‘Do 10 things, see which five work’: How media should tackle the rise of AI
MediaCom global head of paid social Renee Mellow chaired a panel on how AI and messaging are influencing buyer behaviour at this year’s ‘Commerce Disrupted’ event in London.
A key topic of discussion focused around whether AI would replace human workers, to which Digital Genius founder and chief executive Dmitry Aksenov felt that the tech was not currently good enough to replace workers, only to assist.
Looking at chat bots, Blis vice president of EMEA Paul Thompson described machine learning as the gateway to AI.
Why Social managing director and founder Dom Burch recounted Asda’s experience dealing with snow around six years ago demonstrating the need for a mechanism to talk to lots of people in real time. “Ultimately the customer doesn’t care about technology – they just want food,” he said.
Thomson TUI UK and Ireland head of innovation Joanne Hickson discussed her company’s strategy to test whether a chat bot with an element of personality would take in feedback and help customers get to their ideal holiday quicker. She referred to the chat bot as the company’s “secret sauce”, helping a new type of consumer that wants a very different thing from a holiday.
“We’re future-proofing the business,” she added.
Aksenov commented that a chatbot need not stand alone, and should be complemented by normal human messenger conversation, with an aim to get 60-70% automation in the next few months on the most common questions. He also stressed the importance of fitting data back into a system.
According to Burch, consumers will always attempt to game systems to get higher up lists, looking back to when [UK TV show] Watchdog’s Anne Robinson used to get copied into emails. However, Aksenov said Unilever had counteracted this by introducing a one minute delay, creating less reason to play and try to break the AI.
Burch discussed when businesses first accepted that social media was here to stay and the importance of funneling people to one place for customer service interactions.
“Being quite narrow is quite smart thinking,” said Burch. “Do 10 things, then see which five work. Keep it narrow and then give it a go but don’t try to rip up your contact centre and get everyone to go to messenger just yet.”
Voice search and control
The panel agreed that voice was a really interesting space, both in search and control. Burch continued that by looking at the biggest pain points, like interacting with the public services, businesses could learn from commerce and services.
“A lot of kids end up in A&E because they don’t know how to use an inhaler – that might be quite a good use of tech,” said Burch.
Metrics that the panel were looking at, aside from the obvious, included measuring oestroes after AI deployment to demonstrate consumer satisfaction, and emoji analysis, particularly as people move into messaging environments.
Going back to the topic of AI taking jobs, Burch added that the decision is really down to those in charge as to whether they reinvest the money saved into other bits of the company, while Hickson reinforced that she didn’t see computers replacing staff with 40 years’ experience in the business.
Aksenov concluded that AI should allow agents to do more creative work, like spending time trying to up sell and build relationships instead of just answering the initial query, opening up revenue generation opportunities.
To read the original article click here.
Tags: AI, Commerce Disrupted, Machine learning, Panel, Paul Thompson