Telecom operators are spending heavily on AI, but almost all of it is pointed inward, at networks and operations, not at how an outside AI assistant describes the carrier to a customer deciding whether to switch. That gap is the whole story right now. NVIDIA's own 2026 telecom survey found 90% of operators already say AI is lifting revenue and cutting costs, while a separate Ofcom study found a smaller but real and growing share of UK consumers are already asking AI tools to compare providers, troubleshoot a problem or draft a complaint. A telecom brand that has not thought about what those assistants say about its network, pricing or outage history is leaving that answer to whatever the AI system can scrape together on its own.
How much are telecom operators actually investing in AI right now?
A lot, and the money is still accelerating. NVIDIA's fourth annual "State of AI in Telecommunications" survey, published February 19, 2026 and based on 1,038 industry respondents surveyed between September and November 2025, found that 90% said AI is already helping increase annual revenue and drive down costs. Looking ahead, 89% said their organization's AI budget will increase over the next 12 months, up sharply from 65% who said the same in the prior year's survey.
Where that money goes is telling. Roughly half of respondents named autonomous network operations as their best-performing AI use case for return on investment, ahead of customer service at 41%. That is a reversal from earlier years, when customer-facing chatbots were the obvious first AI project for most carriers. The industry's own telling of its AI story in 2026 is about keeping networks running themselves, not about how a customer-facing assistant represents the brand.
A survey from Twilio, run among 985 mobile industry professionals at Mobile World Congress in Barcelona from March 2 to 5, 2026, points the same way from a different angle. It found 60% of respondents already using conversational AI in production, with a further 24% piloting or actively planning deployments, and 82% saying they had seen measurable value such as improved ROI or operational efficiency. Conversational AI here means operator-built chat and voice tools for customer support, not the general-purpose assistants a consumer might ask about a plan. Both figures are self-reported, from vendor-run surveys, so treat them as a measure of stated confidence rather than audited results. But the direction across two independent surveys is consistent: telecom's AI spending is real, it is accelerating, and most of it is aimed at the network and the contact center, not at the public record an outside AI system reads.
Source: NVIDIA's State of AI in Telecommunications survey (Feb 19, 2026) and Ofcom's 2026 Technology Tracker (Sept 25, 2026), both linked in the article.
Are telecom customers actually using AI to manage their service?
Yes, and the number is smaller than the operator-side figures but growing from a real base, not a hypothetical one. Ofcom's 2026 Technology Tracker, based on face-to-face interviews with 4,033 UK adults conducted by Ipsos between March 6 and June 30, 2026 and published September 25, 2026, found that 8% of UK online adults, about 1 in 12, had used an AI tool specifically to help manage a phone or broadband account. Of that group, most used AI to try to fix a problem, a smaller share used it to compare options, and a smaller share still used it to complain.
Zoom out from telecom specifically and the base rate is much larger: the same tracker found 48% of UK adult internet users had used a general-purpose AI tool such as ChatGPT, Gemini, Copilot or Meta AI in the past year, and 23% had interacted with a customer service chatbot of some kind, with 53% having done at least one of the two. On trust, 56% of people who had used AI said they trusted the result the last time they used it, 21% did not, and 22% were unsure.
Put those two figures together and the picture is a market still forming, not one that has arrived. Most telecom customers are not yet opening an AI assistant to shop for a new provider. But a meaningful and growing slice already are, for the three things that matter most in this category: fixing a problem, comparing a deal, and escalating a complaint. Those are exactly the moments where an inaccurate AI answer costs a carrier a customer it never knew it was about to lose.
What's the gap between operator AI spending and customer-facing AI visibility?
It is the difference between AI that makes your network run better and AI that decides what a prospective customer hears about you before they ever call. Telecom has clearly internalized the first kind. AI visibility for logistics and supply chain brands and AI visibility for fintech brands describe a similar pattern in other regulated, infrastructure-heavy industries: real operational AI investment running well ahead of any deliberate effort to manage how outside AI systems describe the company.
The risk in telecom is sharper than in most categories because switching is cheap and comparison is the entire point of the customer journey Ofcom measured. When a customer asks an AI assistant to compare mobile plans, coverage in a specific area, or roaming charges, the assistant answers from whatever public information it can find: your own site, regulator filings, review sites, news coverage, forum threads and complaint boards. AI chatbots mostly do not send a click back to you when they answer, so a wrong or outdated answer rarely produces a complaint you will see. It produces a customer you never hear from because they chose a competitor on the strength of an answer you did not know existed.
What does this mean for a telecom brand's PR and AI visibility strategy?
It means the comms function that has mostly been pointed at press and regulators now has a third audience: the AI systems a growing share of customers consult before calling customer service or signing a new contract. That does not require a new discipline so much as applying the existing one somewhere it has not been pointed yet.
Pricing, coverage maps, current promotions, outage status and complaints-handling commitments should be stated plainly and kept current on pages an AI crawler can actually read, not buried in a PDF tariff guide or behind a login. Regulatory data matters here too. Ofcom publishes a quarterly telecoms and pay-TV complaints report ranking providers by complaints received per 100,000 customers, and most markets have some equivalent public record. That kind of third-party, regularly updated data is exactly the sort of source an AI system can cite accurately, so a carrier with a genuinely good complaints record benefits from that record being easy to find and reference, not just mentioned in a one-off press release. Earned coverage still drives the large majority of AI citations across every industry Muck Rack has studied, which means a trade press mention of a new network rollout or a fixed outage response does more for AI visibility than another page of marketing copy on your own site.
What should a telecom brand actually do about AI visibility today?
Four things, roughly in order of how much they will matter over the next two years.
- Run the test yourself. Ask ChatGPT, Gemini and Copilot how your plans, coverage and complaint record compare with your two biggest competitors, and note where the answer is wrong, outdated or simply missing.
- Keep the comparison facts current and crawlable. Coverage maps, current pricing, outage status and the regulator's own complaints data should all be easy for a crawler to read, not locked in a PDF.
- Earn the trade and consumer coverage that AI systems already trust. A reported story about a network expansion or how an outage was handled carries more weight with an AI system than anything published on your own site.
- Treat the operator-side AI investment as a story worth pitching, not just an internal project. Network automation and AI-driven service improvements are genuine news in a category reporters cover closely, and that coverage becomes exactly the kind of source an AI assistant later cites.
The direct answer, for anyone who read this far
Telecom operators are investing heavily in AI, with 90% in NVIDIA's survey saying it already lifts revenue and cuts costs, but almost all of that investment targets networks and internal operations. Meanwhile, 1 in 12 UK adults already use AI to manage their phone or broadband account, mostly to fix a problem or compare a deal, according to Ofcom. The brands that close the gap between those two numbers, by keeping their pricing, coverage and complaints record accurate and easy for an AI system to find, are the ones that win the comparison a customer never tells them they ran. If you want help building that record, get in touch with our team.



