Customer service is having its Google moment
The new economics of customer service are flipping the old rules of user engagement on their head. AI agents are rapidly evolving from agents of deflection to agents of resolution. And potentially on to something bigger…
In the mid 90s, web portals like Yahoo!, Excite, and Lycos were everywhere. But these so-called doorways to the web often felt more like locked rooms. They were optimized for stickiness: keeping you on their site as long as possible so that you’d view as many of their banner ads as possible.
Even though search was their core feature, web portals were more focused on promoting content and advertising, keeping users in a limbo of news, weather, horoscopes—anything for the eyeballs.
Then along came Google.
They flipped the portal model completely on its head.
Google’s founding insight was that they should get users off their site as quickly as possible. They saw that speed and forward motion to your destination were key, and those principles showed up in the product choices they made: a single search box, fast loading results, zero clutter.
Because these principles were better aligned with what users actually wanted when searching, Google won. And this gave rise to entire industries and remade the web into what we know today.
New products often flip the script in this way, prompting businesses to fundamentally rethink what they are optimizing for. Today, this is happening with AI, and customer service is a prime example.
Engagement versus deflection
Imagine you walk into a bricks-and-mortar store, and the sales rep’s goal is to block you from leaving. This is what portals were back in the 90s.
Now imagine another store, and the customer service rep there tries their hardest to completely avoid answering your question. This is what customer service on the internet was like for the last few decades.
Both cases were driven by economies of scale: of banner advertising and the cost to manually serve a support request, respectively. Show me the incentive and I will show you the outcome.
And so we’ve had years of do-not-reply email addresses, black hole feedback forms, canned auto-replies, and various other user-hostile anti-patterns.
But along came AI Agents to flip the script again—our own Fin first among them. Now customer service is having its Google moment.
Rather than optimizing for deflections (a.k.a. dodging the customer’s question), AI Agents allow us to think about this completely differently. A human-resolved support query costs an average of $15. Fin costs 99c per resolution. At that cost, most businesses can serve every single customer’s question.
And it doesn’t matter how much back and forth is required. Fin will stick with the problem and keep helping until the problem has been thoroughly solved.
That means that the unit economics of AI Agents are far better aligned with what users actually want: impeccable, patient, thorough customer service.
Where to next?
Switcheroos like this don’t just stop with the first innovation. Google has changed a lot since those early days, and broadened out to cover many adjacent types of search. A dramatic moment of reframing can provide rich new product paths for years to come.
It’s exciting to think of how customer service agents might go from here. Sure, it’s great that we now no longer need to keep customers at arm’s length thanks to agents like Fin. But think of what the world’s best customer service experience might look like. Would it stop there?
You might have noticed that ChatGPT and Claude responses have started to subtly evolve in recent months, now almost always ending their answers by prompting a helpful next suggestion. They’re not trying to shut the conversation down, they’re trying to keep it going.
Fin can also sometimes follow up in more open-ended ways, at times checking to see if there are any other things that it can help with.
This product choice may seem like a minor quality of life improvement at first. But it is in fact a broadening out of the entire concept of customer service—from a ticket resolution machine to something much more holistic and customer-focused.
AI is changing the actual role of customer service along with the economics. CS teams of old were sequestered away from adjacent areas like customer success, sales, and marketing—not because it was good for customers (it wasn’t) but because it was efficient to organize humans into specialized teams.
It doesn’t have to be like that anymore.
We’re watching the first blurring of the lines between customer-facing roles, and a reframing of how we should approach the very idea of serving customers.
Every once in a while, technology shifts present us with a Google moment, where it’s possible to spot the direction of travel and capitalize on it—or to not, and go the way of the web portal.
We’ll be exploring these ideas and more at Pioneer, our annual summit, on October 9th — in person in NY and streamed live online. We’d love to see you there!