Contact centres are the places where a company's agents answer customers' calls, chats, emails and messages, together with the software they use. In India, a contact centre is usually called a customer care centre, or a call centre. A large one has hundreds of agents on one floor, or working from home on the same system. Banks, airlines, telecom operators and online shops all run them.
The software behind a contact centre does five jobs. It sends each contact to the right agent. It records the contact. It measures how each agent is doing. It forecasts how many agents will be needed, for example on a Monday morning. And, more and more, it answers the contact itself, before any agent is involved.
Consider a credit card helpline on the first working day of the month, when bills arrive. Thousands of customers call about charges they do not recognise. Contact centre software receives each call and sends it to a free agent who knows the caller's language. Workforce engagement management (WEM) had already forecast this rush and scheduled extra shifts. A conversational AI platform runs the bot that answers the simple questions, such as the due date. An AI customer service agent goes further and solves a whole request, such as blocking a lost card and ordering a new one. During each call, the agent can look up how a similar problem was solved before, in knowledge management for customer service. After each call, conversation intelligence turns the recording into text and scores it. The contact centre also makes calls of its own, through a predictive dialler, for example to remind customers about a payment.
This software used to run on equipment in the company's own building. Over the last decade, much of it has moved to the cloud. The industry calls the result contact centre as a service, or CCaaS. Companies usually pay for each agent who uses it, and each agent's licence is called a seat. Contact centre software runs on top of the company's phone system, the cloud phone system.
The biggest change now is the contact that no agent sees. Most platforms now sell a bot that answers first and passes to a person only what it cannot handle. So some vendors now charge for each problem solved, instead of for each seat, as explained in SaaS pricing models.