Salesforce Signs Deal to Buy Voice AI Startup Cresta in Reported $2.7 Billion Contact Center Bet

Salesforce has spent much of 2026 rebuilding its customer service stack around artificial intelligence, moving from text based support bots toward systems that can handle the messy, high stakes reality of a live phone call. On 2 November 2026, the company took its most direct step into that territory yet, announcing a definitive agreement to acquire Cresta, a nine-year-old startup that describes itself as the voice-first operating system for customer experience. Salesforce did not disclose financial terms in its announcement.

The silence on price did not last long. The Information, citing a person familiar with the deal, reported the figure at $2.7 billion, a number that would represent a meaningful premium over Cresta's last public valuation of $1.6 billion in 2022. That valuation was set when Tiger Global led an $80 million round, according to CXM World. Neither Salesforce nor Cresta has confirmed the number, and the gap between an announced agreement and a confirmed price tag has become a familiar feature of large enterprise software deals.

Cresta's pitch is narrow but technically demanding. Its voice agents are built to resolve complex customer calls end to end, holding natural real-time conversations, handling interruptions, and coaching human representatives while analyzing every interaction on the same platform. Salesforce said the software already operates at enterprise scale for brands including Marriott, Comcast and Alaska Airlines, a customer list that spans hospitality, telecommunications and aviation, three industries where call volume is heavy and the cost of a bad interaction is high.

The acquisition lands in the middle of an unusually busy stretch for Salesforce, which has been assembling customer-facing businesses at a pace that rivals its largest acquisition sprees. It also arrives at a moment when the entire contact center software market is being reorganized around the question of whether AI agents can be trusted with the telephone, the oldest and least forgiving channel in customer service.

Key Facts

Reuters reported on November 2 that Salesforce announced the agreement, with the company describing Cresta as the voice-first operating system for customer experience. Salesforce did not disclose financial terms in that announcement, and the company said the transaction is expected to close in the first quarter of Salesforce's fiscal year 2028, subject to customary closing conditions including receipt of required regulatory approvals. That window runs roughly from February to April 2027, according to CXM World.

The reported price of $2.7 billion, first surfaced by The Information, compares with Cresta's last public valuation of $1.6 billion in 2022. CXM World put the multiple at roughly 1.7 times that mark. Dealroom's company record lists Cresta's total disclosed funding at $276 million, meaning the reported exit would return a substantial multiple on capital raised over nine years. Dealroom cautioned that the price is not confirmed by Salesforce.

Cresta's leadership has changed hands as the company matured. The company was born in 2017 at Stanford's AI Lab, co-founded by Sebastian Thrun, the Udacity co-founder and founder of Google X, together with Zayd Enam and Tim Shi, an early researcher at OpenAI. Ping Wu, a co-founder of Google CCAI, became chief executive in 2023. In a statement, Wu said the combination with Salesforce would let Cresta realize its full vision at scale, embedding real-time conversation intelligence and agentic experiences directly into Salesforce's first-class customer service platform.

Kishan Chetan, executive vice president and general manager of Agentforce and Service Cloud at Salesforce, framed the deal around the difficulty of voice. Voice is where service is hardest to get right, and where it matters most to customers, he said, adding that Cresta has built voice agents that handle complex, high-volume calls for some of the world's largest brands. Salesforce also committed that after closing, Cresta will continue to support the leading contact center and telephony platforms its customers rely on today.

The price tag sits inside a much larger spending pattern. Salesforce completed its roughly $3.6 billion Fin acquisition, formerly Intercom, in September 2026. It agreed to buy Braze for about $5.66 billion, and it has also moved on Rune Technologies, a defence software company, making Cresta its third deal in a fortnight according to CXM World. Cresta itself reached $100 million in annual recurring revenue in 2026, according to its own company history, and grew its team past 300 employees in 2025 with expansion into APAC and EMEA.

Analysis

What this really means is that Salesforce is no longer treating voice as a legacy channel to be defended but as the next frontier to be conquered, and it is willing to pay a premium multiple for a company that has spent nearly a decade building the hard engineering underneath natural sounding phone conversations. The reported 1.7x markup over Cresta's 2022 valuation is not a spectacular return by the standards of the 2021 software boom, which tells you something about how the market has repriced growth stage enterprise AI companies since then. Cresta raised a Sequoia-led Series B in 2021, a Tiger Global-led Series C in 2022 and a $125 million Series D in 2024. A $2.7 billion exit is a solid outcome, but it is not the kind of number that resets expectations for the category.

The strategic logic, though, is straightforward. Salesforce has assembled a portfolio of service assets that covers digital messaging through Fin, customer engagement through Braze, and now real-time voice through Cresta. Pairing Cresta's voice agents with Fin's service agents and the existing Service Cloud install base gives Salesforce a single story it can sell to large enterprises: one platform that handles a customer's question whether it arrives by chat, email or phone. The alternative, letting a specialist vendor own the voice layer while Salesforce owns the surrounding workflow, would leave a gap that competitors could exploit.

The competitive map makes that gap more dangerous than it looks. Constellation Research reports that the contact center platforms Cresta supports include Avaya, Five9, NICE, Cisco and Genesys, a list that reads like Salesforce's rivals in the service infrastructure market. Genesys and Five9 invested in Cresta in 2022, and in 2025 Salesforce and ServiceNow committed $1.5 billion between them to Genesys. That history of interlocking investments means the acquisition is not simply a purchase of technology but a repositioning within a web of partnerships, and Salesforce's promise to keep supporting third party platforms after closing is a recognition that Cresta's customers do not live exclusively inside the Salesforce universe.

There is also a credibility question about integration. Dealroom cautioned that stitching together multiple costly acquisitions carries real integration and execution risk, and the 2026 pace at Salesforce is aggressive by any measure: Fin at roughly $3.6 billion, Braze at about $5.66 billion, Rune Technologies, and now Cresta at a reported $2.7 billion. Each of those deals brings its own engineering culture, its own customer contracts and its own roadmap. The bigger picture here is that Salesforce is betting it can absorb all of them faster than its customers can notice the seams, at a moment when enterprises are actively renegotiating their service software contracts around AI capability.

Why It Matters

The contact center is one of the largest and most stubbornly expensive pieces of enterprise software, and voice is the portion of it that has resisted automation the longest. Chatbots have handled routine text queries for years, but phone calls involve accents, background noise, interruptions and emotional customers, all of which break naive systems. Cresta's claim is that its agents can hold a natural real-time conversation and resolve a complex call end to end rather than handing it off at the first sign of trouble. If that claim holds at enterprise scale, the labor economics of customer service change substantially.

The early evidence cited in the materials is suggestive rather than conclusive. Achieve reports that Cresta cut after-call work by 75 percent, a figure drawn from Cresta's own case study, which means it should be read as a vendor supplied benchmark rather than an independent audit. Still, the direction is consistent with what the broader market is promising: fewer minutes of human wrap-up time, faster resolution, and coaching that happens during the call rather than in a quarterly review. For buyers at Marriott, Comcast and Alaska Airlines, the question is whether those gains survive the integration into a much larger platform.

The deal also matters for the startup ecosystem. A $2.7 billion outcome for a company that raised $276 million in disclosed funding and reached $100 million in annual recurring revenue is a signal that enterprise AI companies with real customers and real revenue can still command premium exits, even after the valuation reset of recent years. Cresta's investor list, which includes Accenture, Andreessen Horowitz, Comcast Ventures, Five9, Genesys, Greylock, J.P. Morgan, LG Technology Ventures, Qatar Investment Authority, Qualcomm Ventures, Sequoia, Tiger Global, WiL and Workday Ventures, shows how deeply strategic capital had already permeated the voice AI category before this exit.

Next Up

The immediate test is regulatory. The transaction must clear required approvals before it can close between February and April 2027, and while voice AI has not attracted the same antitrust scrutiny as some other technology sectors, a deal of this size involving a company whose software runs on platforms owned by competitors will draw questions. Salesforce's commitment to continued support for existing contact center and telephony platforms will likely be part of those conversations.

After closing, the harder work begins: folding Cresta's real-time conversation intelligence into Service Cloud and Agentforce while keeping the third party integrations that customers depend on. If Salesforce manages that, the reported $2.7 billion will look like a reasonable price for the last major channel it did not fully own. If the integration stumbles, the deal becomes another data point in the argument that the fastest route to AI capability is not always the cheapest. Either way, the contact center market now has a clear signal that the voice layer is in play.