Pick up a phone at almost any company and something quietly disappears: the actual words. The deal that slipped, the promise a rep made, the reason a customer was furious - all of it evaporates the second the call ends. Dialpad's whole business is the refusal to let that happen. Since 2011, the San Ramon company has been building phone software that listens, transcribes, and, increasingly, talks back.
Today Dialpad is an AI-powered communications platform that folds four products - business phone, video meetings, team messaging, and a cloud contact center - into a single system. What ties them together is not the calling. Plenty of vendors do calling. It is the layer underneath: Dialpad's own speech-recognition and natural-language engine, which transcribes conversations in real time, writes the summary the moment you hang up, tracks whether the customer is getting angrier, and feeds live prompts to whoever is on the line.
01 / THE FOUNDERSA serial founder who keeps rebuilding the phone
To understand Dialpad you have to understand its CEO, Craig Walker, who has now built the same thing three times. His first internet-phone company, Dialpad Communications, was bought by Yahoo. His next, GrandCentral, was bought by Google in 2007 and became Google Voice. Walker stuck around as an entrepreneur-in-residence at Google Ventures, then left to do the obvious thing: start a third telephony company - this time keeping the data and building the intelligence himself.
He co-founded it in 2011 with Brian Peterson, a fellow Google Voice engineer who became CTO. The company launched as Switch Communications and shipped a quirky web conferencing tool called UberConference - the one that riffed on callers by name in its hold music. In 2016 the whole thing was rebranded Dialpad, and the AI ambitions moved to the center of the story.
02 / THE MOATOwning the data instead of renting the intelligence
The pivotal move came in 2017, when Dialpad acquired a speech-analytics startup called TalkIQ. Most software companies in 2026 rent their AI - they pipe conversations to an outside model and pay per token. Dialpad went the other way early. It bought a speech engine and trained it on the one thing it had that competitors did not: years of real business conversations flowing across its own network.
That is the quiet advantage behind the flashy features. When Dialpad transcribes a support call, detects that a customer is about to escalate, or slides a coaching card onto a sales rep's screen mid-objection, it is running models tuned on the specific, messy vocabulary of business calls - product names, account numbers, half-sentences, hold music. The intelligence is native to the call, not bolted on after it.
03 / THE PRODUCTSFour tools, one bill
Dialpad sells its platform in overlapping pieces. Ai Voice is the business phone system. Ai Contact Center (branded Support) handles customer service across voice, video, live chat, and social messaging, with real-time agent coaching and automation. Ai Sales (Sell) is the outbound dialer with live coaching cards for revenue teams. Dialpad Meetings - UberConference's second life - covers video. And Dialpad Connect knits phone, chat, and video into one internal-comms product.
The pitch is less about any single feature than about the bundling. A mid-sized company juggling a phone vendor, a video vendor, a chat tool, and a separate contact-center platform can, in theory, collapse all four into one login and one invoice. That consolidation argument - one platform instead of four contracts - is a large part of why the company grew.
Reported customer outcomes with Dialpad AI
04 / THE CUSTOMERSFrom the Padres to Motorola
More than 30,000 businesses run on Dialpad, spanning small teams paying per seat and enterprises signing large contracts. The named roster reaches across industries - T-Mobile, Motorola, RE/MAX, Equifax, and sports franchises like the San Diego Padres and Florida Panthers. T-Mobile is an unusual case: it is both a customer and, through T-Mobile Ventures, an investor.
The users who feel the product most are contact-center agents and sales reps - people who live on the phone and who, before AI note-taking, spent a chunk of every call typing instead of listening. Dialpad's own 2025 survey of agents found 52% already leaning on AI-written call summaries to lift customer satisfaction scores.
05 / THE MARKETThe underdog against the giants
Dialpad plays in two crowded acronyms: UCaaS (unified communications) and CCaaS (contact center). The incumbents are bigger - RingCentral and Zoom on the communications side, Five9 and Genesys in the contact center, with 8x8, Nextiva, Aircall, and Vonage crowding the middle. Against that field Dialpad has generally competed on being AI-native rather than AI-retrofitted, and on price at the entry tiers.
The analysts have started to notice. In 2025, Gartner named Dialpad a Visionary in its UCaaS Magic Quadrant and handed it an Honorable Mention in a newly created category within the CCaaS Magic Quadrant. For a company far smaller than its rivals, out-shipping them on AI features is the whole strategy.
06 / THE MONEYHow the funding stacked up
Dialpad raised steadily rather than spectacularly. The pandemic-era Series E, a $100M round in 2020, pushed it past a $1.2 billion valuation as remote work sent demand for cloud calling vertical. Then in December 2021 came the $170M Series F, led by ICONIQ Capital, that nearly doubled the number to $2.2 billion. The business is reported to have crossed $200M in annual recurring revenue.
07 / WHAT'S NEXTAnswering before you ask
In 2025 Dialpad laid out its next bet: "pre-emptive customer service." The idea is an agentic layer of autonomous AI that anticipates why a customer is reaching out, handles routine tasks on its own, and hands off cleanly to a human when the situation needs one. It is ambitious, a little unnerving, and a direct move into the territory Five9 and Genesys have long owned.
Whether autonomous agents live up to the demo is the open question hanging over every company in this market right now. But Dialpad is walking in with something most rivals cannot buy off the shelf: a decade and a half of business conversations it already recorded, transcribed, and owns. In an AI market where data is the constraint, that is not nothing.