A renewals desk for a UK energy brokerage now runs on an AI call center that answers more than 350,000 calls a month, according to xAIa, the Dubai voice-AI company. xAIa does not name the brokerage, but its own blog describes the deployment as the largest single engagement it has put into production, and lays out in detail what the system does on every one of those calls.
xAIa, which describes itself on its own site as the AI consultancy and implementation arm of PRO TECHnology Group, is based in Dubai. On its blog, the company states that a utility and energy brokerage in the UK runs “350,000+ calls a month through xAIa Voice AI, with close to 100% answered,” running on what the same blog elsewhere calls three AI employees sharing one operational layer.
What xAIa says the AI call center does on every renewal call
According to that post, the system is built around four jobs its AI calling agent is designed to do on the renewals desk: reading the consent script, chasing signed authorisation to act on an account, respecting a customer's request not to be called again, and working every account inside its scheduled renewal window.
- “The consent script is read in full, every time, exactly as your compliance team approved it,” xAIa's blog states — and it says every consent call is recorded, transcribed and stored.
- Letters of authorization are chased to signature, per the same post.
- “Do-not-call requests are honoured instantly and permanently. Calling windows are enforced by configuration, not by memory.”
- “Every account is called inside its window, on schedule, with the contract details already loaded,” and an interested customer is booked “straight into a broker's diary with the full context attached.”
Those four tasks sound administrative, but in an outbound renewals operation a mistake in any one of them becomes a compliance problem rather than a missed sale — calling a number that opted out, or skipping a step in an approved script, carries consequences a lead-qualification error does not.
Why consent and do-not-call rules carry real weight in this market
Outbound sales calling in the UK energy sector sits inside a well-enforced rulebook, and recent enforcement shows why a brokerage running this volume of calls would want every script followed to the letter. In May 2026, the UK's Information Commissioner's Office fined Energy Prices Direct Limited £160,000 over more than 700,000 unsolicited marketing calls made to numbers on the Telephone Preference Service and Corporate Telephone Preference Service registers between January 2024 and January 2025, according to the ICO's own announcement. “Compliance with TPS and CTPS requirements is a legal obligation for all organisations conducting marketing calls,” the ICO's head of investigations, Andy Curry, said in that announcement.
That case has no connection to xAIa or to the brokerage its blog describes — no regulator has examined xAIa's system, and nothing in the ICO's announcement suggests the fined firm used any AI calling platform. It is cited here only to show the scale of penalties UK regulators have levied against energy-sector telemarketing in this period, which is the backdrop against which a vendor markets scripted-consent and do-not-call features as selling points rather than footnotes.
The bigger numbers analysts are citing for AI call center adoption
Across all of its deployments, not only the UK brokerage, xAIa says it has processed more than 1,000,000 production calls and now runs above 600,000 calls a month through 63+ AI agents working in 90+ languages, averaging about 500ms of response latency, according to a company blog post published 6 September 2026. Those figures are xAIa's own, not independently audited.
The broader market xAIa is selling into is also moving, on the numbers analysts publish. According to the trade publication CX Today, reporting a Gartner forecast on 11 March 2026,“conversational AI deployments in contact centers will reduce agent labor costs by $80 billion globally” by 2026, as automation absorbs a growing share of routine interactions. That prediction describes the wider contact-centre industry, not any single vendor, and is cited here only to size the trend an AI call center product like xAIa's is riding — not as evidence of what its own renewals desk has saved.
The honest limits
xAIa does not name the brokerage running its largest deployment, publish how many renewal calls convert to a signed contract, or disclose any churn, complaint or regulator-contact rate tied to the system. Those figures would have to come from the brokerage's own records, not a vendor's blog, and nothing published lets an outside reader check the “close to 100% answered” claim against an independent log.
Frequently asked questions
What is an AI call center?
An AI call center is a contact-centre operation in which software agents — rather than, or alongside, human staff — answer or place calls, follow scripts, and complete tasks such as booking, verification or renewals without a person on the line for every interaction.
Is AI replacing call centers?
Analysts generally describe the shift as automating specific call types rather than a wholesale handover of the job. The Gartner forecast reported by CX Today, cited above, frames the change as AI cutting labour costs by absorbing a growing share of routine interactions — not as replacing contact centres outright.
What does xAIa's AI call center do about do-not-call requests?
According to xAIa's own description of this deployment, do-not-call requests are honoured “instantly and permanently,” and calling windows are “enforced by configuration, not by memory,” meaning the system is built to stop calling a number the moment it opts out, rather than relying on staff to update a list by hand.
Who is behind xAIa?
xAIa says on its own site that it is the AI consultancy and implementation arm of PRO TECHnology Group, which it describes as serving MENA enterprises and governments since 1998 across five companies, with xAIa its newest addition, founded in 2025.
Product details and capabilities in this article are as published by xAIa and the other organisations named at the time of writing.
