TrackBack extend technology lead with AI and natural language processing

At TrackBack we are proud to be regarded as the European leaders in tracking and reporting to OEMs about the lead follow-up activity of their dealers. We are also excited to announce our R&D results in natural language processing and AI.

Whilst lockdown has typically meant a reduction in output for many businesses over the past few months, we decided to use this period to experiment with a variety of leading-edge technologies to push the boundaries of what they can learn from the contents of dealer calls and emails with customers. As 2017 winners of a Queen’s Award for Innovation (and a second Queen’s Award for International Trade) we have always believed that whilst it’s important to know when a lead has been followed up, and the sooner the better, it is also important to discover how that lead was followed up. Commenting on the results of this development activity, Gareth Thomas, Managing Director, added “a high quality customer experience is usually the difference between a sale and the dreaded  ‘I’ll think about it’ response”.

Whilst the TrackBack SCAN service, where highly trained analysts score selected calls, is popular in helping to identify training needs, we realised there was also demand for a faster and much higher volume automated version of this specialised analysis service.

“We already knew that there were a number of ways to transcribe recorded conversations” he added, “but what we needed to find out was which method was the most accurate, and then how might we automate a deeper discovery of what was, or was not being said”. There are two main reasons for this project; Compliance, with regard to whether certain things were being asked or checked during the conversation for legal and regulatory reasons – vital in the financial service sector where heavy fines can be applied to dealers who were discovered to have failed to ascertain whether the caller was talking to right person for example – and Best Practice, where failure to offer a test drive or invite a visit to the showroom for example, might result in a weakened sales opportunity. “Our SCAN service can assess an infinitely wide variety of factors including the tone and sentiment of the conversation. Automation can’t yet report to that detail, but our development work has proved that we can detect with a very high degree of accuracy, and in near real-time, whether specific questions were being asked and topics discussed. This then allows us to alert very quickly not only the OEM about compliance levels and other performance measures across their network, but also specified people at dealerships who can immediately investigate possible issues and take action to improve customer experiences.”