The impact of AI adoption for telecoms customers - Which? response

Use of AI in the Telecoms' customer journey 

In March 2026, Which? responded to an Invitation to Contribute consultation for an Ofcom research project examining the impact of Artificial Intelligence (AI) on the experience of telecoms customers in the UK. The full response can be found by clicking the link below. The response is divided into three sections, a summary of each section follows. 

Current adoption: opportunities and risks

• Current adoption: Consumer-facing AI is currently deployed in two main ways within the UK telecoms market: assisting call-centre agents and powering customer-facing chatbots (e.g., EE’s AIMEE, Vodafone’s tobi). It is possible that telecom providers are using AI for other consumer-facing use cases, such as automation of marketing campaigns. However, we do not examine this in depth.
• Opportunities: Customer service chatbots may provide a route for customers to resolve issues without having to contact their provider via a phone call.
• Risks: AI chatbots may face limitations when applied to customer service, such as their capacity to interpret customer messages accurately, or provide accurate and relevant information. Increased reliance on AI chatbots could lead communication providers to reduce human-centred customer service resourcing, which could adversely impact vulnerable consumers or any consumer that does not feel confident using digital technologies.

Future developments: opportunities and risks

The future adoption of AI presents both opportunities and risks for the UK telecoms customers. The future adoption of AI in the UK telecoms market could result in greater levels of personalisation and uses of agentic AI. 

• Opportunities: Personalisation could create benefits for consumers by presenting more relevant products and services to customers.  Personalised pricing could also be used to benefit customers if customers are offered better value services aligned with their usage needs. Similarly, Agentic AI—autonomous systems acting on behalf of customers–could mitigate customer inertia and high switching costs by automatically finding better deals for customers as well as automating complex requests. 
• Risks: However, increased personalisation, specifically personalised pricing, could result in fairness issues which could have disproportionate impacts on vulnerable consumers. Although telecoms companies do use some forms of automated personalisation, we are not currently aware of personalised pricing being applied consistently within telecoms markets. Agentic misalignment, where an AI agent's actions diverge from a customer's true preferences (e.g., choosing a good value package but with poor contract flexibility) poses possible future risks for the telecoms markets unless mitigations are developed. 

The adequacy of Ofcom’s rules 

We recommend that Ofcom ensures regulatory collaboration by working with the ICO and CMA to address overlapping regulatory issues that arise from the use of AI in the telecoms sector. This may include issues related to consumer protection, data protection.

We also recommend that Ofcom follow the example of the CMA and their recently published guidance on complying with consumer law when using AI agents by issuing guidance for telecom providers relating to the deployment of AI agents within telecoms markets. 

With particular reference to Ofcom’s General Conditions, we recommend:

• General Condition C2 (Transparency): covers information and transparency requirements for telecoms providers. It is important that providers are transparent with customers about the use of AI on their platforms and notify them when they are interacting with an LLM. 
• General Condition C4 (Complaints): Ofcom must ensure its obligations for providers to maintain multiple customer communication channels continues to be adhered to. Customer service chatbots, when in use, must clearly notify customers how to escalate their issue to a human representative if necessary.