Client
This client is a leading UK life insurance broker, protecting more than 1.8m lives for the past 25 years. A 250-strong, UK-based customer-facing team offers support and advice to around 50,000 customers every month.
Challenge
A lot of the work they did was time consuming – sales teams worked in siloes and cherry-picked the best leads. The aim was to centralise all data, modernise the tech and implement customer-led and data-fed processes. Searching for that North Star – strategically implementing a modern data platform to deliver connected customer (and staff) experiences.
Starting out
We built the necessary infrastructure and created ML processes, pooling various code and lead data from legacy, bespoke systems into an easily accessible central hub. This newly built infrastructure was first used to test a foundational propensity model, scoring lead data and delivering this to contact centre agents to ensure the best leads were prioritised for contact sooner.
But these changes didn’t happen overnight – gaining internal stakeholder buy-in was crucial for success and this took time. A six-month pilot scheme, with dedicated ‘process champions,’ was put in place. The pilot was hugely important in terms of winning the support of the customer advisors and team leaders who’d be using the new data-led lead management processes and ensure the financial success of the initiative.
Pilot problems
Full disclosure, it wasn’t all plain sailing; the initial impact of the pilot wasn’t wholly positive. Income fell to around half the level of the previous month! There were many reasons for this but trying to launch a completely new way of managing leads without properly understanding and then adapting existing working practices clearly didn’t work.
Despite extensive business process and customer journey mapping ahead of the pilot, it turned out over many years there’d been several undocumented ways of working. Once the problem was realised, the issue was ironed out and the situation turned around.
At this point, it would’ve been easy for senior leadership to lose faith, but this wasn’t the case. There was enough positive feedback coming from the advisors to give the senior team enough hope to stick with the pilot.
Achieving results

Just four months later, the team’s revenue was exceeding expectations (up 40%!) and confidence was fully restored. Six months after the trial started, it was rolled out for real and the positive impact was immediately evident.
In four weeks of launching the ML propensity model, there was a 16% increase in conversions for the customers the model was applied to, which has so far generated £400,000 in additional revenue – equating to an impressive £4.89m revenue boost over a year.
Buoyed by that success, we then used AI to display information on monitor dashboards, ‘StarScreen’, around the office, for the first time enabling the visualisation of near real-time performance data every 15 minutes. The data has helped democratise information around the business and ingrain data in the day-to-day culture of the advice teams.
A similar integration allowed for more timely and reactive automated customer communications. The data feeds into a communications system, triggering appropriate messages and providing daily analytics of channel interactions, such as opens, clicks and consent preference updates.
Optimised lead scoring throughout the day has helped deliver the most appropriate leads to the right team members. This, along with integrated communications, led to better outcomes for customers and the team feeling more motivated by seeing up-to-date performance metrics. There was less downtime and greater accuracy too.