Yet the messaging from the top of the organisation is all around technology, especially the adoption of generative AI (genAI), which is being talked about as the cure for these problems in its own right. Embed these new tools into our processes and everything will improve, is the mantra.
If you found yourself nodding at these statements (even secretly), then you are not alone. Many commercial organisations and leading brands are facing into exactly this set of issues. Yet this problem statement was actually laid out in a recent issue of The Lancet Oncology in relation to the NHS and its cancer services.
The authors expressed their frustration that novel solutions, such as AI-driven diagnostic tests, are being presented as magic bullets for a crisis in cancer treatment, yet fail to address underlying problems in the system of care. They argued that NHS leaders have convinced themselves that new technologies can reverse inequalities whereas, according to the paper, they can create additional barriers for those with poor digital or health literacy.
It does not take a huge leap of the imagination to substitute health for financial services, online platforms, retail and e-commerce, or even fashion as the grounds where this gap between vision and reality is being played out. The top-down rush to adopt genAI as a provider of competitive advantage and internal efficiency has not been matched by a similar appetite for aligning it with current systems and processes which might otherwise be overwhelmed. That includes people, of course.
None of this is unusual in the context of transformation programmes, the likes of which have been taking place across the last two decades to embrace digital technology, cloud technology, big data and now genAI. Rarely do the anticipated benefits get delivered in full and it is often the failure to take a more measured approach, rather than an urgent, accelerationist, tech-first one, that causes the problem.
For this reason, DataIQ argues that adopting a six-step sequence can increase the critical success factors and reduce the friction and risk of failure:
- Strategy – develop a genAI strategy based on identifiable business problems for which it offers a relevant solution.
- Technology – evaluate the solutions available not solely for their claimed capabilities, but also for their compatibility with your core IT environment and tech stack.
- Quality foundations – addressing data quality issues within any data sets that provide the ground truth for training LLMs will mitigate against hallucinations later on.
- Governance – establish guardrails and processes to enforce them for all users of genAI, with clear accountability and responsibilities.
- Leadership – adopt a leadership stance that talks of genAI as an enabler, not a disruptor, and recognises that it is a team sport, not a silo.
- Literacy – provide AI literacy training alongside data literacy programmes at an appropriate level for key roles.
A healthier approach like this could bring about the recovery in performance and profits that so many organisations are looking for. Otherwise, the risk is that, to quote a German saying, the operation was a complete success, but unfortunately the patient died…
“Winning with AI – The essential guide to embedding artificial intelligence into data-driven brands” by David Reed with Paul Hatley is available now.
Book a knowledge briefing with David Reed, here.
