Dialpad and Ada announced on November 4, 2026 that they are joining forces, with the transaction expected to close on or about November 12. Dialpad sells an AI platform for customer experience on its own voice network. Ada sells customer service agents, with more than 500 in production, it says. The combined company is meant to build an agentic operating system for customer work, and the announcement does not put a price on it.
What the announcement fixes
- The deal is expected to close on or about November 12, 2026.
- Craig Walker, Dialpad’s co-founder and CEO, leads the combined company.
- Mike Murchison moves from Ada CEO to Chief AI Officer.
- Current customers keep their contracts, pricing, account teams and support channels.
- Financial terms were not disclosed.
What was actually released
What shipped on November 4 is a statement of intent, not a product. Nothing new became available, no migration path is described, and both product lines keep running as they are. The document is dated San Francisco, and its most concrete commitments are the ones a customer should read first: current Dialpad and Ada customers keep their contracts, pricing, account teams and support channels, and existing integrations continue to operate as configured.
Leadership is settled in the document. Craig Walker stays on as CEO of the combined company, Mike Murchison becomes Chief AI Officer, and Mike Gozzo reports to him as Senior Vice President of AI Product and Engineering. Both plan to extend the model past customer service into sales, without naming a date or a product.
The release lists three outcomes the combined platform should deliver: every customer treated as the most important one, people and agents working as one team, and each interaction feeding what the business learns. All three are ambitions, and none carries a number.
A second block is scale. Ada cites more than 500 agents in production across voice, messaging, email and social, in more than 130 languages. Dialpad says it has analyzed more than 14 billion minutes of business conversations since 2018. Together the companies say they serve more than 50,000 businesses, and they name some: Randstad, Motorola Solutions, Netflix, Square, Pinterest, Sky, Canva and Monday.com.
| What the release states | Figure |
|---|---|
| Deal close | on or about November 12, 2026 |
| Deal price | not disclosed |
| Dialpad business conversations analyzed since 2018 | more than 14 billion minutes |
| Ada agents in production | more than 500 |
| Languages Ada covers | more than 130 |
| Companies served by the two together | more than 50,000 |
| CEO of the combined company | Craig Walker |
Owning the voice network and its own speech models, the release says, gives the combined company more control over core costs and performance.
What the claim rests on
The release rests on two kinds of evidence. The first is production scale: minutes analyzed, agents deployed, languages covered. Those numbers can be checked, but only with the companies, and the release gives no audit and no definition of what counts as an agent in production.
The second is endorsement. Three quotes carry the argument, from the two CEOs and from two investors, Marc Andreessen at Andreessen Horowitz and Brian Feinstein at Bessemer. An investor’s view of a merger is a position, not a measurement: both firms back the companies involved.
Then there is the phrase it is built on. An agentic operating system for customer work is not defined: no architecture, no data model, no API, no date, no product name. The claim that matters to a buyer is the difference between answering a customer and resolving the request, and the release argues it with an analogy about a good employee rather than a resolution rate.
On the price, the document is explicit: financial terms were not disclosed. That line is more useful than a number: it says the announcement was written to land a strategic story, not a valuation. Any figure you see attached to this deal in coverage is a report from outside both companies. We do not have it from Dialpad or Ada, so we do not repeat it as fact.
What the document does not say
- No purchase price, no valuation and no revenue multiple.
- No integration roadmap, no combined product name and no ship date.
- No statement on headcount, retention or which engineering teams stay.
- No customer evidence with measured resolution or deflection rates.
- No detail on closing conditions, or on what happens if the close slips past November 12.
On this page
What it changes for teams buying a customer service agent
Very little changes before November 12, which is the point of the contract language. If you run Ada or Dialpad today, the release commits to keeping your contract, pricing, account team, support channel and integrations as configured. A renewal in the next few weeks has a written baseline.
What does change is the number of vendors on a shortlist. Telephony, contact center and customer agents were three buying decisions; this turns two into one supplier with its own voice network and speech models. Fewer contracts, and one less alternative to play against the other at renewal.
Procurement should ask three things in writing before the roadmap lands: which product line your workflows sit on, what happens to the API you built against it, and what additional roadmap details in the coming months means for your contract term.
For what an agent costs, ElevenLabs prices all paid voice plans at $0.08 a minute shows a published rate card; this release gives none for the combined platform. On what buyers accept, our retail evidence is closer to the ground: what shoppers choose after trying a retail AI agent.
The Neural Minute take
The sentence worth keeping is the one about the transfer queue: it states the thesis better than the phrase agentic operating system does.
They don’t drop you into a transfer queue. They bring in the right colleague, make the introduction, and stay on the line to help.
That is a problem of orchestration rather than model quality, which is why two vendors holding different halves of the stack would merge. What is missing is the price, and its absence is the story: a merger announced without a number is making a strategic argument, and the claim is that one platform beats two best of breed tools. The cost of that argument is optionality.
Our reservation is the one we apply to every platform announcement: the release asserts resolution and shows a count. More than 500 agents in production and 14 billion minutes analyzed are inputs. Neither is an outcome, and neither tells you what share of requests the combined stack closes without a human. Ask for the resolution rate, not the agent count.
Two positions follow. Treat the months between the close and the roadmap as a freeze: do not sign a multi-year expansion on the promise of a product that has no name. And keep a second vendor qualified even if you do not deploy it, because the shortlist for a voice-first agent is shorter than it was a week ago.
Sources
- Dialpad and Ada combining to build the agentic operating system for customer work, Dialpad press release, November 4, 2026: https://www.dialpad.com/press/dialpad-and-ada/
- The page Dialpad points readers to for the combined company: https://www.dialpad.com/ambitious/
- Figures and commitments are read in that release on November 9, 2026: closing date, leadership, contract continuity, 14 billion minutes, 500 agents, 130 languages, 50,000 companies, and no financial terms disclosed.