Dialpad AI Agents Alternative UK: What the Ada Deal Changes

Your phones already run on Dialpad, Dialpad now sells AI agents that answer those phones, and on 4 November 2026 Dialpad announced it is joining forces with Ada, the AI agent company. So the question on your desk is no longer only whether Dialpad's AI agents are good enough. It is whether to commit your phone work to a product whose roadmap the companies say will be set out in the coming months, or to run that work on a separate voice AI platform beside the phone system you keep.

This guide is written for the Head of Customer Operations or COO whose business phone system is Dialpad. It reads Dialpad's own pages and help articles as they stood between 9 and 11 November 2026 and turns them into the checks that matter: what the product is now called, what the combination does and does not promise, how an agent reaches a UK phone line, which design details need an owner, and where your call data sits. It does not quote any Dialpad price, and our 2026 buyer's guide carries the cost comparison across vendors.

The boundaries are deliberate. Contact-centre suites adding AI agents are covered in our Genesys route map, our Five9 alternative guide and our Amazon Connect build guide. This post is about the other shape of buyer: a business phone system first, with a contact centre and AI agents added on top, and a vendor in the middle of a combination. The generic build, orchestrate or buy decision has its own guide on this blog.

This guide is shipped by the team behind Dilr Voice, a multi-agent voice AI platform that chains specialised agents into a single phone call. Or see DATS, our five-stage AI consulting system, if the open question is where AI belongs in your operation at all.

What is Dialpad Resolve, and is it the same as Dialpad AI Agents?

Dialpad Resolve is the current name for Dialpad's agentic AI offering, which Dialpad previously called Dialpad AI Agents. Dialpad describes it as AI agents for voice and digital interactions that answer questions, take configured actions, trigger workflows and escalate to a person with context. It sits beside Dialpad Connect for business calls, Dialpad Support for contact centres and Dialpad Sell for outbound sales.

The rename matters when you search, because Dialpad's help centre article on agentic agents still says AI Agents throughout, while Dialpad's own product summary says Resolve and lists its pricing as contact sales. Treat them as one product. This guide uses Dialpad's AI agents for the thing you configure in the admin settings and Resolve for the commercial product.

The help centre also splits the agents into two types, and the distinction decides what you are buying. A Generative (Classic) agent answers from your knowledge base. An Agentic agent can also perform tasks and run workflows through connected systems. Only the agentic type is relevant to a phone line that has to book, update or look something up, so everything below concerns agentic agents on voice unless a sentence says otherwise.

The product ladder is the other thing to hold in mind. The UK Resolve page says every conversation in Connect builds the customer intelligence that powers Support, Sell and Resolve. That is a deliberate platform story: the agents are designed to be most useful when your calling, your contact centre and your AI all sit with one vendor. For some teams that is exactly the attraction, and we come back to when it wins.

What changes for Dialpad customers when Dialpad and Ada combine?

Dialpad and Ada announced on 4 November 2026 that they are joining forces, with Dialpad's chief executive Craig Walker leading the combined company and Ada's chief executive Mike Murchison becoming Chief AI Officer. The companies said the deal was expected to close on or about 12 November 2026. Existing customers keep their contracts and integrations for now, and roadmap details are promised in the coming months.

The continuity commitment is the sentence most Dialpad customers will want in their file. In the joint announcement, the companies wrote:

"Current Dialpad and Ada customers will continue with their current contracts, pricing, account teams, and support channels, and existing integrations will continue operating as configured."

Dialpad and Ada, joint press release, 4 November 2026

Read that as what it is. It is a statement about current contracts and current configurations. It says nothing about which agent builder survives, whether the two agent products converge, or what a renewal looks like once the roadmap is published. The same release says the companies will share additional roadmap details in the coming months, and that financial terms were not disclosed.

The release also tells you where the combined company thinks its strength lies. Dialpad says it has analysed more than 14 billion minutes of business conversations since introducing real-time AI in 2018, and that Ada has deployed more than 500 AI agents in production across voice, messaging, email and social channels. Both figures are the companies' own. Together they describe a voice network and proprietary speech and language models on one side and an agent product on the other, which is a sensible combination, and also a sign that the agent layer you configure today may not be the agent layer you run next year.

None of this is a reason to leave Dialpad. It is a reason to sequence the decision. If your AI agent build is about to start, the useful questions for your account team are about the next twelve months, not the last twelve:

Our guide to voice AI contract clauses sets out the wording that protects you through a change of ownership, whichever vendor you choose.

How does a Dialpad AI agent get onto a UK phone line?

A Dialpad Agentic AI Agent is created in the admin settings by a company or office admin, given knowledge sources and skills, validated and published, tested in a text preview, and then attached to a voice target through call routing. Dialpad says the agent can sit on a department, the main line or a contact centre, which suits a business phone estate more than most contact-centre suites do.

The sequence below condenses the Dialpad help article as we read it on 9 November 2026. It is shorter than the equivalent AWS assembly in our Amazon Connect guide, because the agent and the call routing that reaches it are both configured in Dialpad's own admin settings.

The skill model is where the real work sits. Dialpad says a skill grants the agent access to specific actions, which run through connectors, and each API endpoint inside a connector is called a data operation. Each skill carries its own guardrails: what the agent must never do, what data it must not touch, when it must ask for confirmation, and when it should escalate. The connectors Dialpad names include Salesforce, ServiceNow, Zendesk, HubSpot, Google Calendar and Workspace, Outlook, Calendly and Zoho CRM. Anyone who has designed tool calling for a voice agent will recognise the shape.

Two details deserve a line in your plan. Validation can end in "Publish with warnings", which is sensible for iteration and risky as a habit, so decide who is allowed to use it. And the Test feature in the article is a text preview: you enter a question and check the reply. A phone agent should also be tested on the phone, with the accents, background noise and interruptions your callers bring, before real traffic reaches it. A golden-set regression test is the natural gate for every later prompt or skill change.

Which Dialpad details should shape a UK voice design?

Four details from Dialpad's own help article should shape a UK voice design: agentic agents are available in English with other languages coming soon, connected knowledge syncs every 24 hours while uploads never sync, long behaviour prompts make the agent monologue on calls, and escalation fires when a caller asks for a person. None is a defect, but each needs an owner.

Language is the first test for many UK teams. Dialpad's help centre says agentic agents are available in English, with other languages coming soon. If a share of your callers prefer Welsh, Polish, Punjabi or Urdu, that is a gap today, and it is worth asking whether the deal changes it, because the release's figure of more than 130 languages is Ada's own and describes Ada's product across channels, not Dialpad's voice agents.

Knowledge freshness is the second. Connected sources such as Zendesk, Document360, Box, Guru and Google Drive sync every 24 hours, and uploaded documents do not sync at all: you remove the old file and upload the new one, up to 5 MB each. A price change, an outage notice or a policy update made at nine in the morning can therefore be wrong on the phone until the next sync. Our voice knowledge base design guide covers how to keep time-sensitive answers out of a slow-syncing store.

Voice behaviour is the third, and Dialpad is admirably candid about it. Its help article answers a question about agents that monologue on voice, saying long, dense behaviour descriptions produce essay-style answers that are especially painful on a call. The fix it gives is short, step-by-step behaviour prompts, with detail held in skills and knowledge. The same article explains that some instructions, such as asking the agent to wait a few seconds before greeting, are ignored because the model has no real timing control. The article adds that agents end a voice session after 90 seconds of caller inactivity. Plan your pacing work, including how silence is handled while a tool runs, with dead-air masking patterns in mind.

Escalation is the fourth. Dialpad says that once an agentic agent is connected, a caller who says they want to speak to an agent is escalated into your contact centre automatically. That covers the explicit request. The cases that matter more in the UK are the implicit ones: a caller who sounds vulnerable, a complaint, a failed identity check. Each needs a guardrail in a skill and a test call, and Dialpad's own write-up of agentic AI from pilot to production lists the context a human receives on handover, including reason for contact, authentication status and sentiment. Our warm transfer design guide sets out the cases to test.

The same operating logic underpins our AI operating model consulting, which assigns a named owner to every prompt, skill and escalation rule before an agent takes a live call.

Where does Dialpad store call data for a UK customer?

Dialpad's own pages give two versions of the answer. Its UK trust page says all permanent customer data, including call records, recordings and transcripts, is stored in the United States on Google Cloud Platform. Its privacy policy also names the United States as the default, unless your Dialpad administrator has selected a different region. A UK buyer should get the region for their account confirmed in writing before the data protection review closes.

Both statements were live when we read them on 11 November 2026. The UK trust page answers the question where and how is my data stored with the United States, backed up daily, and adds that transient data passing through Dialpad's data centres is held for no more than 72 hours. The Dialpad privacy policy repeats the United States default and adds the administrator-selected region as an exception, and also says Dialpad may temporarily process personal data in other regions for support and sales. The difference is probably one of emphasis rather than conflict, but it is your controller's job to know which applies, and to which data types, including recordings and transcripts produced by AI agents.

If the answer is the United States, the transfer route is a recognised one. Dialpad's privacy policy says it has certified to the U.S. Department of Commerce under the UK Extension to the EU-U.S. Data Privacy Framework, and GOV.UK records that the UK adequacy regulations for that extension came into effect on 12 October 2023. Check that the certification is current when you sign. Even then, the route does not finish your work: your records of processing, your privacy notice to callers and your retention settings still need to reflect it. Our guide to international transfers sets out the review, and the ICO's expectations for voice agents are covered in our UK and EU compliance guide.

One more line in Dialpad's terms belongs in the same review. The Agentic Services section of Dialpad's legal terms says the services may use third-party large language models and infrastructure that Dialpad does not control or develop. That is a disclosure, not a criticism. It does mean your sub-processor review should ask which models an agentic voice agent uses, where they run, and whether that changes after the Ada combination.

Security credentials are a strength Dialpad documents publicly, and fairness requires saying so. Its UK trust page states SOC 2 Type II compliance, ISO 27001:2022, ISO 27017 and ISO 27018 certification, and HIPAA readiness with a signed business associate agreement, all on Dialpad's own published account. If certification is a gate in your procurement, Dialpad clears it on its own published evidence.

When is staying on Dialpad the right choice?

Staying on Dialpad is the right choice when your calling, contact centre and CRM connections already run there, the phone work is English-only, and you want AI agents on the same routing, recording and reporting as your human teams. It also suits buyers whose procurement requires published security certifications, which Dialpad states on its UK trust page. In those conditions a second vendor adds a contract without removing much work.

The single-vendor advantage is real. An agent that sits on the main line or a department through ordinary call routing is easy to switch on and off. Handover into a Dialpad contact centre stays inside one system, with the context Dialpad describes passed across. And Dialpad lists real-time and historical analytics among its capabilities, so a baseline taken before the pilot can come from the same reports the agent will later be judged on. Dialpad also says that 91% of calls receive an AI CSAT score, a vendor figure, but one that hints at how much measurement you get without extra tooling.

If you are in that position, we would stay and spend the effort on design and sequencing: write the skills, write the guardrails, test on the phone, and time the build so you are not rebuilding after the post-deal roadmap lands. Our voice AI ROI framework sets out the arithmetic for proving the case either way.

When does a separate voice AI platform make more sense?

A separate voice AI platform makes more sense when the phone work is a contained set of inbound or outbound calls, your callers need languages beyond English, your outbound campaigns need per-country calling rules, or you would rather not tie a new build to an agent product whose roadmap is changing. It also suits teams whose AI work spans phone systems other than Dialpad.

The trade is direct. You give up one admin portal, one report and one supplier, and you take on another contract, another data protection review and a handover boundary between two systems, usually a transfer to a Dialpad number or queue. In return you choose an agent layer on its own merits and keep it independent of your phone system choice. Whether that is worth it depends on who owns the work: prompts, knowledge, escalation rules and integrations still need a named owner on your side, whichever route you pick. Our wider enterprise voice agents guide sets out those operating questions, including the third path of assembling your own stack from developer tools.

Timing is the quiet argument here. A contained workload, such as an after-hours front desk or a reminder campaign from a contact list, can be piloted on a separate platform in parallel while your Dialpad account team publishes the post-deal roadmap. If the roadmap answers your questions, you can bring the work back. If it does not, you have a working line that never depended on it.

How does Dilr Voice differ from Dialpad's AI agents?

Dilr Voice is a multi-agent voice AI platform from DILR.AI that chains greeter, qualifier, knowledge, customer care and action agents into one phone call, each with its own prompt, model and tools. It runs on Twilio or Vobiz telephony rather than inside Dialpad, supports 30+ languages, and ships per-country compliance rules by default, so it sits beside a Dialpad estate rather than within it.

Dilr Voice applies recording consent, DNC registry checks, opt-out recognition, permitted calling hours and full audit trails on every call by default. Response time is under 500 milliseconds on our own on-platform measurement. Knowledge bases are uploaded documents, chunked automatically and searched during live calls, so agents cite from your own data. Outbound campaigns run from an uploaded contact list with scheduling windows, retry logic and auto-pause at the configured daily end time, and they need your own Twilio or Vobiz account. Post-call automation updates CRM records and sends email, SMS and Slack notifications without code, with HubSpot, Salesforce and Google Calendar named live.

The honest limits matter as much. Dilr Voice does not appear in your Dialpad admin portal or Dialpad reporting, no live Dilr page names a Dialpad integration, and a transfer into a Dialpad queue should be tested as ordinary telephony on your own numbers before you rely on it. We show no security certifications on our live pages, where Dialpad shows several, and hosting details for enterprise, including dedicated tenancy and regional data residency options, are set out in the Dilr FAQ. There is a free trial with no credit card required.

Frequently asked questions

What is the best Dialpad AI agents alternative for UK teams in 2026?

The best Dialpad AI agents alternative for UK teams in 2026 depends on the workload, not the brand. For a contained English inbound line already on Dialpad, staying with Dialpad often wins. For multilingual callers or outbound campaigns needing per-country calling rules, a separate platform such as Dilr Voice, PolyAI or Synthflow is worth testing, judged on phone tests rather than demos.

PolyAI and Synthflow each have their own alternative guide on this blog, and our voice AI RFP template turns the criteria into questions you can send. The voice AI category page lists every comparison we have published.

Can Dialpad AI agents make outbound calls?

Dialpad's help centre says a Greeting Skill runs on every new launch, including inbound, web chat and outbound bot conversations, and names outbound appointment reminders as one use case. That shows Dialpad AI agents can open outbound bot conversations. Confirm with Dialpad whether that covers outbound voice calls on your UK numbers, and which calling-hour and consent controls apply.

For outbound work in the UK, the rules on consent, opt-outs and calling hours carry as much weight as the agent itself, and under PECR they fall on the organisation that makes or instigates the calls, which is you rather than the software vendor. If you are unsure whether your outbound plan is ready for any platform, a short scoping call is the quickest way to find out, and our DATS methodology covers the wider question of where AI should sit in your operation.

Want to see this in production? Try Dilr Voice live, book an AI placement diagnostic, compare vendors in our best AI voice agent guide, or read about our approach to placing AI inside enterprise systems.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed.