BREAKING   Regal raises $40M to put AI phone agents inside enterprise contact centers Zendesk holds ~88% of the IVR software market, per category trackers TCN has run cloud call-center tech since 1999 from St. George, Utah Puzzel builds on ~20 years of European cloud contact-center heritage in Oslo UJET went smartphone-native years before AI voice was fashionable Bright Pattern's founder built Genesys before rebuilding the category
The CX Report · Enterprise Software

Contact Center Software Is Learning to Answer the Phone Itself

Zendesk, TCN, Puzzel, Regal, UJET and Bright Pattern are chasing the same bet - that the next voice on your support line won't be human. Here is how six very different platforms are racing to get there.

Retro pen illustration of a contact-center headset with AI soundwaves
The headset is still the symbol of the industry. The voice inside it is the thing being replaced. — Illustration: YesPress Newsroom

Call a company in 2026 to argue about a bill, confirm an appointment, or ask where your package went, and there is a real chance the calm voice answering was never hired, never trained, and never took a lunch break. It listened, understood the accent, checked your account, and decided - in the length of a breath - whether to solve the problem or pass you to a person. The call center did not disappear. It grew a new voice, and six companies are fighting over whose voice it becomes.

The category has an ugly name - CCaaS, for Contact Center as a Service - and a reputation to match. This is the software behind “your call is important to us,” the hold music, the menu tree that never has the option you want. Nobody grows up dreaming of building it. And yet it is one of the most interesting places to watch artificial intelligence actually land, because a contact center is where a demo meets a customer who just wants a refund and does not care how clever the model is.

Six names keep coming up: Zendesk, TCN, Puzzel, Regal, UJET and Bright Pattern. They are not the same kind of company. They started in different decades, on different continents, solving different problems. What unites them now is a single question, and their answers are more different than the marketing suggests.

1999
Year TCN began running hosted call-center tech - a 25-year head start on the AI natives
$40M
Raised by Regal in 2024 to bring AI phone agents to enterprise brands
~88%
Zendesk's share of the IVR software market, per category trackers

01 / THE QUESTIONWhen the software can talk, what is a call center for?

For twenty years the pitch in this industry was “omnichannel” - put voice, email, chat and messaging in one screen so an agent could follow a customer across all of them. It was a real improvement and a boring one. The channel everyone assumed was finished, plain voice, is the one that got interesting again, because a language model can now hold a phone conversation well enough that most callers do not immediately hang up.

That changes the product. The old contact center was built to route calls to humans efficiently. The new one is built to decide which calls a human should ever see. Every vendor here is drawing that line in a different place, and where they draw it tells you what they actually believe.

The question is no longer whether AI can answer the phone. It already does. The question is when it should stop and get a person. The central design problem of the 2026 contact center

02 / THE MAPSix bets, plotted

One useful way to see the field is on two axes. Left to right: is the platform built mostly for inbound support, where a customer calls you, or outbound engagement, where the software reaches out first? Bottom to top: how AI-led is the core product today, versus how much it still assumes a human at the desk? Nobody sits in a corner alone, but the clusters are real.

The CCaaS Positioning Map

Where each platform sits by primary use case and how AI-native the core product is. Approximate, based on public positioning.

INBOUND SUPPORT OUTBOUND ENGAGEMENT HUMAN-LED AI-LED Zendesk Bright Pattern Puzzel UJET TCN Regal
Regal sits alone in the outbound, AI-led corner. Zendesk anchors inbound support. Most others cluster in the pragmatic middle.

03 / THE INCUMBENTZendesk owns the ticket. That is the point.

Zendesk started in a Copenhagen loft in 2007 and became something rare in software: a verb. People say “file a Zendesk” the way they say “send a Slack.” Its power is not the phone call at all - it is the ticket, the record of a customer's problem that follows them across email, chat and voice. By some category trackers it holds close to nine-tenths of the IVR software market, which is less about telephony brilliance and more about the fact that everyone already lives in its workflow.

That is the incumbent's advantage and its trap. When the AI wave arrived, Zendesk did not need to invent a new customer base. It needed to bolt automation onto a system millions of agents already open every morning. The risk is that owning the workflow makes you defend the workflow, while a challenger with nothing to protect goes after the one moment you were quietly ignoring: the live phone call.

04 / THE VETERANSTCN, Puzzel and Bright Pattern - the deep bench

Three of these companies were building contact centers before “AI agent” meant anything.

TCN, founded in 1999 and run from St. George, Utah - about as far from the marketing gravity of San Francisco as an American software company gets - has spent a quarter century on the unglamorous, high-volume end: outbound dialing, compliance, collections, the calls that have to go out by the million and cannot break a regulation while doing it. That is exactly the terrain where a reliable AI voice is worth the most and the least forgiving of a bad one.

Puzzel, headquartered in Oslo and formerly called Intelecom, carries roughly twenty years of European cloud-telephony heritage. It is the omnichannel option for organizations that care about data residency, local languages and the particular way European support desks work. It rarely makes noise in the US press, which is part of why it keeps winning contracts the US press never notices.

Bright Pattern is the connoisseur's pick. Founded in 2010 by Konstantin Kishinsky, who held senior roles at Genesys - the legacy giant this whole category is trying to unseat - it is built for teams that want to configure everything and are willing to do the work. Reviewers score it highly for omnichannel depth. It is the platform for people who know exactly what they want a contact center to do.

Bright Pattern's founder built Genesys first. The best software categories are full of insiders who left to rebuild the thing they knew was dated. On why pedigree hides in unglamorous places

05 / THE MOBILE ONEUJET bet on the phone in your pocket

UJET, founded in 2015 by Anand Janefalkar - an alumnus of Jawbone and Motorola - made a contrarian call early: the important device is not the agent's desk phone, it is the smartphone in the customer's hand. So UJET built support around what a modern phone can do, folding in the camera, location and biometric sign-in so a customer can show a problem instead of describing it. That instinct aged well. When the whole industry pivoted to smart, context-rich interactions, UJET had been designing for the pocket for a decade.

06 / THE INSURGENTRegal, and the scar tissue behind it

Regal is the youngest name here and the one that reset the conversation. Founded in New York in 2020 by Alex Levin and Rebecca Greene, it did not try to out-Zendesk Zendesk. It went straight for the phone call - specifically outbound and event-driven ones: the lead worth calling back, the appointment reminder, the payment that needs collecting, the customer worth re-engaging before they drift.

The detail that makes Regal credible is not the funding, though the $40M it raised in 2024 for AI phone agents got attention. It is the origin. Before building the software, co-founder Rebecca Greene ran a 3,000-person contact center at Angi that handled close to 100 million customer interactions a year. Most voice-AI startups have never had to staff a Monday morning queue or explain to a regulator why a call went sideways. Regal's founders lived inside the mess they are now automating.

Who Answers First

A rough read on each platform's founding era and its center of gravity in the AI shift.

PlatformFoundedHome baseCore strength
TCN1999St. George, UtahHigh-volume outbound & compliance
Zendesk2007San FranciscoTicketing & support workflow
Bright Pattern2010South San FranciscoConfigurable omnichannel
UJET2015San FranciscoMobile-first, smartphone-native
Puzzel~2000s (Intelecom)Oslo, NorwayEuropean omnichannel & data residency
Regal2020New YorkAI-native outbound voice agents
Founding years and locations from public company sources. Strengths reflect stated positioning, not a ranking.

07 / THE REAL FIGHTThe handoff is the product

Here is the thing the demos hide. Making an AI answer the phone is now the easy part. The hard part, the part that decides who wins, is the handoff - the instant the software realizes it is out of its depth and passes a frustrated human to a human. Do it too late and the customer is already furious. Do it too early and you have just paid for a robot that solves nothing.

Every platform in this story is really competing on that seam. Zendesk can lean on the context in its tickets to make the handoff seamless. Regal can lean on knowing exactly why it placed the call in the first place. TCN can lean on 25 years of knowing which calls should never have been automated. The winners will not be the ones with the smoothest AI voice. They will be the ones that know, reliably, when to shut it off.

AI did not kill the call center. It split it in two - the part a machine handles, and the part it hands to a person. Everyone here is drawing that line differently. The 2026 CCaaS thesis, in one sentence

08 / WHAT TO TAKEHow to actually choose

If you run a support team and have to pick, ignore the AI marketing and start with the job. If your world is support tickets and you want the broadest ecosystem, Zendesk is the safe center. If you dial out at scale under heavy compliance, look at TCN or Regal. If you want AI-first outbound voice specifically, Regal is the sharpest instrument. If you operate in Europe and care about data residency, Puzzel. If your customers live on their phones, UJET. If you want to configure every last detail yourself, Bright Pattern.

The larger lesson has nothing to do with call centers. It is a clean example of how a category matures under AI: the incumbent that owns the workflow and the insurgent that owns the moment can both look right at the same time, for a while. Zendesk owns where the work happens. Regal grabbed the one interaction Zendesk was not built around. Watching which one moves onto the other's turf first is the whole game - and it is playing out in the most boring software you will ever be grateful someone else has to maintain.