Suncorp automated processes that were previously executed through a business process outsourcer in India after losing its offshore workforce to COVID lockdowns in March.
Insurance automation delivery manager Vicky Leo told Automation Anywhere’s Imagine conference that her team worked on several “fast, tactical short-term projects” to deal with the fallout from COVID-19.
Some of that work enabled Suncorp to continue to function despite losing "the majority" of its business process outsourcing (BPO), a scenario never envisioned in the company’s business continuity plans.
“Our business continuity planning at Suncorp never accounted for a BPO partner ... located in India going into lockdown due to a global pandemic,” Leo said.
“We'd always relied on being able to shift work from one location in India to another, so when we lost the majority of our offshore operations workforce in March, we had to act fast and relax some of that stringent governance that we had in place to get stuff done quickly.
“In the past few weeks my team has been incredibly busy … automating all of that work that we needed to do because we no longer had a BPO doing it for us.”
Leo’s team - and, in particular one robotic process automation (RPA) developer - also spent a weekend standing up a bot to handle the large volume of financial hardship and relief requests that Suncorp fielded from its customers as the impacts of COVID were felt.
“One of the things that I love about Suncorp - and why I've worked for them for 12 years - is its fast decision-making when it comes to assisting our customers in a time of need, and this came in the form of premium relief to our customers suffering financial hardship due to COVID-19,” Leo said.
“With this decision also came very heavy manual processing for our policy operations area. We didn't have the labour force to process hundreds of relief requests coming through every day. And this is where my team came in.”
Leo said one of her senior developers - “who I won't name because there's a lot of poaching that goes on in this industry, especially in Brisbane” - spent the entire Easter weekend standing up a robot “to process hundreds of premium relief transactions for our customers.”
“So when the customer rang up to cancel because they'd been financially impacted by COVID, we offered them a 25 percent discount or a three-month waiver for their policy,” Leo said.
“The actual processing of that - so the balancing of the policy and the movement of the money to our write-off account - took about 10 minutes per transaction, and we were getting hundreds through each day.
“We would have needed to find an extra at least 10 people to process the volume coming through. [Instead] the robot is now processing 75 percent of those transactions, with the 25 percent of exceptions going to a human team to process.”
Automating 6 million transactions
In insurance, Suncorp started testing robotic process automation (RPA) in late 2016 via a partnership with IBM, which built the automation in Blue Prism.
Leo said the proof-of-concept automated “a fairly simple insurance policy process where the input was a list of transactions in an Excel file from our motor dealer network that required processing into our policy mainframe.”
The project ran for three months.
While it achieved “everything we had set out to accomplish in terms of automation rate, we realised along the way that this was something we didn't want to rely on a partner for, and we made the decision all the way back then that we wanted to build an internal capability,” Leo said.
“Automation was absolutely something that we wanted to get seriously behind - not just dip our toes into - and we could see the potential for our organisation from the start.”
At the beginning of 2017, Suncorp started to spin up what would become Leo’s team.
“Structurally, the decision was made to sit within our operations department,” she said.
“We knew that we would need to be closely aligned with our IT team but more importantly we needed to sit in the business, as this was where most of the opportunity sat.
“Additionally, we borrowed a couple of developers from our IT area, and I handpicked a couple of people from our business improvement team who I knew were process-minded and could process map to the right level of detail we would need for RPA.”
The team spent six months recreating the proof-of-concept bot in Automation Anywhere, together with a second bot, “while setting up what would become our ways of working, our governance framework, and our operating model.”
“The first few processes we automated generated significant capacity savings, but more importantly, and what we didn't realise at the time was, it halved the turnaround time for both the processes that we automated, and we lifted the quality to 100 percent,” Leo said.
The team also spent several months building a pipeline of internal processes that could be automated.
“It was at this point where we actually went to the Group [for money].”
Busy bot building year
The 2018-18 financial year would go on to be a major year for action on the automation front within Suncorp’s insurance business.
“This was probably one of the hardest years of my career and I think if we'd kept up with that pace of delivery, I'm not sure I'd still be in the team, let alone Suncorp,” Leo said.
“But this was actually the year that gave us most of our learnings. I'm super proud that we managed to deliver 22 unique processes in just under 18 months.”
One of the bots delivered in that year was ‘Ruby’, a workers compensation wage payment robot.
“The challenge that we had with Ruby was we had a process that had been functionally aligned to our operations area,” Leo said.
“The process was to pay injured workers a wage amount weekly or fortnightly, and we needed to reduce the cost of that process and decrease the turnaround time because often they were taking longer than ... 24 hours.”
Leo said Suncorp wanted to be able to run the process outside of business hours as well.
Once stood up, the bot automated 90 percent of wage payments, or so the team thought.
“Ruby logs into our claims system, just the way a human would, picks up activities directly from her own queue, and she makes between six and eight validations before she actually makes the payments,” Leo said.
“What we didn't realise was we thought we were automating 90 percent of those wage payments, but what we were actually automating was 90 percent of the wage payments that the claims handlers chose to send to our operations area.
“It was only after six months that we realised that we were only getting about 50 percent of those payments, and it was after we automated that the claims handlers had more faith in the robot than they did in our human workforce to get payments paid accurately on time, so we had to spin up another bot to manage the volume coming through.
“We basically doubled the amount that we had set out to, within six months of going live, which was a really good achievement.”
Today, Leo said Suncorp remains focused on process and decision automation.
“We have automation teams in each of our lines of business: my team being the biggest in insurance, then we have another one in [the] bank, one in New Zealand and then one to manage group services,” she said.
“We've automated - at last count - over 6 million transactions across the Group a year, which is something that we're really proud of.
“We've made strong progress in automation of back office processes but we've been a bit slower in augmenting customer-facing roles so that's where we're starting to turn our attention to now - spending more time in our contact centres and that 'human in the loop' approach.”
Not about 'cost out'
Leo recommended against basing a business case for automation around ‘cost out’ - that is, reducing staff numbers and therefore labour costs.
“You will absolutely gain [workforce] capacity back, which you can release to decrease costs which is a major driver of automation, but don't get hooked on making this the sole focus,” she said.
“Think about the things you can do with the additional capacity: what processes don't you do today that you could introduce to delight the customer? What are the things you put on the backburner because you've never had the capacity to do them before?”
That mindset also made it easier to deploy the automations in the first place.
“[It's] much easier to roll out change when culturally people are excited about what you're doing,” she said.
“You're giving them capacity back in their day as opposed to threatening their jobs.”
Additionally, Leo said there were “less tangible benefits” to automation that should not be minimised.
“A robot is only going to execute the steps it's programmed to do, and it won't deviate from the process the way a human will,” she said.
“This means 100 percent quality and consistent processing every time.
“There's no rework, there's less complaints, there's no leakage, [and] there's no privacy breaches.
“Personally, I would take all of these things over cost-out any day.”