Insight of the Week: 3 Paths to the AI-First Contact Center

By Kerry Robinson

Bloomberg ran an interesting article on the impact of AI on the contact center agent?outsourcing industry in the Philippines?last week.

It included the following chart, showing the trajectory of global Business Process Outsourcer (BPO) stock prices. As you can see, there’s a bit of a peak at the end of 2021, followed by losses across the board in 2022 and then a divergence of fortunes, with Accenture and Wipro outperforming the others over the past 18 months.

The simple argument goes: Gen AI is proving its ability to replace junior agents, so the companies that supply them are becoming less valuable. But the reality is more complex.

The Bloomberg article points out that BPOs - and the government of the Philippines - are busy learning about and leveraging Gen AI. Just like we are. And our partners are. And you.

Hopefully, you’ve been persuaded that the future of customer service and contact center is AI-first: where AI agents handle most of the calls and chats you receive, escalating to senior agents for help. But what options do you have? I see three main paths to get there: roll your own AI, pick an AI suite, or use an AI-first BPO.

Roll your own

Nobody has a monopoly on AI capability. Anyone can get low-cost access to powerful, production-grade AI from big clouds like Google, Microsoft, Amazon, and OpenAI. ?For a few cents per minute, you can attach speech-to-text and text-to-speech models. In a few month’s time you won’t even need to do that, as the streaming multimodal models like Open AI’s GPT4o come online. But remember the 90:10 rule: it’s easy to build a cool prototype, but to get from prototype to production - and ROI – safely, is something else entirely. If you haven’t spent the last 18 months building your own AI runtime environment, AI application build process, AI simulation and evaluation tools and AI reporting, this option isn’t for you. You’ll get left behind. And as I talked about last week: AI delivers exponential advantage, so you can’t afford to wait. By all means, start rolling your own environment, but you need to pick one of the other 2 options to get started now…

Pick a suite

We’re starting to see Gen AI automation capabilities pop up in major Conversational AI platforms. Parloa, Kore.ai, Cognigy, PolyAI, and newcomer Sierra raised a stack of cash this year too. They vary a lot in terms of how much focus they’re putting on truly un-scripted Gen AI automation, but the fact is that these platforms now make it possible to build Gen AI-powered automation without all of the glue-code and cloud architecture and administration work you’d need to roll your own. If you want ‘one throat to choke’ when it comes to your contact center AI infrastructure, these could be a good option. And if you buy my exponential advantage argument, these platforms potentially allow you to buy back time and get deployed quickly. Doesn’t stop you rolling your own in future, if that turns out to make sense for your business.

AI-first BPO

As we discussed at the start of this article, the BPO market is experiencing a shakeup. I expect that those that survive - and certainly those who thrive - will pivot their business model towards delivering a large part of their services with AI agents (by rolling their own, or partnering with a suite provider) and/or they’ll leverage their low-cost human resources to do the quality assurance that’s required to keep AI agents out of trouble, or the tagging and evaluation that’s needed to optimize the AI agents, or even fine-tune their underlying models.

So how do you decide?

I think it comes down to the key question when it comes to any business decision: what’s core to your business? We can think of it from the perspective of customer services, and tech:

Is your customer service organization a differentiator? Maybe you provide high touch concierge services, or technical support. Maybe as a business, you win with customer experience?

Is it a profit center? Does your contact center drive customer retention and upsell?

If you’re nodding along here, then BPO probably isn’t for you.

What about rolling your own vs suite? That’s where you need to consider how core your technical capabilities are.

Are you a tech-forward organization? Do you differentiate with your tech and AI capability? Maybe you built your own recommender engine? Perhaps you use Twilio instead of a more fully featured CCaaS platform?

If that’s the case, think about rolling your own. But don’t let that slow you down. If you’re not inclined to outsource and you didn’t already roll your own, then you need to consider a suite like Parloa, Kore.ai, Cognigy or PolyAI. I’ve had direct experience with three out of those four. There are good reasons for choosing each. If you want to know my pick, you’re going to have to wait. More on that – hopefully – next week ??

Good thing is, with all the investment these AI suite providers are getting, if you do choose the suite option, every dollar you put in is being doubled-up (maybe even more) by deep pocketed venture capitalists, eager to win big in AI!

Kerry Robinson is an Oxford physicist with a Master's in Artificial Intelligence. Kerry is a technologist, scientist, and lover of data with over 20 years of experience in conversational AI. He combines business, customer experience, and technical expertise to deliver IVR, voice, and chatbot strategy and keep Waterfield Tech buzzing.

Subscribe to Kerry's Weekly AI Insights