The CEO of Observe.ai, Swapnil Jain, recently spoke with No Jitter to discuss the idea behind the birth of the Intelligence Workplace platform, what AI can teach us about sentiment in the contact centers, and paints a picture of what an AI-based contact center might look like. Jain also shared a sneak peek into what Enterprise Connect attendees can expect in a few weeks at the show.
Responses have been edited for conciseness and clarity.
In October 2021, Observe.ai unveiled its Intelligent Workforce Platform that aimed to optimize agent performance and automating repeatable workflows. How has it addressed and resolved challenges that contact center managers are facing? What are some of the challenges you’ve observed?
SJ: Before I started the company, I spent a lot of time in contact centers in India and the Philippines. For about a year, I sat on those floors, watching what agents and supervisors were doing. I then realized the contact center world is still very clunky, using legacy on-premises systems and manual software. So the idea was, how do you build a company where the software for these people in the contact center helps them do their job better and faster. That’s how we started the birth of the Intelligent Workforce platform.
The persona that we’ve already made intelligent by a workforce platform is the quality manager persona. For example, most of us have called our bank or mobile service provider and hear, ‘Your call is being recorded for quality and assurance.’ Contact centers record these millions of calls, then a quality manager goes and randomly selects some calls, downloads them on their laptop, listens to them, fills out a spreadsheet, then emails that spreadsheet to a supervisor who would then run the coaching. The process is 100% manual and costly because humans are doing it. They can only analyze one percent of interactions.
Giving quality managers a tool that analyzes 100% of interactions happening in the contact center using speech and natural language understanding can help them find what they were looking for, then automate the entire workflow. Now, the job as a quality manager isn’t to download files and listen to them but to analyze the data that Observe.ai generated to write effective coaching plans for my team. I can also do this faster, cheaper, and I can do this, not only on 1% of the calls, but 100% of the calls. I'm also able to provide much more data driven, faster coaching to my agents.
The second persona we’ve made intelligent is a supervisor persona. The supervisors are coaching the agents once every month or twice every month based on the spreadsheet being sent to them from the quality assurance team. With Observe.ai, all the data is available in the dashboard. Supervisors can quickly come in and write effective coaching programs for their teams by automating their workflow so that they don't have to do anything manually. The only job remaining is to use the data and practice coaching and behavior change conversations with their teams.
In the same release, you said, “Our goal is to usher in a new generation of contact center workers— AI-empowered, ultra-productive, and constantly improving.” What steps is Observe.ai taking to achieve that goal long term?
SJ: The contact center workforce is agents, supervisors, quality managers, and compliance people. Our mission is to make software for each of these personas so they can do their job ten times better and faster using artificial intelligence (AI). We truly understand the agent persona and want to do the same thing for them [as we did for supervisors and quality managers]. [That is] build a new kind of agent application empowered to use conversational AI and have automated workflows integrated, so [agents] don’t need to put you on hold to solve problems. That’s the direction we’re going.
We also went omnichannel last year—starting with voice [channel] at 70% of the volume today. New-age companies are omnichannel and require chat and email support. With the acquisition of Scope.AI, we'll be launching more channels this year. I think by going into omnichannel, by focusing on more personas—we have been able to realize our vision.
What AI capabilities could contact centers benefit from that don’t currently exist but might in the future?
SJ: I expect conversational AI to continue, and agent improvement software will continue. We’ll also start seeing more proactive reach-out from the business to consumers—and that’s one area we don’t [currently] see.
I also feel like we’ve had this debate in the contact center for quite some time around—‘Is AI here to automate or augment? I believe AI will do both and take care of simpler tasks. There’s a vast majority of tasks that exist which machines don’t do right from this perspective. Let’s say you call your mobile service provider for a certain problem. You’re a well-informed, educated buyer—probably using an iPhone or Android device. You already have the application, know how to use it, how to log in to their website — so you’re not calling for these simple things that you can do in the app or on the website. When you’re calling, it’s for something critical, a complex task — something that requires an agent, that requires empathy.
I think people will come to accept that AI is going to automate simple tasks. But at the same time, you need to build tools for the agent to do their job better for situations when machines can’t solve problems.
What can AI teach us about sentiment in contact centers?
SJ: [We can] learn from the data and enable your contact center agents and supervisors to be more empathetic when talking to the customer. Perhaps a customer says, ‘”Hey, I’m stranded at this airport, my flight is canceled, what do I do?” You take a moment to say, “I understand what you’re going through. I apologize for the experience.” These are very small things that agents can do to connect with a customer having a problem and ensure that someone will solve that problem. What we've learned over time is we're experiencing customer frustration at high speed. Regarding consumer expectations from a business— consumers want (and expect) everything to be perfect, all the time. Obviously, things can’t be perfect.
At Observe.ai, we help agents by finding situations in interactions where the customer is frustrated or angry. We’re able to assist the agent by identifying those specific moments and informing the user of situations where they need to be more effective. A supervisor could then coach the agent and say, ”This is where you have to speak slow. To build trust, build empathy.”
What can Enterprise Connect attendees expect from Observe.ai?
SJ: We’ll have some big product announcements to make, which focus on the true use of AI and automation. In terms of the overall roadmap for 2022, we heavily focused on understanding the conversation post-interaction and helping the agent. Throughout the year, they can expect announcements around real-time agent assist, and real-time agent applications.
What’s one thing Observe.ai did particularly well last year that you are most proud of? What could Observe.ai do differently over the next 12 months?
SJ: I’m proud of our innovation on the product side with the Agent Performance and Coaching launch and further innovating that as well. I'm proud that we have maintained through product first company that our customers have loved and that will help us you know, continue building this company to large scale in the contact center market.
Blue-sky question: We’ve seen AI embedded in everything from fitness equipment to refrigerators. Can you paint a picture of what an AI-based contact center would look like?
SJ: Let me start with the ‘AI everywhere’ concept. I feel like this is one of those things where people are taking a hammer and trying to find nails. What do you do with [AI?] Alright, let’s put it in a Fitbit. Let’s put it in a fridge. Why do you need AI in a fridge? AI went through a phase where it was everywhere—it can change the world, it can cure cancer, it can make food for you and do everything.
Now we’re thinking—alright—what are the real problems AI can solve? The contact center of the future I envision, will have AI from pre-interaction to post-interaction. We spoke about pre-interaction earlier, I.e., businesses mining data, and understanding that an order is delayed. When you call your e-commerce company, they should predict why you’re calling. Additionally, we will see advancements on the voice side used to understand why a consumer is calling.
From pre-calls to post-calls, we’ll see automation, speech understanding, natural language understanding, and conversational AI coming into the picture, to change the nature of the contact center. We are living in a world where you can connect with your contact center over chat, email, full-context of what you said last time on the chat. We are living in a world where there’s no distinction between channels that you are seeing, but every function from beginning to end has AI, so none of us have to wait, and none of us have a bad experience.
