1. Why did I decide to form DI originally?
My first company was in Consumer Electronics, an industry where I was witnessing huge amounts of waste be it in the manufacturing process or in retail. Back then was the dawn of 'Big Data' and I realised that my industry's inefficiencies could be solved with using data properly. I was a hardware product guy and new nothing about software or data, so set out to learn everything I could. I built an app called Skim.it that used Natural Language Processing to summarise business documents and make them easier to consume and share and through that fell in love with data science and Artificial Intelligence, realising the huge potential it has to make businesses more efficient and to innovate.
2. How is DI currently using AI and ML?
In two ways; through our data science consulting, specifically Natural Language Processing (NLP) and Computer Vision (CV) and through data processing on our Unstructured Data Intelligence (UDI) platform. Two of my favourite projects at the moment are work we're doing with a Medical Writing company that needs to process vast amounts of unstructured clinical trial data. The other is our work supporting a very large UK retailer to improve efficiencies in back office operations and the customer experience.
3. What sets DI apart from its competitors?
We have ground breaking technology that automates the pre-processing and structuring of data from very complex formats; web, pdf, image, video etc… This unlocks the opportunities of what our clients can do with data. Then we have a team of very knowledgeable data science and AI experts that can create value from the data. Our experts have not only experimented with AI in the labs or academia, but have actually deployed AI/ML models to production at scale. Being an AWS partner we have the best tools in the industry at our disposal to do that, and we provide a managed service to our customers that guarantees support throughout the life of their AI solution.
4. What does the future hold for adoption of AI globally?
AI is now entering the early majority stage in the adoption curve. We're seeing more and more businesses from a wide range of industries from art market to life sciences start to implement AI solutions within their organizations. As the technology itself becomes more democratized through the incredible work from the open source community, plus the cost of training models comes down through breakthroughs in transfer learning thereby reducing the amount of data, time and resource you need, these factors make AI much more affordable and accessible globally.
5. Top tips for orgs wanting to adopt AI?
We're seeing with our clients, who started integrating AI into their business 3 or 4 years ago, that they are now far more efficient, profitable and carving out more market share from their competitors. So my number one tip would be, don't wait! It takes time, good datasets and a good understanding of the art of whats possible to pull off a successful AI implementation, so businesses need to start now if they don't want to get left behind.
I would also recommend you get buy-in from everyone in the organisation. AI can be seen as a scary new technology, but if its designed properly it should be symbiotic with the people using it, providing assistance not replacing jobs.
Finally, I would say, don't underestimate the value of your data. You may have many years of business documents that could be invaluable to the training of a Machine Learning model. We help businesses to unlock the value of that data to uniquely train their own AI models that give them a competitive advantage.