Efficiency or effectiveness?

I got a cold marketing call yesterday on my mobile.  Nothing new about that you probably think.  It was from a double-glazing company – which at least in the UK is also nothing new.  My usual response to such calls is to ask what they’re calling about, then having determined it’s a cold call, ask them where they got my number from and to remove my details from their database, but this caller was not interested in having a discussion or responding to my questions.

The call started ‘can you hear me?’ which did not raise any immediate suspicions, but as soon as I responded, the caller launched into an introduction, and from that point on was not going to be interrupted until it finished with ‘do you have any doors or windows over ten years old?’ at which point I terminated the call.  As I had tried to interrupt the flow, the fact that the tone and delivery of the caller did not change made me realise that I was being called by agentic AI.  Any response to the first question would have set off the sales pitch – I certainly did not answer with an appropriate reply, and the closed question at the end of the initial pitch was obviously intended to steer the remainder of the call in a relevant direction for the next step.

The experience concerned me at two levels.  At a personal level, it is a sign that I am now going to get nuisance calls from bots in increasing numbers, none of whom are going to be deterred by my protests, threats or ending the call.  Might I end up in a situation where I can’t make and receive the calls I want to make, because the phone is permanently busy with AI-generated nuisance calls?  The second is the implications at a business level in automotive retail, where there are a huge number of agentic-AI applications coming to market against a background of continuous pressure in manufacturers and retailers to cut costs.  The siren call of a vendor who promises a perfectly compliant, always cheerful, hard-working agent who will follow up on leads or work a database from 9am to 9pm seven days a week for a pittance and no commission sounds like an offer you can’t refuse.  But I believe you should.

With my dealer hat on, we have been an early adopter of AI because of the huge volume of leads that we have been receiving, the time pressures on staff no matter how many we recruit, and because you wouldn’t expect a business run by the MD of ICDP not to be at the  leading edge.  One application sits in the background and does an excellent job of recording and assessing all the phone calls, creating and savings summaries onto the CRM and generating a proposed email response.  It struggles occasionally with heavily-accented callers to correctly capture names, and we need to make sure the sales and service team are using the IP phone system and not their mobiles, but basically we are very happy users.

We have a second AI application which is agentic-AI so trying to directly replace a human with an agent, to route incoming calls and handle out-of-hours calls.  This is proving to be much more challenging to refine and set up so that it supports customers without leading to frustration on their part.  I am not convinced that we will ever reach the stage where the edge cases of strong accents or unusual requests are totally eliminated, but certainly in the early days we could monitor calls and find customers getting frustrated with the AI agent not understanding their request or handling it in a way which was clumsy to some extent.  Any level of frustration is undesirable, but if it gets to the point where the customer finds the experience to be negative, then it is unacceptable.

I think we will address that with our inbound calling as in the end, the customer is to some extent already committed to speaking to you as a current or potential customer, and there are also clear advantages that the agent will handle the calls promptly at all hours, seven days a week.  However, if I come back to the use of agentic AI in outbound calling, then that connection to the prospect may be much weaker, and the potential frustration higher.

I am concerned that manufacturers, dealers or repairers will succumb to the appeal of a tireless agent with a permanently positive attitude that becomes first point of contact with new prospects.  In the product demonstration, I am sure the agent will perform perfectly because the target prospect will have been selected so that they do not respond as I did to the call from the double-glazing sales agent, but be compliant with responding to the questions in way that always gives a successful outcome.  However in the real world, the calls will be at a time that is inconvenient to the prospect, or they will have forgotten that they ever made an initial enquiry that generated the call, or they will have now decided that they’re more interested in a different model and can’t interrupt the agent’s flow trying to book them in for a test drive on a model they’ve now decided they don’t want.

Given how quickly technology moves on, I hesitate to say ‘never’ but I will want an extended trial period before I replace our two living, breathing outbound sales agents with a bot.  I will happily carry that employment cost rather than have a bot working in the background, recording ‘lost sale’ without me understanding that this was self-inflicted.

Image: MS online sources