Originally posted on Medium, Feb 26
We are all starved of human contact. The churn of crowds feels alien yet people continue to achieve amazing things together; we have designed tests, developed vaccines, and rallied to help the hardest hit. Human kindness is thriving. And it is fuelled by technology.
Technology is revolutionary. It has driven social change since our ancestors first picked up rocks to butcher animals and spark fires. From the sextant to steam power, the internet to artificial intelligence, technology both triggers and tames huge historical forces. Today, it plays a critical role in both the spread of Covid and our response.
Inspired by this, I began experimenting with robots. I was intrigued by the idea that we could amplify human kindness with automation technology. Could the British Red Cross make people happier and deliver better services by working with machines?
Robots get a hard time. They are mocked and feared in equal measure. Videos of inept cyborg footballers flash across our phones, followed by predications that automation means mass unemployment and riots.
This is because the word ‘robot’ covers a huge span of technologies, from tiny insect-inspired drones to huge manufacturing machines. They don’t need a physical body at all — they can operate from the cloud as pure software.
The common thread is autonomy; robots carry out a series of complex tasks automatically, adjusting their actions in response to feedback from the world around them and the data they encounter.
Artificial Intelligence (AI) — the brainpower that drives most robots — is developing at an astonishing pace.
AI has beaten Go players so comprehensively that our understanding of the game has changed forever. One AI duo unexpectedly invented a new language to chat between themselves. New molecules are being designed by to accelerate drug discovery.
AI and robots can also do much more mundane stuff — the kind of things we hate doing as part of our day-to-day jobs, rather than the kind of intellectual feats we aspire to.
To me, this is a promising use of technology — subcontracting the chores we hate so we can concentrate on the things humans do best, like finding creative solutions to tricky problems. It’s what we did with washing machines, tractors and looms — built machines to help with the work.
The impact of robots will be enormous but even Elon Musk, the world’s most ardent technophile, is uncertain what we face, saying: “AI will be the best or worst thing ever for humanity”. So, it makes sense to investigate it now.
Robots and work
So how might ‘robots’ help a humanitarian organisation?
I began by experimenting with robotic process automation (RPA). This largely lives down the shallow end of the artificial-brainpower pool, but it has huge potential for complex organisations like charities because digital technology has infiltrated everything we do.
We create, process and share vast amounts of information every day, hunting it down in databases, feeding it through complex networks of hardware and software, and firing it around messaging systems.
We spend a lot of time doing this. More than we should. More, in fact, than we spend using the skills we are trained in on the job we were hired for.
According to the 2021 Anatomy of Work report, in UK offices we only spend 26 percent of our time on skills-based work, 14 percent on strategic planning, and a whopping 60 percent on “routine tasks like inputting, moving and reporting information, arranging meetings, following rote processes”.
On average, UK workers spend 60 percent of their time on routine tasks like inputting, moving and reporting information, arranging meetings, following rote processes
These processes are often poorly mapped or only exist in people’s heads. The rely on legacy systems and lists of log-in details. They mean downloading and moving files, shifting formats, copy and pasting details from one place to another.
They suck the will to live, make people miserable and are hugely inefficient and risky.
So why not improve the processes and give a chunk of the work to robots? They don’t get bored or dread doing the essential admin that builds up as deadlines approach. Could they help us make better decisions and spend more time using our skills?
Experimenting with RPA
RPA is pretty straightforward. Software robots — or ‘digital workers’ — are trained to run complex processes using the same IT applications, interfaces and access permissions as staff. They work fast and accurately, 24 hours a day.
Multiple robots can be seen as a virtual workforce — new capacity which works alongside your staff. One digital worker can be trained to help multiple teams, optimising its own workload.
These robots can be dumb — running through an algorithm until it encounters something unexpected or the process changes — or they can use AI to assess information and make decisions.
The technology is well established in the banking sector. Evidence suggests the benefits of RPA adoption can be significant; payback was reported at less than 12 months. Users also report improved compliance (92%), improved accuracy (90%), improved productivity (86%) and cost reduction (59%). [Deloitte figures]
To see if these benefits would translate to the Red Cross, and how we might scale this technology to help humanise our work, I ran a proof-of-value (PoV) project.
Will robots help our people?
A small team — Technology Innovation Lead, Technical Project Manager and Business Analyst supported by an agency partner — ran experiments over four months. We looked at:
· The desirability — what did our people think to working with robots?
· The technical feasibility — from infosec to server specs
· The financial viability — the volume of work versus the costs of setting up RPA
We examined the market and decided to run the tests with the Blue Prism platform. We then built a bot to help our Accounts Payable team, examining the impact on their daily work.
We also studied processes across the organisation, from HR to frontline programmes, to understand if there was a way to successfully scale this technology, while improving processes and data flows at the same time. We also examined the best way to ensure human oversight.
A communications and engagement strategy was tested and iterated to make sure that these robots would be seen as an asset instead of an invasion.
All the evidence was turned into a business case, which detailed the investment and resources needed to make RPA a useful tool. It was a good way to make the decision ‘eyes open’, which can be hard with a new technology.
Sharing the work
The impact was impressive. The Accounts Payable team are sent up to 400 invoices every day. They check the details against our records before they make the payment. It takes a lot of time, so we trained a bot to manage the five suppliers who sent the most invoices.
It immediately saved the team time. But the benefits went beyond the volume of work delegated to the robot. The people in the team were freed-up to do other things. While the bot checked the invoices, the humans could tackle backlogs and work on process improvements. For example, they managed to reconcile 100 statements in a single month, when they usually struggle to find time to do any.
They were also happier. The robot — which they named Optimus — ran the checks overnight, so the invoices were ready to be paid when the team started the next day. One person talked about the bot “doing all the boring tasks we had to do before we even got to the job we’re supposed to be doing”.
We immediately began to scale Optimus to manage a much larger number of suppliers.
Welcome to the team
We are now on the brink of rolling our RPA. The evidence convinced the leadership to invest in a new capability, lead by an automation specialist with the technical staff to build robots to support teams across the organisation.
It will be a gradual expansion, with automation applied to predictable areas of work like the finance, HR and IT functions first. But this will create the skills and experience necessary to support our frontline staff, hopefully freeing them up to spend more time face-to-face with our service users. And it will create capacity to respond fast when we face new challenges which place huge demands on our organisation, as Covid has.
I hope that RPA will allow our staff to use their time and expertise where it is needed most by freeing them up from routine admin and supporting their decision making with accurate data.
Over time, it may help constrain the growth of the organisation, as we will be able to add capacity without recruitment in some areas, but we do not intend to replace our brilliant people with robots or automate-out the human kindness which is fundamental to everything we do.
We all face new challenges as we emerge from the Covid crisis. There will be huge demand for charity services even as the sector battles to repair the financial damage.
But the crisis has also shown us that change is possible, and that technology is a critical asset for everything we do. It is the perfect time for new ideas, innovative approaches, and a reimagining of how we work to deliver social impact.
I believe robots can help us, humanising our organisations, streamlining our bureaucracies and allowing us to reach more people in need.