Our Research Assistant Thomas recently attended the AI & Big Data global Expo, an event typically held in London and hosting 5,000 people, and writes on his experience.
The expo consists of 4 main pillars; IoT, Crypto, Cybersecurity, and AI. It’s an opportunity for businesses and leaders to get together, share ideas, and learn. We’ve attended a few, let's say, ‘subpar’ online conferences, so our expectations were pretty low, yet the AI & Big Data Expo managed to keep up the same energy typically found at one of these big-scale events.
We always seek to explore new opportunities that AI can bring to our clients and we’re naturally curious, so we love to attend these events to learn from great innovators as it ensures we stay at the forefront of our industry.
For a while now, companies have talked about AI and the tremendous future benefits it’ll bring, but what’s exciting us is how we can use AI now to start generating value for our clients, so we were keen on the fact that Implementing AI was the central theme at the AI & Big data Expo.
First up, we had Anjali Dewan from AmEx talking about using AI for fraud detection. AmEx believes they have built the most extensive machine learning system in financial services, analysing $1.2 trillion worth of transactions, and machine learning powers 100% of their risk models. They were early adopters of AI, so they have plenty of experience.
Anjali emphasised using AI to improve the customer experience, not replace human jobs. They have the lowest fraud losses in the entire industry (50% lower than their competitors), have doubled their digital resolution rates, and consistently reduced their customer disruption.
They’ve achieved all this by having very reliable and robust data sets, growing a great team of data scientists, and continually monitoring and maintaining their machine learning models.
After Anjali’s talk, I veered off from the AI track to join a talk that was part of the IoT Expo. Here, Rui Pedro Silva shared his thoughts and insights on innovative product management.
The key to a great product is genuinely understanding your user’s problems. He posed that a product is not defined by its features and capabilities but by how much customers understand, relate, and feel passionate about it. For a product to achieve long term success, it must provide value to its users.
“Forget the buzzwords, speak the human language. We should strive to be universally understood.”
Rui shared his five fundamental product management principles:
It’s all about finding out where you can add value to the user. If you add value, even if your product is similar to products already out there, you have a market.
I always find panel discussions are where you find the most valuable insights: the free-flowing structure lends itself to interesting, fresh conversations.
The speakers were all experts on AI, but between them, they worked in a wide range of different industries. The discussion centred around the theme of implementing AI. It was refreshing to talk about AI from this grounded perspective instead of always thinking big picture.
There were so many nuggets of wisdom throughout the discussion, I certainly didn’t manage to note them all down! So, here I’ve picked out the key points and insights from the panel:
Hopefully, these insights will help you get started on your AI journey. Getting started is the most challenging part but remember, “done is better than perfect”.
H&M has used AI across their entire business for a while now, but they see this as the start of their journey. Their vision is: “Making H&M Group the industry leader in applied AI with scalable and integrated solutions covering the entire value chain.
They’ve proven AI’s value within their company by creating internal proof of concept projects for individual use cases. But now, to achieve their vision, they need to industrialise and scale their use of AI across their entire business.
H&M created the Fountainhead, a platform for “democratising” AI within the company to work with “hundreds, if not thousands” of use cases throughout the enterprise. The initiative's main objectives are to reduce the time to market for AI use cases; make AI tools accessible for all product teams; contribute to a higher AI skillset in Business Tech, and increase model output availability for all (e.g. demand forecasts).
Already the framework has delivered business value; it’s reducing time-to-market for use case development by 50% (from twelve months to six), identified potential use cases across the entire value chain; and fostered a co-creation and sharing culture.
Rolls Royce is also an early adopter of AI with over 20 years of experience in advanced analytics. Recently, they’ve been using AI to monitor and analyse around 8,000 aeroplane engines per day in real-time to provide proactive maintenance. But this presentation wasn’t about their AI use cases but rather about scaling AI across their enterprise and the technical and social challenges they faced.
They found that different AI deployments bring various trustworthiness issues. For example, facial recognition can be fraught with bias, or predictive maintenance systems may start to provide untrustworthy outputs. To challenge AI mutation and bias, they began thinking about AI integrity assurances. They analysed AI decisions the same way they would investigate a human colleague’s decision. He went on to say, “When your colleague makes a decision, we don’t have a look at which neurones fired”. Here are the steps they take when evaluating an AI’s integrity:
This new take on AI led them to create the Aletheia Framework, which is now the industry standard AI ethics framework. The framework is a toolkit to guide the practical application of ethical AI projects. It is designed to go beyond theory to be a straightforward, 32-step process that any organisation can follow so that its AI is accurate, well-managed and has a positive impact on the world. Importantly, it does not seek to influence the AI algorithms themselves, as these can constantly be learning and evolving. Instead, it provides a checklist of measures to ensure the AI application's initial design is ethical and that its resulting outputs remain unbiased and true to the intended design.
My experience over the two days of the expo was excellent. AI is rapidly maturing, so to have many early adopters and pioneers share their experiences with implementing AI is extremely valuable. We’re looking forward to taking these learnings, applying them to our initiatives, and solving some of our client’s problems.
I look forward to attending the event next year, in-person rather than through my screen, to experience the conference with all the energy I’ve heard about.
If you would like to hear more about what we learnt at the event or how we’re helping our clients with AI, please get in touch, we always fancy a chat!