Is AI Delivering Productivity Gains in Contact Centers?

October 2026

AI innovation continues at an extremely rapid pace regardless of regulation, bringing both enormous potential and risk to enterprises. The question businesses must ask is: how can they leverage AI to realize measurable productivity gains and a lasting reduction in operating expenses? Contact centers worldwide are considering their options as GenAI and Agentic AI are ideal for these people-intensive organizations. GenAI is useful for its ability to understand unstructured language, accurately identify “intents,” and apply this information to drive desired outcomes. Agentic AI is emerging as a powerful tool because it can apply automation and workflow logic, reducing the need for humans to handle routine customer inquiries and interactions. However, for AI to be viable in contact centers, it must cost less than the work it replaces and must also deliver sustained, measurable productivity gains. Additionally, leaders must factor in practical issues, such as the time, effort, and risk of implementing AI solutions. 

DMG’s research has found that AI delivers real productivity benefits in specific and well-defined use cases, but the gains are smaller and slower than vendor claims, and they do not generally result in significant agent reduction. The applications delivering these benefits are:

1. Post-interaction summarization, also known as automated after-call summarization and wrap-up – this activity was not previously automated and therefore is an ideal task to address with AI. Companies have reported a 20% – 60% reduction in after-call wrap-up, depending on the contact center business function, complexity of topics discussed, and time allocated to after-call wrap up. Additional automation opportunities are extending the automated after-call wrap-up capability to include and automatically kick-off follow-up activities using agentic AI. (And there are potential benefits by using the after-call application to identify and document the follow-up activities, but to have them initiated by an agent or back-office employee.)

2. Conversational AI (CAI) – customer-facing self-service – first time users of CAI and companies that previously experienced a low automation rate (20% or less), such as many in health care, are experiencing a measurable increase in self-service utilization and call displacement when automating routine, high-volume inquiries. However, to realize full call displacement on routine calls, the organization must follow emerging CAI implementation best practices, which include starting with 1 or 2 high-volume use cases, actively involving the organization’s employees in the testing on an ongoing basis, creating and actively maintaining appropriate guardrails, and then scaling each use case. In most cases, the savings allowed organizations to scale and did not result in major agent layoffs.

3. Agent assist/agent augmentation, the agent-facing side of CAI – position these bots to identify and deliver appropriate case information, knowledge articles, policies and procedures, workflows, and sales and collections approaches directly to human agents as they interact with customers. This allows human agents to spend their time helping customers instead of searching for the information they need to resolve inquiries or sell. Sales and collections opportunities present a much larger upside potential than the savings coming from reducing agent average talk time in service organizations. However, to realize benefits, the solution must be highly customizable and easy to integrate and implement. Additionally, companies need a single source of data to ensure that customers receive the same answer whether they interact with a CAI solution of an agent. Organizations utilizing agent augmentation tools are benefiting by scaling their resources, rather than reducing their agents. 

Final Thoughts

AI is delivering productivity gains in contact centers, but generally not in the form or at the size many enterprises were promised. The benefits are real for targeted and well-defined use cases, and organizations that implement them with discipline and best practices are realizing measurable results. In general, the gains are being used to absorb growth, improve the customer experience, and free agents to handle more complex and valuable work, not to drive large reductions in staff. However, for any AI initiative to be successful, it must cost less than the work it replaces and deliver gains that last.