Lessons for universities from using ‘bots’ in the NHS

There are parallels between the service expectations and challenges faced by the NHS and those of universities. Using integrated artificial intelligence, or “bots”, can help both sectors rapidly improve efficiencies and the quality of key services. Here we share some knowledge gained from working with Robotic Process Automation (RPA) in the healthcare sector, which could be applicable for higher education institutions.

The challenge

In 2019 KFM was tasked with using bots to streamline and digitally automate a series of legacy systems and processes within King’s College Hospital NHS Foundation Trust, which soaked up 10 days of staff time and effort per month on manual, repetitive and boring tasks. 

From 2020 this was done against the backdrop of the pandemic. King’s College Hospital had the UK’s largest Covid-19 wards in the country during the first peak of the pandemic. This put huge pressure on services to be optimally efficient and responsive and to leverage technology to do this, where possible.

An automated solution

We worked with procurement, finance, HR and clinicians to design and deploy bots to automate many time-consuming processes. This has resulted in an estimated saving of 70 days per month of staff time, cost avoidance of more than £500,000, and a reduction in data errors by 10 per cent, which equates to more than 5,000 errors avoided. 

What we’ve learnt in the NHS shows there is opportunity within higher education to automate processes which are repetitive, rule-based, high volume and low in variability. Early adopters in universities first used RPA for back-office functions, however it can now be used to ease the administrative burden of admissions processes, course scheduling and staff rosters, sickness reporting, managing invoices, student performance data, staff training, support services and career advice and more.

Bots can also help transfer data between old IT systems that do not interact, saving staff the task of manually downloading then uploading information.

Our advice

• Start with a Discovery Meeting with a cross-section of academic and non-academic support staff to identify institutional “pinch points”.
• Reassure colleagues bots will enhance their jobs, not eliminate them.
• Bots work best on repetitive, rule-based, high volume, and low variance tasks.
• Start with a well-defined, simple but repetitive process. This will help prove the concept with internal customers and garner support for more complex projects.
• Get the IT team onside from the outset because bots interact with a variety of systems and IT infrastructures. You will need your IT department’s support to ensure the bots have sufficient access to do their tasks.

A couple of warnings

• Don’t try to automate processes that don’t currently exist.
• Don’t automate processes with a high level of variability because you will have to program the robot to deal with each variation.

Getting started

The starting point is a Discovery Meeting with a cross-section of managers to identify pinch points and processes most in need of updating – these will become the prime candidates for RPA.

Bots are interoperable across multiple systems so they can add enormous functionality without incurring great cost or revamping whole IT systems – this is helpful for institutions with old legacy systems that are not worth updating.

Bots can extract data from different departments, to create a dashboard that can produce multiple reports for different managers. The bots then react to systems in the same way a human would, they see what needs to be done and apply themselves to the task.

The process to implement RPA is set out below. It is best to start with a single process as a pilot or proof of concept.

The journey from initial Discovery Meeting to first automation could be within a month. 

How bots can be used to improve efficiency: living examples

Covid-19 pandemic response

RPA has supported King’s College Hospital’s staff Covid-19 swab testing scheme. Patient-facing staff are required to be tested regularly. The test results need to be recorded by hospital admin. Previously someone needed to collect and compile all the results before laboriously entering the details into the hospital’s rota system. RPA has been used to automate the compiling and data-entry processes. Staff send results to the bot’s email address where it picks up the information and verifies it to ensure it is accurate and error-free. The bot updates the system to ensure that the most up-to-date details of each staff member’s test results are readily available. This is useful for staff and rota management because the bot alerts managers of mandatory self-isolation periods following a positive test result. At the height of the pandemic the robot helped save four hours of staff time per day.

Staff training and rostering reminders

The bot extracts HR data and uses it to remind staff to complete their training and for managers to complete their rotas prior to the next shift cycle. The bot sends about 100 reminder emails every month to staff who are nearing the expiry of their Mandatory and Statutory Training. This again frees up HR staff time to focus on other areas of personnel support and management.

In HR bots can also be used for new starter and leaver processes. Most businesses have multiple systems they need employees to interact with throughout their employment. RPA can be used to take details from one form that a new or departing employee can complete and then enter the details of that employee into all the systems they are, or were, required to access.

Closing old POs

The bot worked its way through 35,000 aged purchase orders that would have otherwise required manual review and intervention to close. The bot completed this process in just six weeks; closing more than 34 purchase orders per hour, 24 hours per day, seven days per week. Not only did the finance team benefit from not having to complete such a manual, repetitive task but the bot ensured that this process was carried out error-free. The likelihood of human error in processing 35,000 items is very high.

Carol Glover is head of communications at KFM, a subsidiary of King’s College Hospital NHS Foundation Trust.