Technology, big data systems, architecture and data science algorithms are now common place in the business world and are widely used to solve problems and provide insights. However, on their own, they are not sufficient to achieve the business advantage and return on investment (ROI) required to excel in today’s highly competitive world. Instead, businesses now need to concentrate on digital governance, scalability and production-ready solutions to fully utilise the potential AI has to offer.
Algorithms and specialisation versus production-ready solutions
So far there have been a lot of proof-of-concepts (POCs) and siloed implementations in the areas of data science and big data. However, in order to reach their full potential, these now need to be turned into scalable solutions in the production environment before being expanded across the enterprise. In effect, technology is only good enough to capitalise on investment if it can be rolled out into production. This means that in the future, developing a mature software house capability will be paramount to delivering the business vision.
To do this effectively and efficiently, automation is required. DevOps methodology, comprising shorter, faster software releases, has been proven to aid this approach, speeding the path to production, with less risk and broader collaboration across the business, data science and development teams. See Brickendon’s DevOps page.
How to reap the rewards of embracing change
At the lowest level, firms should seek to use localised solutions, take the low-hanging fruits and incorporate routine tasks and traditional change that allows them to fix common business pain points. At the next level the
While this approach is not favoured by all, those who are not scared of change and are prepared to embrace new business models and approaches, will reap the rewards of long-term growth creation and increased competitive advantage.
True disruption requires proper implementation
While the potential for AI as a disrupter is large, for it to be properly effective it needs to be implemented correctly. A dalliance in the domain simply because it’s a hot topic is not enough. Adequate research and preparation is required and a proper process should be followed in order for the full benefits to be reaped. It is important to measure the impact and cost of a solution in relation to the complexity, firstly in relation to the level of innovation, and secondly, to understand when to buy from vendors and when to build in-house.
AI: as innovative as you can make it
The holy grail is no longer just about predictive and prescriptive analytics. Businesses need to go one step further to decode the black box of algorithms, provide an explanation to the user and give context around predictions. Known as insight, or causality, this is becoming a key feature in the financial services data world and is an essential focus for any business seeking to reap the full benefits of AI.
The business problem versus the application of technology
Client-driven innovation will most likely require the use of robotics or recommender systems, whilst business-driven innovation might incorporate fraud detection, next-best action or compliance monitoring. Meanwhile ops-driven innovation is an opportunity for increased automation and the use of robotic process automation (RPA).