AI and outbound outreach: Menace or marvel for the modern contact centre?

Imagine a world where AI doesn't just assist but fully replaces human agents in contact centres. With ChatGPT's breakthrough launch in late 2022, this vision became more than just a dream for business leaders—it suggested an efficient new reality. The promise? Unmatched profits and a new era of customer satisfaction driven entirely by machines.

Now, as we look towards the winter of 2024, two years on from ChatGPT’s arrival, the fever for AI in contact centres has somewhat cooled. The promise has not been fulfilled.

It is my view that AI isn’t here to take over contact centres—it’s here to make them smarter. Harnessing the power of Large Language Models (LLMs) and generative AI, outbound contact centre activities will become finely tuned, ensuring every call hits the right person at just the right moment. Does this sound slightly unsettling? Perhaps. But soon, the relief of receiving only calls that matter will make us wonder how we ever lived without such a transformation.

A decade of change

The last ten years has seen outbound campaigning evolve from mass, unsolicited outreach to a more scientific, data-driven strategy. Random calls and messages that once alienated the public are now being replaced by carefully targeted campaigns, where every contact is deliberate and timed for maximum impact. Pausing to think before you dial, text or email someone is now the order of the day. Misdirected efforts are more than just annoying—they’re costly.

Artificial intelligence has also advanced during this time. Once concerned with decision trees and simple predictive analytics, AI now engages in human-like conversations and uncovers deep insights from massive datasets, transforming how we interact with technology and with one another.

AI in contact centres today

For all the advances seen in sectors from finance and fashion to healthcare and construction, AI’s seismic shift has yet to have a major impact on the modern contact centre.

AI bots suffer from what are known as ‘hallucinations’. These are incidents, admittedly relatively rare, where the AI model confidently returns erroneous or misleading answers, to the despair and confusion of those authoring even the most advanced LLMs. It is notoriously difficult to prevent such AI hallucinations, and the fallout from trusting bots over human agents could be catastrophic for companies that take the risk.

A more successful, if a bit mundane, application of AI in contact centres has been the streamlining of Quality Management. Voice Analytics automates the repetitive task of evaluating agent performance, but the advancements are far from revolutionary.

AI’s future potential

AI should theoretically pinpoint the ideal way to pitch a proposition and identify those who won't respond positively. However, the effectiveness of Large Language Models (LLM) hinges on the sheer volume of data they consume. The more data, the better the model. But what happens when data is limited? This, for me, is where things start to get interesting.

The most valuable asset that companies like Meta, Google and Amazon own is you. Your data. Every purchase you have ever made online is recorded. Every web page you’ve ever visited is kept track of. Store loyalty cards are rudimentary data collection devices. Alexa and Siri listen to every word spoken in your living room. We are sharing valuable information about ourselves on a daily basis.

This endless flow of data is being fed into the voracious appetite of an LLM and is already being used to deliver personalised ads in your browser. A simple Google search for a pair of trainers will almost instantly ensure that the next page you visit will contain ads for sports footwear.

With the rapid advancement of AI, this type of highly personalised advertising will become much more subtle and guileful, with the LLM deducing specific character traits and subliminally steering you towards products you never realised you wanted. I envisage that the leap from there into the world of contact centres will be an obvious and rapid one. The very same technology will be able to deliver unprecedented precision and accuracy for contact centre outbound campaigns be they voice, SMS, email or other channels.

Where do we go from here?

Think of an LLM as a genie that would not only grant your wishes but be able to predict them. Eradicating relentless, irrelevant sales calls for good may tilt the balance towards embracing AI in outbound contact centres and potentially render outdated regulations obsolete.

Over time, I foresee that we’ll get used to this technology to such an extent that consumers will complain when brands fail to anticipate their needs and flock to those that do. Like a powerful narcotic, the confluence of outbound outreach and AI has the potential to charm as well as harm us. Menace or marvel? The answer might lie in how we balance its power.

By Danny Singer , Founder, Noetica