Sultan leads the UK Financial Services sector in our Technology Practice. He specialises in large IT enable transformations across various industries, as well as helping clients embrace the future of Artificial Intelligence.
AI is really big right now. Big news and big possibilities, but some big questions too, not least about how to get the best out of it. Could it really eliminate whole categories of jobs, as research by Oxford University suggests?1 Or could it help make all of us smarter, more efficient, and more fulfilled in our work by doing the essential but repetitive tasks, and mundane activities that human beings do not excel at?
It probably won’t surprise you that I’m in the latter camp. Transformation is an overused word in the business world, but this is one case where it’s justified. I genuinely believe AI will transform the way we live and work, and open up new potential for organisations all across the world. That’s why at PwC we are practising what we preach, and putting the power of AI to work to transform our own services, so that we can share these experiences to help our clients do the same.
We employ some of the brightest and best in professional services, but too much of their time and talent is taken up in research tasks that fail to inspire them. Take our Financial Services Risk and Regulation team – their role is to provide accurate, up-to-date advice on this crucial and increasingly complex area, but the sheer volume of information that needs to be assimilated to do that is growing exponentially. This is one area where AI comes into play – we can now use an AI powered assistant to do the first stage of information processing, using technology tools that can cope with both structured and more importantly unstructured data. That helps speed up research tasks, and frees up time to focus on the intellectual challenges of an engagement turning the gathered information into deep insights. And the best bit is that the tools get cleverer over time: thanks to the inherent machine learning capabilities, the more you use them, the more precise results will be delivered at the first pass.
So how have we approached AI in PwC, and what have we learned that our clients can benefit from? The first key point is that we’ve looked at it holistically: we haven’t implemented multiple one-off apps in a scattergun way, but put time and thought into exploring what an integrated AI platform would look like for a professional services firm. Every AI project we’re now working on is the result of that thinking, whether that’s the automation of accounts payable, the development of intelligent ‘assistants’, or automating rule-based tasks in our shared services centre. The second key point is that we’ve implemented it collaboratively. Hundreds of people have been involved across the whole PwC global network – far more so than would be typical in a firm like this.
1. Focus on the big picture
AI can do so much – whether that’s advanced data analytics, software robotics, or natural language understanding – that it’s easy to lose sight of the wood for the trees. So always keep the big picture in mind – what do you want this technology to do. Or put another way, develop the use case first, and the application after.
2. Practice, not perfect
With technology developing as quickly as this, the key is to develop early familiarity across the organisation - getting people comfortable with the concept of AI and its future direction. Don’t get locked into cycles of perfection - waiting to release once the solution is 100% perfect – get a prototype out there fast for people to use, then adapt and iterate as you go.
3. Be open-minded, not dogmatic
This whole area is evolving fast, so stay open to new ideas, and be prepared to change direction.
4. One final thought
AI may have spawned some scare stories but we’ve found that people really embrace the possibilities, once they get the chance to see what it can do. Like so many other aspects of technology, familiarity really does breed content.
1 Research by the University of Oxford predicts that nearly 84% of office and admin support work is ripe for computer automation.