Market insight

Customers still expect a route to a person.

Automation is accepted when it solves the issue and when a person is reachable if it does not. Enterprises are redesigning human roles rather than removing them.

Key points

Customers accept automated service on condition that a person remains reachable, and a failed automated interaction makes them markedly less willing to try again.

Service organisations are, on the whole, redesigning human roles rather than removing them; some that reduced staff because of AI have already rehired.

Escalation design – when automation stops and how the case is handed over – is part of the service, not a fallback.

What customers expect

  • 87% of 3,566 customers said companies using generative AI for service must provide access to a human agent.

    Gartner, press release, 4 August 2026. Survey February–March 2026; a customer preference, not a legal requirement.

  • Only 27% of customers would try a chatbot again after a negative experience; 7% had used a chatbot in their most recent service interaction.

    Gartner, press release, 2 September 2026. 3,566 B2B and B2C customers, February–March 2026; chatbots, not only voice.

  • 84% of consumers consider service poor if AI does not resolve the issue within three attempts.

    Zendesk, CX Trends 2026. Vendor research.

  • 49% of respondents described their experiences with AI in service as positive (up 9 percentage points year on year); 54% trusted AI only for simple issues. Easy escalation and being told they were speaking with AI were named as key conditions.

    Intercom, 2026 AI Sentiment Report, 9 September 2026. Vendor research; more than 1,000 respondents.

  • Around one in twelve UK internet users used AI to help manage their telecoms services, while human contact remained valued.

    Ofcom, Consumers using AI to help manage their telecoms services, but human contact still valued, 25 September 2026. United Kingdom only.

How service organisations are responding

  • 21% of 275 service leaders had reduced human-agent headcount because of AI; 7% had already rehired. Among those that reduced headcount, the average reduction was 16.7%.

    Gartner, press release, 5 November 2026. Survey July–August 2026; the 16.7% applies only to organisations that made reductions.

  • 85% of 321 service leaders were expanding human agents’ responsibilities; in a separate survey, 31% had implemented or planned AI-related staff reductions by Q1 2027.

    Gartner, press releases, 28 April 2026 and 27 April 2026. Survey September–October 2025; international sample.

  • 68% of medical-practice leaders had not changed roles or structures because of AI; 26% had.

    MGMA Stat poll, June 2026. 260 responses; US medical practices.

  • 77% of organisations using AI agents allowed handover to a human.

    Salesforce research (3,075 respondents, 13 countries), as reported by ZDNet, 24 June 2026. Vendor material; secondary report.

Longer-term forecasts

  • Forrester forecasts a 13.5% fall in US customer-service representative jobs by 2031 (from 2.59 million to 2.24 million) and recommends managing reductions through attrition rather than mass lay-offs.

    Forrester, The Quantitative Employment Impact of AI on Customer Service Jobs, July 2026, as reported by CMSWire (5 June 2026) and No Jitter (3 August 2026). Forecast; US only; secondary reports.

  • Respondents expected agentic AI to handle 68% of customer-service and support interactions with technology vendors by 2028; 89% said the human element remained important.

    Cisco, The Race to an Agentic Future, 27 May 2025. Expectation of 7,950 decision-makers in 30 countries; concerns customers of technology vendors.

Designing the route to a person

Vee's interpretation. The analysis below is our own reading of the sources above, not a finding of the organisations cited.

A well-run automated process defines in advance where it stops: a stop condition when the customer asks for a person or the case falls outside the agreed scope, a callback when no one is available, and a transfer that carries the case history so the customer does not have to repeat it.

The Gartner finding that only 27% of customers would try a chatbot again after a bad experience implies that a poorly designed escalation is costly beyond the single interaction: it reduces the share of future contacts that automation can handle at all.

Escalation rules are therefore part of the specification, tested and measured like the rest of the process: how often transfers happen, why, and what the person receiving the case can see.

By sector

Telecommunications: Ofcom’s research shows UK customers beginning to use AI tools themselves to manage services, while still valuing human contact. Contact flows are likely to include both automated and human steps for some time.

Healthcare: the MGMA polls describe US practices adjusting front-office work rather than replacing staff. Appointment and patient contact often involves people under stress or with limited digital skills, which raises the importance of a clear route to a person.

Banking, consumer finance and insurance: conversations about arrears, restructuring or claims can lead to decisions with financial consequences. Where such decisions are taken, the point at which a person takes over is both a service-quality and a regulatory question (see Regulation and trust).

Workforce implications

The survey evidence does not describe a simple substitution of people by software. It describes a shift in what people do – more complex cases, exceptions, supervision and quality control – alongside selective reductions, some of which have been reversed.

Long-range forecasts, such as Forrester’s 2031 projection, are scenario estimates for the US market. They are useful for direction, not as a measure of what any individual organisation will do.

What to watch

The 31% of service leaders who, in Gartner’s April 2026 survey, had implemented or planned AI-related staff reductions by Q1 2027, and whether the 7% rehiring rate reported in November 2026 rises.

Forrester’s Predictions 2027 (30 October 2026): up to 10% of consumers may use a personal AI agent as their main service interface in 2027, with growing bot-to-bot traffic. Forecast.

How national authorities apply the EU AI Act’s Article 50 duty to inform people that they are interacting with an AI system, in force since 2 August 2026.

Vee in context

Vee’s Q3 2026 report (letter to shareholders, p. 12) states that Vee is developing the control, supervision and security mechanisms needed as automation is used more widely, that the division of tasks between specialists and the systems supporting their work is changing, and that analytical and managerial competences are gaining importance. The letter adds (p. 14) that the growing role of automation affects the organisation of teams, the responsibilities of specialists and managers and the way work is delegated and supervised.

Market data on this page come from third-party publications and are attributed to their authors. They describe the market, not Vee’s results. Forecasts are the publishers’ own and are not forecasts by Vee S.A. Figures were checked against the publishers’ pages on the date shown.

Other market insights

The AI value gap

Spending on AI in customer service is rising, but few organisations report the return they expected. The gap sits between deploying a model and running a process that delivers a measurable result.

Read the insight

Integration and legacy

In large organisations, AI meets existing telephony, CRM, collection and policy systems. Most barriers reported by enterprises are about connecting to those systems safely.

Read the insight

Contact data

Organisations record almost every call, yet analyse only a small fraction. Process data is where measurable improvement comes from.

Read the insight

Regulation and trust

In the European Union, transparency in interactions with AI is now a legal obligation. Regulated sectors expect controls, documentation and clear disclosure that a customer is speaking with an AI system.

Read the insight

Service economics

Budgets are moving towards AI, but the cost of automated resolution is not falling to zero. Buyers increasingly ask what a process achieves, not how many minutes it uses.

Read the insight