In our role as consultants we work with a number of different businesses in different industries, and what is always interesting are the different views different organisations and industries have on the same issues.
Matt RiddallSVP, Client Solutions, arum
In particular are the different views around power dialling (or the "left message machine" as it can also be known as by many frustrated Collections agents and managers).
With the proliferation of channel choices available at the moment coupled with a heavy regulatory overhead in relation to using predictive diallers, we see some organisations who are looking to phase out their use completely and instead use automated solutions to place right party contacts to a collections agent (call fronting), without having the overhead of a large number of agents in order to make a traditional dialler efficient. On the flip side we still see some organisations where effective dialler management is still in its infancy, with a desire to do more predictive dialling.
Predictive diallers have been around for years - I used my first one back in 1989 - and while manufacturers will claim their technology has come on in leaps and bounds most users will agree that they still have some serious shortcomings, especially around voice detection and even simple reporting. I find it incredible that some diallers today still do not report penetration and saturation as standard metrics, for example.
Let's make some numbers up. If your right party contact rate is 30%, what is happening to the 70% that have not been contacted? No doubt they have been hit with letters and SMS in addition to the call fronting or predictive dialling, so what happens next? How do we get to this 70% - where the big opportunity is?
For me organisations need to look at how good their overall contact strategy and channel management is and what contact methods replace those that have failed (an effective "no contact" strategy). Without doubt the call fronting approach has great benefits in the high volume front end collections space by reducing the number of agents required as they are only handling right party contacts, but we continue to see different results in different organisations with some saying call fronting works better than the dialler and some saying the opposite.
Let's not forget that not everybody will react positively to call fronting or any other type of automation. How many of us would truly give personal security details to an unknown machine? Many customers will still expect to speak to a human when they pick up the phone, or immediately hang up when a machine makes a call. This is unavoidable and organisations should recognise this and ensure they react accordingly - to satisfy customer experience and to mitigate against losses.
And how far through the delinquency cycle should a customer travel before this is recognised? In my last job before I was a consultant, my pre charge off team at circa 150 days delinquent was achieving a higher right party contact rate through manual dialling than the dialler was achieving in the 1/29 day delinquency space - a last wring of the cloth before the account charged off, but effective nonetheless. Segmentation and profiling for this type of activity earlier in the delinquency cycle will undoubtedly provide benefits, but there is an increased cost so it has to done right.
So should organisations continue with traditional predictive diallers or instead move to newer call fronting technologies, with a hand-off to an agent, as a dialler replacement?
For me, the answer is that traditional dialling techniques, especially using power dialling technology in progressive dialling mode, or manual dialling as a backup, will still be around for a while, although the technology may look different (cloud based diallers for example). Overlaying analytics and segmentation or balance / delinquency cut offs will add substantial value. But the essence is in human interaction to review accounts and to identify additional contact opportunities - this will always play a part.
From the same author
More from Matt Riddall.
Digital and AIWhat three weeks at MIT taught us about building agentic AI
Ella and Matt took MIT's agentic AI course to pressure-test the AI-assisted system arum was already building. Here is what held up and what did not.
Matt Riddall and Ella Grice · 17 June 2026
Digital and AIIs your AI agent actually ready? 11 questions to ask first
Eleven questions to answer before an AI agent handles a customer in arrears, from a team that has been building them since 2019.
Matt Riddall · 2 April 2026
Digital and AI7 common mistakes to avoid in collections implementations
The seven mistakes creditors make most often when implementing a new collections system, and how to avoid each one.
Matt Riddall · 30 July 2025
Keep reading
More from arum.
Digital and AIBuy vs. build for debt collection platforms
In today's economic and regulatory environment, transforming collections capability is high on organisations' strategic imperatives.
Reuben Gates · 13 May 2026
Digital and AIPreparing for AI in collections and recoveries
Artificial Intelligence is rapidly moving from concept to practical application across financial services.
Forid Meah · 16 March 2026
Digital and AIShared ETL code libraries: improving data quality and delivery
Shared ETL code libraries stop the same logic being rebuilt in every pipeline, and lift data quality and delivery speed in collections operations.
Stephen Wright · 19 February 2026
Talk to arum
