An AI based virtual agent that can take on a real-time customer facing role in debt recovery and autonomously negotiate debt repayment is due for trial in mid to late 2019 thanks to an Innovate UK funded Knowledge Transfer Partnership programme.
Advanced Collection Systems’ (ACS) collaboration with the University of Hertfordshire has proven extremely fruitful. It has led to the design of an ‘Adaptive Virtual Agent (AVA)’ through the KTP, four separate Master’s projects, and the development of an AVA based commercial solution for the debt collection sector by independent AI development company Insight.
AVA is a machine-based debt collection software system designed to work alongside human debt collection agents. Through big data analytics and machine learning, the ‘intelligent’ software automates many of the calls, negotiations and transactions currently conducted by human agents, allowing them to focus on more complex interactions and expanding the capacity of the call centre to increase the volume of calls it can take.
ACS contracted University of Hertfordshire PhD student, Paul Moggridge, as the KTP Associate to work on the project and to assist AI development company Insight to integrate AVA into its own commercial solution. ACS has also worked with various departments at the University, engaging four Master’s students with expertise ranging from psychology to data mining, machine learning, programming and cyber security; each contributing to the unique, custom-built technology.
“For a business like ours, taking on a project of this size and level of complexity would usually be really difficult – add in the amount of research we’d have to source externally and it would be near impossible. The University of Hertfordshire has been an invaluable resource, equipped to support all facets of our business.” Martin Eaves, Company Owner and Technical Director, ACS
Initially, Martin was hesitant to enter into what seemed like a daunting process of getting the KTP underway: ‘A huge contributing factor in us deciding to undertake the KTP was the amount of support that the University provides in assisting us to complete our application. This could otherwise be extremely time consuming, but we were very well supported by the team at the University.’
Now 18 months into the 3 year project, Paul has found his role as the KTP Associate to have been highly rewarding: ‘It’s been a challenge balancing study with working full-time, but they feed into each other really well. I can apply what I’m learning during my PhD study directly to my work and vice versa.’
‘What I’ve also found beneficial is the opportunity to get additional vocational training that I wouldn’t have otherwise had during my studies. I’ve been able to attend a management school for two weeks, covering things like project, risk and test management.’
For ACS, the benefits experienced by their Associate have reinforced their interest in finding other ways to collaborate with the University.
‘Not only has our business gained a lot from the partnership in terms of very specific, tailored research, but it’s also been great to give back to the University and feed into the teaching cycle. The intellectual property generated throughout the project, as well as the data, are all fed back into the curriculum for related courses, so students are able to learn from the work that we are doing,’ says Martin.
ACS plans to partner with more of the University’s business and research services to work on other areas including designing the Cyber Security Centre to test the software in a secure setting.
ACS will use AVA to scale up its business and provide an enhanced service to clients. Insight, which specialises in providing AI solutions to the credit and collections sector, plans to integrate AVA into a full service repayment negotiator solution which it will offer to credit collection departments on a commercial basis.
Most recently, the Innovate 2 Succeed programme delivered by Exemplas, has been used to develop a plan to maximise the business benefits from deploying AVA within ACS and assisting Insight to identifying a strategy for the integration of AVA into its commercial product and follow on launch.