As I’ve delved deeper into the world of Generative AI implementation, one thing has become crystal clear: a platform strategy is crucial for enterprise AI deployment. It’s not just about jumping on the AI bandwagon; it’s about choosing the right platform that can handle a variety of models for different use cases. This flexibility is key to staying ahead of the curve.
From my conversations with CIOs, it’s evident that asset-intensive industries and the automotive sector are gearing up for widespread Gen AI deployment in the next 18 months. But it’s not a one-size-fits-all approach. Take the banking sector, for instance. They prioritise internal use cases like fraud detection, code generation, and contact centre optimization. Why? Because for them, the source of truth and data integrity are paramount. It’s a lower-risk strategy with a quicker time to value.
Healthcare, on the other hand, is all about operational efficiency. They’re looking at Gen AI for everything from data analysis and ambient note-taking to automating clinical documentation and coding. It’s about streamlining processes and improving patient care.
Across the board, productivity is the name of the game. Executives are eyeing Gen AI applications across the organization:
- Front Office: Enhancing customer experience, boosting sales, and empowering contact centres and frontline staff with AI copilots.
- Back Office: Supporting code development and fraud prevention.
- Products/Services: Refining product recommendations.
- Core Capabilities: Streamlining processes like loan processing.
Enterprises must evaluate whether Gen AI will lead to significant technical debt or if it offers composable and extensible solutions. Adopting a test-and-learn approach is crucial, allowing organizations to stay agile and capitalise on market innovations.
But here’s the million-dollar question: Will Gen AI lead to a mountain of technical debt, or is it a flexible, extensible solution? My take? We need to adopt a test-and-learn approach. The Gen AI landscape is evolving rapidly, and we need to be agile enough to pivot and capitalize on innovations as they emerge.
It’s crucial to remember that LLMs are only as good as their training data. Expert review is non-negotiable, and without strong prompting and guardrails, we’re treading in risky territory. Current high-impact use cases – from customer experience and sales to legal compliance and software engineering – all require human oversight.
So, what’s the way forward? Here are my key recommendations:
- Start narrow with Gen AI. Don’t try to boil the ocean.
- Get a grip on your organization’s data management.
- Embrace experimentation over perfection.
- Frontload robust governance.
- Remember that augmented intelligence (humans + AI) is currently superior to AI alone.
- Shift your mindset – Gen AI is expected to deliver a 100x return.
- Assemble your best internal expertise to drive AI initiatives.
- Prioritize infrastructure flexibility to avoid vendor lock-in.
- Don’t underestimate the importance of non-technical staff in Gen AI production.
- Focus on scalable use cases.
- Build a composable AI platform.
- Adopt a platform and product approach to Gen AI.
- Solve real problems, don’t be a solution in search of an issue.
- Create an AI factory with dedicated resources and funding.
- Accelerate success with human-in-the-loop approaches.
- Improve solutions through user co-creation.
- Prepare for the convergence of quantum computing and AI, especially for blockchain products.
- Leverage your partner network for robust, scaled growth of your AI strategy.
One final thought: don’t overlook the power of academia. Universities are hotbeds of AI innovation, and tapping into these specialist epistemic communities can be a game-changer. Consider establishing a PhD leadership network to gain access to world-leading expertise in strategic business areas.
The Gen AI revolution is here, and it’s moving fast. By adopting a platform-first approach and keeping these recommendations in mind, we can harness its power to drive real business value