UK roadside assistance and vehicle recovery provider Green Flag Breakdown is part of the £3bn Direct Line group, a brand that specializes in selling insurance and other financial services directly to consumers by telephone and the Web. Operating what it calls a "smart network of rescue specialists to help get you moving again, no matter what," the company is one of the UK's largest roadside assistance businesses. On average, it rescues 2,000 customers per day and has over 3 million customers using its services.
An unmistakably successful business, then. But as Green Flag's Technology and Delivery Lead Shakeel Butt explains, a major push has been going on in recent years to rebuild all of its core legacy IT systems, which his 135-strong team has been doing with the help of UK-based technical consultancy Contino.
Butt told us that this technology transformation-porting all core systems from the older architecture into the cloud and simultaneously transforming them into a fully API-based, micro-services architecture-is seen as the very necessary foundation of a wider group-wide business transformation), as well as being expected to be an asset set to yield significant cost savings and additional revenue streams in the coming years. It is also, it turns out, merely the first step in what is envisaged as a comprehensive, five-year long, tech and business realignment for Green Flag Breakdown.
50 years of legacy
Ambitious stuff, but Green Flag customers hadn't been known for being unhappy prior to all this: so what had prompted such a serious project? Butt says the decision didn't come from IT at all, but from the very top of the organization:
A new MD, Dean Keeling, joined Green Flag about four years four years ago, spending the first 12 months just understanding the business and pulling together his plans to transform it. But he kept being told technology was the blocker to progress, no matter what part of the organisation he was talking to. And with any 50-year-old business, there were indeed a number of problems there: it was very old, it was expensive to maintain, there'd been no continuous investment across the estate over the 50 years-new bits added, old bits removed.
It was quite a heterogeneous set of different technologies that had been plugged in, there were also some stability issues for various systems, and all in all it was very difficult to improve and enhance.
What did "the zoo" look like? On one end there was a rescue platform, or claims engine in internal parlance, that was a mainframe system with an Oracle database backend, where all rescues got managed. There was then a second but separate mainframe application for policy sales and management, plus finance systems at a group that had to be integrated into and fed into for reporting for Business and Management Intelligence systems, as well as a couple of pieces of software his frontline colleagues used to support their customer-facing roles. Butt says:
So you've got a lot of legacy there, a mix of custom/bespoke and in-house written applications and off the shelf systems as well, all working together. There's complexity there, but then you have the other challenge of continued knowledge of all of these systems; people move on, knowledge starts to deteriorate within the organization, and so on.
Butt states that he quickly found himself being directed to supply a tech stack to support immediate transformation today, but which would also future proof the business for tomorrow. He adds:
Once you've got the business moving to the place that you want it to go in, there's all kinds of ideas and new products and services and partnerships that come to the fore. This is all about the business knowing that the technology is flexible and future-proofed so that it can pivot and move with the business, based on its specific needs.
Up and running in five weeks
This all got off to a strong start, he said, with the new serverless public cloud platform, which was up and running in a mere five weeks. That quick burst of speed came after a careful evaluation and planning process, however; the initial business case was put together in early 2018, again led by the MD who crafted what was called a value-creation plan for the business. All this had to go through the Green Flag Breakdown C-suite before ultimately going to the board to get approval and funding.
And a jump straight to cloud was far from being the only option, either. Butt says:
There were a couple of options that we looked at, because we needed to conduct full due diligence, like refactoring what we'd got already versus building from off the shelf products, or greenfield-building a completely new stack from the ground up.
What finally made this third approach the most attractive was the idea of instead of getting everything done by third parties, which a business of Green Flag's size would traditionally tend to do, all the new technical capability would be recruited for in-house, and the entire transformation would be run as an agile operation at scale, with the new teams then transitioning from build to run, immediately looking after the platforms that they had personally built.
Once this pathway was chosen at the end of 2019, recruitment commenced while a suitable vendor partner for augmenting and accelerating development and build was begun-with that choice being simplified when Butt and his team asked the chosen cloud platform, Amazon AWS, for recommendations, as Butt didn't really want to go to a very large, multi-national faceless consultancy but "someone that was similar in size and ambition to us".
That AWS choice was also very deliberate, incidentally, he says:
One of the things I didn't want to do was to be juggling with lots of a multi-tenancy cloud landscape, because that just creates complexity and workload for engineers, who have to remember how to work with the Azure versus GCP, Google Cloud versus AWS. One of the principles we have also adopted is that if there's a service that we need and AWS has it natively, we'll consume that service rather than build it ourselves, which is acting as a massive accelerator.
Real-time access to lots and lots of data
It's early days for this very new way of working at Green Flag, but what has the impact been so far? Butt declares that change in that core replacement claims engine had historically taken at least six months; since the move to cloud in Summer 2020, 300 changes have been recorded already, while so far zero outages have been reported. And for the call centre team, because the main front end application has been completely redesigned, a 60% reduction in call time needed with the customer has been achieved, as there's no longer any need to navigate multiple systems and access more than one screen. Butt explained:
That's a massive benefit for all our customers, because when you're at the roadside looking for help, anything that we can do to shorten that call time to get the technician to you quicker is going to massively impact your experience.
In terms of next milestones, this year will see the launch of a new sales and service application, final replacement of all legacy, a cloud telephony solution and then optimisation, which may well centre on the use of machine learning to help automate and recommend better recommendations. Butt added:
We've now got real-time access to lots and lots of data, which we never really had before; our MD always used to describe Green Flag as historically very data rich but very insight poor. That's now changed.
So next for us is understanding what benefits we can deliver from the insight that lives in that data, and for me the big one is how we can use big data and real time data to drive benefits and make lives easier for our customers and colleagues.