YesPress · Company File No. 042
UJET /yoo-jet/
The contact center built like a smartphone, sold like enterprise software, and now run on agentic AI. The company that decided "press 1 for English" was a design failure, not a feature.
The contact center built like a smartphone, sold like enterprise software, and now run on agentic AI. The company that decided "press 1 for English" was a design failure, not a feature.
A customer of a grocery delivery app taps a button inside their phone. No phone number. No "your call is important to us." A live agent picks up two seconds later, already looking at the order, the address, the previous complaint about the bruised avocados. The agent sees a photo - the customer took it ten seconds ago - of the dented can of beans in question. Refund issued. Call over. Forty-eight seconds, end-to-end.
This is the contact center as UJET imagines it. Boring, fast, and quietly working in the background of an app you already use. Most contact centers do not look like this. Most contact centers look like 1998 wearing a cloud sticker.
"A call from a smartphone should not feel like a call from a landline. We fixed that."— Anand Janefalkar, founder, paraphrased from his 2019 interview with Sramana Mitra
In 2015, the contact-center-as-a-service market existed. It was just unfortunate. Genesys ran on appliances. Avaya ran on appliances pretending to be cloud. NICE and Five9 had moved real workloads to the public cloud but kept the user experience of a dial-up bulletin board. The customer experience, supposedly the point, was an afterthought.
Anand Janefalkar - Motorola alum, Jawbone alum, the kind of engineer who had spent a decade making consumer hardware feel inevitable - looked at this and saw a category that had been built backwards. Companies were spending billions trying to make the contact center cheaper for the company. Nobody was making it better for the human being on the other end.
So he asked an annoying question. Why does a support interaction lose context the moment it leaves the app? Why can a customer send a photo to a stranger on Instagram in two seconds but not to a support agent in twenty minutes?
"The contact center is the only product in your business where customers arrive already angry. It's a strange thing to neglect."— Editorial paraphrase, drawn from UJET's public positioning
The bet was unfashionable at the time: build a CCaaS platform that lived inside the customer's app, on the customer's phone, with the agent's screen built like a modern productivity tool instead of a Windows 2000 utility. CRM-first - meaning Salesforce, Zendesk, and the rest of the customer-record-of-truth tools were the spine, not a bolted-on integration. Cloud-only - no on-prem fallback, no hybrid concessions, no "we'll meet you where you are."
It was a bet against the entire installed base of the industry. Which, to be fair, is the only kind of bet venture capital tends to reward. GV led in. Kleiner Perkins joined. Citi Ventures, DCM, and Ericsson Ventures came along. UJET raised a Series A in 2017 and never looked underfed again.
"Cloud-only sounds obvious in retrospect. In 2015 it sounded like product suicide."— Common refrain in early CCaaS sales calls
UJET sells software that runs a contact center. Voice, chat, SMS, in-app messaging, and the underrated stuff: photo, video, screen share, biometric authentication. The platform routes interactions to humans or to AI, depending on which one is actually useful, and it does so without making the customer repeat their order number four times.
Two things make it different from the seventeen other companies that will tell you the same thing. First, the architecture: native cloud, native mobile, CRM-first. Second, the Google Cloud relationship, which is closer to a joint product team than a partnership and which produced CCAI Platform - Google's own contact-center offering, sold under Google's brand, powered by UJET.
Customizable desktop with smart reply, automated summaries, and real-time sentiment - the boring-but-loved parts that make agent attrition go down.
Generative AI self-service. Handles tier-one. Escalates to a human with the full transcript and customer context attached.
Forecasting and scheduling built on Google Cloud. Co-developed; sold inside UJET and inside CCAI Platform.
"It's a strange thing to admit, but the most innovative feature in our contact center is that it remembers what you said in the last one."— Imagined product reviewer, captured in the wild
A scrapbook of moments, fewer than the all-hands deck would have it, more than the cynics expected.
Anand Janefalkar founds UJET in San Francisco. The original pitch deck centers on a phrase that sounds quaint now: "smartphone-era support."
Series A closes with GV and Kleiner Perkins. The bet on cloud-only suddenly looks like a category instead of a quirk.
Strategic partnership with Google Cloud announced. UJET joins the co-sell program. CCAI Platform begins life.
Google launches CCAI Platform - powered by UJET - as its first-party contact center offering.
Wins Google Cloud Technology Partner of the Year for Productivity & Collaboration.
$76M Series D closes. Sapphire Ventures leads; existing investors all follow on. Vasili Triant elevated to Co-CEO.
Triant becomes sole CEO. Janefalkar transitions to founder/board.
Onix acquires UJET's professional services unit to scale Google Agentspace delivery.
Google Cloud CCaaS managed service offering launches via Google Cloud Marketplace.
In a category where everyone claims AI superiority and most demos collapse on the second prompt, UJET's evidence is unusually concrete: customers who renew, analysts who score it well, and a Google Cloud relationship that has survived two CEOs and four hype cycles.
Fig. 2 — Numbers don't lie. They do, however, occasionally hint at things their owners would rather you not notice. The 20-quarter G2 streak, for instance.
"100% of customer service agents interact with AI every day. 0% consider it critical to their daily success."— UJET research, 2025. The kind of stat that's only flattering if you sell the cure.
The customer roster has the right shape for a B2B SaaS company between Series C and IPO: Instacart for consumer scale, Turo and Wag! for digitally-native operations, Atom Tickets for entertainment, Nest for the home. Not the Fortune 50 yet. Not trying to be, yet.
UJET's official mission language uses the words "predictive, contextual, conversational." Strip the consultant veneer and the company is doing something simpler. It is trying to make support feel like a private message between two humans, with the AI doing the parts neither human enjoys - the transcription, the lookup, the next-best-action guess, the after-call summary.
The CEO transition makes this explicit. Vasili Triant is not the engineer founder. He is the go-to-market operator - Cisco, 8x8, three years as UJET's COO - who runs the company as a sales machine pointed at the enterprise. Janefalkar designed the product. Triant is selling it to the people whose budgets actually move.
"The best contact center is the one the customer never notices is there."— Industry truism, never quite achieved, occasionally approached
Agentic AI is going to eat tier-one support. This is uncontroversial inside the industry and faintly terrifying to the people doing tier-one support. The interesting question is what happens to tier two and tier three - the calls AI cannot close, the calls where the customer is angry and the situation is novel and the policy is ambiguous.
UJET's bet on the next decade is that those calls become more important, not less. When AI handles the easy ones, the hard ones get harder, the agents get fewer, and the tools they use have to be better. That is the product UJET has been building since 2015, dressed up now in agentic clothing. The architecture choice that looked unfashionable then - mobile-first, CRM-first, cloud-only - is the architecture you would, in fact, choose now.
Back to the avocados. The customer in the opening paragraph - the one who got a forty-eight second resolution - did not know any of this. She did not know UJET existed. She did not know an LLM had read her order history before the agent picked up. She did not know the agent's screen had three pre-written reply suggestions waiting. She knew only that her can of beans was being refunded and her morning was no longer ruined.
UJET would consider that a complete sentence.
"You can tell a company is good at customer experience when nobody talks about the customer experience."— The Oscar Wilde Theorem of B2B Software
If this file was useful, send it onward. The internet has too many opinion pieces and not enough company profiles.