure customers harness the power of AI —tackling complexity, risk, and the cost of diving into and supporting AI now and in the future. By offering flexible consumption models, an end-to-end AI-optimized IT infrastructure portfolio, an open ecosystem of deep partnerships with other leading AI companies, and a commitment to open standards, they can support a GenAI implementation that aligns with a business’s unique needs, risk tolerance and long-term vision. In short, they can help to ensure a strategy that is not just cutting-edge but also pragmatic and sustainable.
We can do that for our customers because of the lessons learned on our own AI journey. By implementing AI within our own operations, we’ve gained first-hand experience of its challenges and opportunities, giving us a deep understanding of what works and what doesn’t in real-world business settings. Our “customer zero” approach, where we become our own first and best customer, ensures that our AI solutions are not just theoretical concepts but are grounded in practicality, refined through real-world experience and ready to deliver tangible results for our clients.
Through that practicality, we developed these five guiding principles to help you more rapidly and efficiently deploy AI technologies that will serve your business today and prepare you for your future business. These pillars for practical GenAI implementation are a testament to our own journey and our commitment to helping customers simplify complex technology.
1. Enterprise data is your differentiator
Never lose sight of the fact that your data is a goldmine of insights, and unlike your competitors, you have exclusive access to it. You have a treasure trove of customer, operational, and market data – information that reflects your company’s unique journey and expertise. This data is the secret to success in the AI race.
By building upon pre-trained models and customizing them with your proprietary data, your differentiator, you can gain a competitive edge through deeper customer insights (AI can analyze your customer data to uncover hidden patterns and predict future behavior), proactive risk management (AI can detect fraudulent transactions in real-time by analyzing customer patterns and flagging anomalies) and enhanced decision making (AI can analyze vast amounts of data to identify trends, forecast demand, and optimize pricing strategies – giving you the insights you need to make smarter, faster decisions).
2. Respect data gravity
Although data may be a treasure trove, it’s never found all in one pot. Data is highly distributed, with most of it residing on-premises and more than 50% of enterprise data generated at the edge.
For data to be effective, it must be near applications and services that rely on it for efficient processing and analysis. It is better to yield to “data gravity” and bring AI to the data (where the majority is on-prem) rather than move enterprise data to available computing resources. Most organizations are finding it more effective and efficient to train and run AI models on-prem to minimize latency, lower costs, and improve security. To turn data into actionable insights with AI, often in real-time, a combination of on-premises, edge and cloud deployments is vital. For this reason, 66% of UK decision-makers prefer to build an on-prem or hybrid approach for AI use and procurement.
3. Right-size your AI infrastructure
There is no one-size-fits-all approach when it comes to AI. I’ve witnessed customers across multiple industries, in organizations of varying sizes, implement their AI in innumerable ways – from locally on devices and at the edge all the way to massive hyperscale data centers. Not all models are large and not all AI workloads run in a data center. Or in the cloud. To avoid massively over or under provisioning, it will be important to right-size the AI solutions you adopt to your use case and requirements, so analyze your use cases and goals to determine the most appropriate infrastructure and model types.
4. Maintain an open, modular architecture
Equally as important is a mindfulness that the AI landscape is constantly evolving, and that no one can predict its future course. This means that a rigid, closed system can quickly become obsolete Therefore, maintaining an open, modular architecture will be crucial to help enterprises adapt to fast-paced changes in AI technologies and avoid being locked into outdated or inflexible architectures.
AI / GenAI workloads are a new class of workload – requiring a new class of open, modern innovation spanning the entire AI estate: data layers and lakes, compute, networking, storage, data protection and AI software applications. But it’s entirely plausible, if not likely, that new GPU infrastructure, algorithmic infrastructure, or inventions could emerge in the future that would require enterprises to adapt. The worst mistake you can make today is to bet on and commit to a closed, proprietary, single-dimensional AI system that is not flexible.
Open-standards AI tools offer flexibility, transparency, and a vibrant community for support and innovation. By integrating open-standards solutions into their AI strategy, businesses can avoid being beholden to a single vendor and customize tools to meet their specific needs.
5. Forge a thriving AI ecosystem
No single vendor can solve every AI challenge; collaboration is key. AI is a composite of many technologies, intellectual capabilities, and services, which enterprises will need to mesh with each other to succeed. Be sure to embrace vendors that enable an open ecosystem of partners, from AI major players like Microsoft to silicon providers like NVIDIA and Intel to open-source leaders like Hugging Face.
Open ecosystems provide equal opportunity across the tech ecosystem, support the creation of new GenAI breakthroughs, and give customers greater access to innovation and flexibility. Access to open models and technologies can accelerate progress and solve problems worldwide, fueling a global “innovation engine” across all corners of the industry, from individual developers and startups to public sector and enterprise organizations.
A real-world approach for real-world results
Successfully navigating a new landscape nearly always requires a pragmatic approach that balances excitement with realism, preparation and careful execution. Being able to realize value from new technologies demands the creation of strategic roadmaps, and when it comes to AI, the preparation, quality and storage needs of the data that feeds it have added importance. Don’t be caught up in the feeling that you need to transform into an AI powerhouse overnight. Start by identifying a specific, achievable goal that has the capacity to generate business ROI, and strengthen the route to success with a clear vision and the right partnerships.
We’ve compiled a list of the best free cloud storage.
This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro