Use AI Bots to Support Human Agents At The New Call Center. Home. | by The Conversation Consultancy Limited | Nov, 2020

On the face of it, things look dire, The bottom line is under pressure. Lowering churn could become a matter of life and death for companies. Each customer is gold. Advertising will mean less than WOM marketing in the future, so your front line has to be slick. But Covid and Better AI arrived at the same time to deliver companies a huge boost. Massive real estate savings, and quality chatbots that mean we can slash our human labor force and cut costs in half! It’s a win-win!

And those people have children, and parents to take care of. We’re insourcing a ton of these jobs already, and now we want to use the pandemic to put US and UK agents working from home out of work too? I got a better, more humane idea. Create better apps that make human agents more effective.

If Amazon can do it, heck, so can we. Amazon, despite its vastness, has an unbelievably responsive customer support infrastructure, which is the ‘impossible to reach’ benchmark of Covidian Era Customer Service. Sure it has the resources, but Amazon has automated very little of its customer service, except behind the front end, where it’s almost certainly made them stunningly efficient. If you want to speak to a rep, you pretty much can can. And first contact resolution of your problem is an expected norm. One of the reasons Amazon became a ruthless business-crushing monopoly is because it always gives the customer the white glove treatment. Plenty of jobs there.

Battered by the virus and its economic impacts, customers, bombarded with fake promises of “great customer service”, will walk at the first sign of anything but the best phone/online support possible. As aspirational purchasing drops, brand equity loses value, and customer service becomes a much more competitive differentiator, a vital engine of survival, recovery, and finally sustained profit. especially in big-box retail.

Only price and discounting can have a higher initial impact, but there’s a limit to how much largess a company can give on the giveaway front. Plus, with products and services expected to be more reliable, customer support will have to take up the slack become the key factor in building customer loyalty, WOM marketing, and brand equity. So The conclusion for company bosses should be a simple one –

…raise the customer service budget, increase capacity, boost salaries, give its executives more muscle, and widen the range of customer service careers.

Customer service should be professionalized, with best practices built around the agent rather than the overseer. More space per agent, better scripts, better soft skill training, better chairs, higher salaries, meaningful promotions, and four or five figure bonuses. And the Karma will be real. All this will make customer service that much kinder and more effective for the public. Customer support will no longer be a destination to dread. Instead it could become part of the joy of the online retail experience. Oh, and those “waste two minutes of your life so I can keep working this job” surveys? Gone.

Simply put, AI must support Agents instead of replacing them.

Better WFH tools, better scripts, better conversational training so the customers connect with them, and better sentiment driven organizational AI tools so agents can triage their workflows more effectively, and autonomously. Part of that autonomy would be to use a call queue algorithm to create more repeat + personalized relationships between human agents and returning customers. AI would provide conversational prompts for agents built from personal, occupation, location and other info that the customer had handed over during previous interactions. (eg, caller from Cincinnati + likely Football fan by demographic is asked about the college draft in general, and the Bengals picks in particular if it’s confirmed s/he is a Bengals fan during the interaction). This would cement relationships further, and adds a message to a customer waiting that s/he can talk to someone new now or wait until Steve is available. (this data can also be used to assess agent’s personability.)

What do we need to know about an agent’s personability? Because Agents with the right communications skills (starred in HR when their data hits assigned targets) could be offered sales courses. Yes they could double up as salespeople very effectively. If they pass, they get to leverage customer relationships, gently upsell existing multi-touch customers, and then flip screens to work incoming sales calls. Suddenly, a customer service career gets more multi-faceted and more profitable. That’s incentive to excel. (Head of sales won’t like it, until his targets are met in record time). Part of that familiarity could be a ‘lounge’ segment of the Customer Service team that literally just chats about products with customers, as if they’re friends comparing items they’d considered or had bought.

Suddenly the career vista of a customer service agent becomes inherently incentivized, so we can spare them the morale-destroying all-seeing-eye of the corporate panopticon that subjects them to those totally subjective “From one to ten how did I do today?” surveys, which suck up our time (for nothing) and make the agents feel worthless, with careers contingent, basically on how we feel that day. Personally, even if the service I get is average or poor, I always give top marks to an agent. How do I know that person wasn’t ill, or under excessive pressure, or just exhausted?

The poor bots will have to work way way harder 24/7/365 for the same amount of electricity. It’s an outage!

Online or on the phone, Let’s quadruple the bot workload — toggling between 1) customer-facing and 2) agent data-retrieval support after their customer is put through to a agent. Also the bot is still part of the conversation — as 3) the agent’s ‘butler” during the customer call. Let’s make the data work for the interaction creatively. If the same person calls, he gets the “same” bot loaded with all that person’s call-supporting data, the same assigned character traits, back story, as well as character/sentiment analysis for the agent. In other words, with just a handful of defined / neutral, randomly assigned character traits it could easily seems as though every bot instance was unique by name and nature. (Some of them might even complain about how hard they have to work these days), just for fun. And one of those extra duties will be…

Phone customers could be given the option of starting their interactions with the bot or listening to the muzak, until an agent is available. If they choose to wait with muzak, an IVR prompt could use various approaches to having a bot enter a relaxed conversation with the customer to pass the time. IN some scenarios, no user data questions should be asked, unless the caller asks for them to be, or proffers personal info. The aim of the exercise is small talk, banter, and bonding, because when the human user gets put through the bot is still in on the interaction.

With all these changes, customer service becomes a much more engaging, AI supported human experience and the connection between marketing, branding and customer service, could be less a matter of using friendly customer service characters to sell product, but showing the power of customer service to demonstrate how much the brand cares for you, the customer.

From a wider economic perspective, whether you see the world from a GDP or happiness index perspective, more humans in more jobs, means more demand, and better lives. For governments and companies, it’s time to look at more than the P + L chart. Remember productivity? The bots serving humans approach can really help boost productivity. And if you’re thinking long-term, which should be, there’s no better metric than that.


More Posts

Send Us A Message