A reminder that AI will spread fastest in the finance area rather than some of the esoteric use cases that catch the eyes of dreamers.
Data is the raw material and platform for all digital transformation and yet is often forgotten as exciting technology startups divert the attention of CEO and the C-Suite. No digitisation and transformation projects will achieve their potential unless data is given as much priority as customer UX and customer journeys.
Machine learning, robotic process automation (RPA) and AI require massive data to perform well. Not just
- depth but breadth
- current but historical
- in ternal nut external
Yet the majority of that data lies outside the metadata of document management systems in emails, unstructured text, photo and video to name a few.
If not accessible how can an AI program deliver a valid decision? Today my car management system told me their was a system fault; I could tell as the engine faltered and power reduced by 90%.
The diagnostic tools showed an injector fault but the human operative smelt a rat. Asked a couple of questions and decided it was a battery fault. All checked out in the end- the symptoms lead to a wrong conclusion from a pretty "intelligent" set of algorithms.
Fitted new battery, engine working perfectly.
The same happens with insurance claims- algorithms apply business rules to detect fraud and automate claims processing. Yet 25% to 35% of low value claims will be fraudulent but passed by "AI" as the fraudsters know how to play the algorithms at their own game.
Best to stick to use cases where complete data, AI and automation can be combined - in the finance area.
As enterprises and public sector organisations learn how to:-
- unlock & access ALL the unstructured data
- join it with structured & external data
- apply machine learning, RPA and AI
- Combine with human intuition......
The long hoped for revolution in productivity will finally come to bear so that disappointed voters everywhere can enjoy the benefits and not feel left behind.
The reason for this is that the nature of finance is quantitative, and advances in computing readily lend themselves to number crunching and data analysis, more so than in other business domains. As such, any time there is a leap in the capabilities of what computers can handle, the practical use-cases in finance are never far behind this expanding frontier. Raw data is the fundamental resource of both finance and computer science – while the more eye-catching applications in AI (like autonomous military agents) are a long way from being practical, AI applications in finance will have a substantial impact in the next 1-3 years.