Conversations about digital transformation often start with lofty goals about wanting to “improve productivity” and “reshape the way we work.” In more recent times, this conversation has gained urgency as companies look to the expected benefits of robotic process automation (‘RPA’), machine learning and artificial intelligence (‘AI’) for significant yield on their investments.
These are all very important objectives and successful companies are realising significant benefits when these technologies are able to be implemented well. McKinsey & Company cite case studies which show a return on RPA implementation of between 30% and 200% in the first year. It’s a hot topic with a growing market and some terrific, innovative vendorshaving “grabbed the tiger by the tail” and enjoying strong growth. Clearly these technologies must be on your technology agenda.
Realising these ambitions is not necessarily as easy buying a robot and throwing data and transactions at it. Like most things in IT, it takes some proper planning to develop a clear set of expectations of what this technology is going to deliver; to work out what “success” will look like.
A big part of this planning is going to need to be a detailed and pragmatic conversation about the data in your business. While managing the status quo, it is easy to pay little attention to the state and nature of the data in your business. Much of the data is created, managed and used in the context of individual systems, with business processes built around them. Of course, we transfer data between these systems through APIs, or file transfers or services or whatever, but it is still a lot about the data, inside a system, comfortably cloistered and happily doing its thing.
Once you start working on digital transformation, and then RPA, machine learning and AI, data starts to take a different role. Data goes from being the input to / output from a process to being one of the core building blocks of your enterprise. You will need to evolve the way you think about the data in your business.
In a digital enterprise, data needs to have a few key attributes:
- Wherever possible it should be stored once – a “Single Source of Truth.” This SSoT needs to be available to every system or process that needs the data.
- Your SSoT MUST be “golden.” As you are only going to have this as the single source of that data, you must make sure that it is accurate and robust.
- It must come from the best root process for that data, and it should be a natural byproduct of that process. For example, get your employee data from the HR system, it is maintained by processes which manage the real world ‘stuff’ about people in your company; who joins, who leaves, who changes roles, who gets a new phone number and so on. Don’t rely on data which need manual update or comes from somewhere which isn’t “natural.”
- It must be based on standards, which are clear and well understood. By making sure that the definition of what goes into your data sets is well defined, documented and managed properly, you will ensure that it is more easily able to be leveraged across your business. Where possible, align to industry or international standards to improve opportunities to integrate to key partners and customers.
- Data is organic. Make sure you have good governance around your data standards. They will change over time (there was once a time we didn’t store anyone’s email address…) and you must make sure to implement changes in a way which is structured. Review your data regularly, and build your systems to enable changes to be embraced easily.
- It must be open and accessible. You will need to make sure that this SSoT data can be accessed by the systems that need it, when they need it. Some of this might be a regular update of data for changes, some of it might need to be high-performance real-time APIs. Plan it well and design right for the need.
At the core of this transformation is the acknowledgement, as I said above, that we don’t implement monolithic systems which stand in isolation any more. Its all about creating these powerful federations among your applications, which leverage the best capability of each to provide better data, and reduce the effort and complexity in maintaining it.
Applications in the future are going to be different. They will be smaller, they will do fewer things (but do them well), and will enable more automation, AI and new uses we probably haven’t even thought of yet. These micro-applications will combine to achieve the things we need to manage our business, be flexible and ready to respond to change.
Good data is the foundation on which this transformation will be built. Good data is the acorn from which your robot oak trees can grow….