Using technology to automate white-collar work is a top agenda item for leaders in industries with significant back-office costs, such as banking and insurance. With the squeeze on the top line, increasing efficiency is critical to maintaining financial performance. But how to separate good practices from hype?

Whether it’s robotic process automation (RPA), machine learning or artificial intelligence (AI), lessons from previous waves of operational and process change apply. Let’s not forget that automating office work began decades ago with the adoption of core IT systems, followed by customer relationship management (CRM), enterprise resource planning (ERP) and human resources (HR) applications such as Workday. When you embark on an office automation project such as implementing RPA, for example, use the knowledge and expertise of the past to make the most of these emerging technologies. Here is some guidance based on my own experience and work with clients:

  1. Think holistically. Automating selected office processes not only improves processes’ productivity, but also enables other non-technical changes that will release further efficiency. I call this “intelligent automation”. A virtuous cycle can be achieved combining the high throughput and standardisation of automation, non-technical operational efficiency gains, more interesting and value-adding work for employees, and the possibility of new or more customer-configured products. Furthermore, the need for expensive office space can be reduced.
  2. Work with a bag of tools, not just one. Organisations spend a significant amount of time selecting an RPA tool. Once everyone agrees the tool is the “right one”, a project is initiated, processes are selected and automation begins. More often than not there are snags; the tool does not handle Citrix, for example. Then the doubt sets in: Perhaps this automation is over-hyped and not for us? This is akin to playing golf with a single club and doubting your skills when your round is poor. A good golfer requires a bag of clubs, each suited to specific circumstances and shots. Then the golfer can play and improve. The same approach applies to office automation. Perhaps two or three automation tools are required to cover the range and type of processes in your business. Be open to using a range of tools, both technical and non-technical.
  3. Spend time studying both your processes and how people actually work. Group processes into those suitable for immediate automation, those which require some re-work first (perhaps using lean sigma), those which are duplicates between departments or functions, those which can be eliminated, plus any other categories that emerge. There are many. Then determine both the sequencing of the planned automation, the nature of your future operation post-automation and the support required to make the change. Work with your employees and get their input into the process, since they have the “know-how” to help make the transition to automation a smooth one.
  4. Think about how you will manage the change and the possible change of roles for affected employees. How can you retain and re-use their skills and knowledge, which were built up over years? You will need people to resolve and process the exceptions that the robots cannot handle.
  5. Set out a clear HR and communications policy. Align your corporate policies and culture from the start of the project so that you can delineate the options for employees affected by the change.
  6. If you are in a regulated business, involve your risk and compliance team from the start. Once a process is standardised, certified compliant and then automated, operational risk is significantly reduced, compared to the previous human-operated process.
  7. Determine whether you’ll handle things in-house or bring in help from outside. Decide early on what you want to do in-house, what capabilities you will bring in (e.g., consultancy, project management and delivery, robot factory, technology forecasts), and how to handle any phases or changes to your model as you scale. Have a plan to scale quickly if your initial efforts are very successful.
  8. Learn from manufacturing. Most factories today are highly automated, more flexible in what they can produce and making higher quality goods at a lower cost, thanks to automation. Factories are now staffed by far fewer people, supported by machines and robots. This level of change and productivity gain is what we could see over the next 5-10 years in offices.

Personally, I implemented my first office automation (a CRM system) in 1993 and used it to transform the department I led and create a significant new revenue stream for the company. Since then, I have led or been involved in many operational transformations both as a client and as a consultant. The new automation wave may well herald a paradigm shift in the world of white-collar work and is one we must implement intelligently.