4 minute read

What HR can learn from EDF Energy on developing robots to support back-office activities

EDF Energy has employed robotic process automation – in effect digital workers – to create efficiencies and cut costs in its back-office activities. Its approach provides some practical lessons to people leaders on how to approach the implementation of intelligent automation

Image by geralt on Pixabay

The challenge

EDF Energy is one of the UK’s largest energy providers and wholly owned by the French state-owned EDF (Électricité de France). It operates power stations and wind farms and provides energy to more than five million residential and business customers in the UK.

Playing in such a competitive market, a focus on costs is always present. In 2017 it decided to examine whether automation could enable it to provide a better service to customers while cutting costs.

Robotic process automation (RPA) – a form of business process automation technology based on software robots or artificial intelligence (AI) workers – was one option that offered potential for cost saving and efficiencies. This works by replicating what humans can do in well-defined tasks and processes.

The company had previously outsourced much of its back-office function to offshore teams in India. It viewed the challenge as an opportunity to automate some of these processes to make some savings.

The key thing is to have people who understand processes and how the business works. You can then train them how to build and use the RPA

What EDF Energy did

In 2017 EDF Energy ran a capability project to see if the opportunity it thought was there was real. According to Robin Parker, business consultant at EDF Energy, the project covered six requirements:

  1. Proof of concept
    EDF Energy partnered with an IT company to help build a couple of demonstration processes onsite to confirm that the technology worked on EDF’s IT systems and did what it said it could. On the back of these demos, the company found that the technology could automate some of the functions it expected
     
  2. Vendor assessment
    The company assessed a number of suppliers, analysing the pros and cons of each one and ensuring it focused on the areas which were most important to EDF Energy
     
  3. Operating model
    The next step was to consider how the company would introduce the automation. It asked questions such as what did it in need in place in order to deliver the results required and how would the governance work?
     
  4. Delivery mechanism
    EDF Energy decided it needed a centre of excellence team to design, build and operate the system. So, it examined issues such as how many people would be required, where the centre of excellence would sit within the business and how the team should be structured.
     
  5. Opportunity assessment
    What was the size of the prize? According to Parker, the easy targets to start with were the processes that had been outsourced. There were some 900 processes that needed to be filtered down so the firm looked at how many people operated these processes day to day. If only one person was operating a process every now and then it was not worth considering. If 20 people were operating a process every day, then that was a good candidate. If the process was likely to change, then it was left to come back to later. This approach enabled EDF Energy to filter down to around 50 processes that offered the bulk of the benefit.
     
  6. Creation of a detailed playbook on how to make it happen
    All the above was brought together as one business case based around headcount reduction and cost reduction. This was baked into the 2018 business plan so that the project had budget and buy-in.

The company then tendered for the automation provider, selecting the Thoughtonomy platform. It established the centre of excellence and recruited most people to it internally. “The key thing is to have people who understand processes and how the business works,” explains Parker. “You can then train them how to build and use the RPA.”
Some experienced developers were recruited externally to “help turn handles faster” and mentor and develop the inhouse team so it could grow its skills more quickly.

The results

The automation was put in place in 2018 and 20 processes are now live with one to two rolling out every week. According to Parker, some of the key lessons include approaching a project like this through collaborative development, with the business leading it, as well as ensuring you start with an area where you have well-documented overall processes.

EDF Energy started with residential where it had a good handle on where people were working and good KPIs, so it could see where it could save on people. Then it had to drill down further.

“In order to do a build you need to examine processes in extreme detail – down to button press level,” he advises. “When we looked in that level of detail we were able to make some efficiencies and improve processes before starting automation and, in some cases, we improved so much it was not worth doing the automation. So just by looking at processes in this detail were able to make efficiencies.”

In addition, the outsourced operation was closely involved in the programme and has been able to identify areas in which it could also get efficiencies.

The company is now looking at other areas of the business including its UK-based call agents, where it will run some pilot trials to check whether the benefits it thinks are there are in reality realisable. For example, a bot could increase the speed at which a new bill is raised when a customer calls with meter readings, enabling the call centre staff to focus on customer service and upselling in that call.

The project has been so successful that EDF Energy is now looking to expand its centre of excellence team to help deliver more automation at a faster rate into the business.

Published 20 March 2019. Robin Parker was speaking at the SSON IA World Series event

What did you think about this content? Use the stars below to give it a rating out of five.

Total votes: 37