Evolution, Not Revolution: What the Bestpractice.ai library tells us about the state of AI (Part 1)
At Bestpractice.ai we have published our library with over 580 real world AI use cases and 800 case studies: What did we learn about the current state of AI from having built this?
We have just gone live with the world’s largest open source of AI use cases and case studies. The aim is to share with everyone an understanding of the scope of the market which currently only exists behind closed doors at the world’s most elite consulting, banking and investment firms. It’s taken the team many months of grafting to get this far and over the coming weeks we will be publishing a series of thought pieces based on what we learnt — how to think about use cases, how to define different AI approaches, and, above all, what the implications for corporate and organisational strategy are.
What is very clear is that the hype focusing on the big structural and philosophical questions generated by potential AI technology is obscuring the very real shift that actual AI technology is powering. Increasing the speed and breadth of CV scanning or reducing the cost of spotting unhealthy potatoes on a production line really lack the glamour of Terminator robots. But they are changing business — and with that the world.
Building the bestpractice.ai library reinforces our belief that there will be three phases of AI take up:
- Optimisation — using AI to replace legacy approaches, typically optimising for cost, operational improvements such as accuracy or process uptime and risk reduction.
- New product and services creation — using AI to break traditional product limitations around scaling; whether cost, speed to deploy or quality management.
- Transformation — using AI to break the structural trade-offs between speed, scale and specialism. Historically organisations have been able to offer at most two of the these (and most compete on one) before the complexity generated overwhelms them. Conceptually, AI allows all three in the same organisation.
Most AI-savvy organisations are still entering the first phase. (Most organisations are still not even there.) Organisations that have clarity on what creates competitive advantage and have invested in the IT and data support tools to optimise this are better positioned to take advantage. The Bestpractice.ai library offers a clear guide on where early wins can be made — and the Benefits section for each Use Case shows how those wins can be measured. The inevitable end result of all of this is that organisations will end up with multiple point solutions that will inevitably need stitching together in a coherent fashion.
A few organisations are already deep in to Phase 2. Some are using it directly to offer new products (Amazon’s Alexa for example) and to help build new revenue streams (Amazon’s AI tool kit offering is helping drive the AWS cloud hosting business). These are typically data-native organisations like Google that have begin to re-orient their entire businesses around AI: if you already visualise your business as a data flow then implementing AI is a no-brainer. If data is something that your business throws off and that needs special time to go look at then inevitably its harder.
In terms of breaking the trade-offs between speed, scale and specialism then few companies have developed this: Amazon is amongst the closest. They offer a wide array of products (allowing for long-tail specialist choice), capture scale in their pricing and distribution approach and operate with a speed of response that sets standards. To get there they are investing in AI to support the complexity required — but talking to executives at Amazon shows that even they are still at an early stage on this journey. More widely, the third phase will remain at the planning stage until the tools from the first phase are in the right place (and the hidden challenges of legacy data architecture can be fixed or dodged). However, the challenges are every bit as much human and cultural as they are technological.
It’s been an exciting few months building the library. Do take a look. We will have got things wrong, missed things — let us know. And if you find it useful let your colleagues and friends know.
