Build confidence in AI
Anyone who has dabbled in artificial intelligence will be aware that the bandwagon is rolling very fast. Perhaps it’s even out of control.
Everyone wants a part of AI: it is talked about as an economic elixir, our technological lifeblood – and even more hyperbolically, the “new electricity”. I think the term was coined by a Stanford professor and has now trickled down. But beyond the hype, what is the truth?
At Secondmind, we are well aware of the many narratives surrounding AI. As scientists and business strategists, however, we are more interested in facts. AI is not new (it’s been around since the 1920s) but even now its impact is not easy to quantify. That’s what Secondmind are setting out to change.
Our goal is to produce AI in which our customers will place their confidence and trust. There is no question that business leaders want to embrace AI, but they are still looking for proof points to be convinced. It’s sometimes difficult for even the most senior of executives to translate the desire for AI into tangible results. In short, we need absolute clarity on what businesses are getting in return for their investment.
Delivering value might sound like a simple enough concept, but a number of leaders that I interact with have had their fingers burned in previous trials. This limited success is resulting in deeper holes in their finances.
Senior leaders in banks, insurance companies, supply chain logistics companies and other institutions regularly express their desire to use AI in their enterprise, but suffer from limited confidence in it.
So how do we – or for that matter any AI company – start building that confidence? More than anything, we need to demonstrate the value that comes with AI. We need to talk in terms of delivering tangible benefits that are measurable and traceable. That is the key proposition. We need to lead from the front.
In this approach, we are different from many companies in that we work together with our clients to define the business problem and agree the key performance indicators that allow us to baseline measures for success. Our technical teams translate these inputs into mathematics, generate decision inputs and validate the impact in the client’s business operation. This is enabling our clients to be more competitive, more agile, more streamlined, and ultimately more profitable.
Not only does that make us unusual, but it also puts us a step ahead of many of our so-called rivals. A lot of companies who want to use AI are swamped with data, but they don’t really know what it represents or if it amounts to anything. Our AI is actually tested and analysed after it has become embedded in a business operation. There’s no hype, it simply works – and we have the results to prove it.
Global supply chains suffer from poor visibility. The very nature of deliveries in any sector means that there are a great many variables, and the lack of visibility can result in operational blind spots and consequently a high level of inefficiencies.
In recent years there has been a huge upsurge in computational power resulting in faster and shorter cycle times at Secondmind. We can test and validate our AI’s performance within an enterprise swiftly and accurately. This enables us to demonstrate – and more importantly, prove – the value of our product.
For instance, take the work we have done with one of our clients in the supply chain sector, an asset pooling company.
It’s a huge operation, with tens of millions of assets moving around the globe daily. Our client has very limited understanding of where those assets are ending up in the extended supplier ecosystem. The assets get broken, they get lost, and they are, to all intents and purposes, untrackable.
An individual asset is not worth a lot of money, but the sheer volume in use means that abandoning them not only depletes the bottom line but also increases the overall carbon footprint.
The AI that we have developed and validated with our client can make more accurate predictions even in the low-data, partially visible environment. Our solution has demonstrated far fewer failed collections and up to 28% improved capacity utilisation. This results in lower transport costs, less wastage and lower overall emissions.
We need clear business metrics which will deliver that all-important demonstrable value. That way we will get buy-in from the key users of technology and make the most of human-machine teaming. And that, as we have seen, is the key. Our clients expect nothing less.