AI Driven Contact Center

Today’s contacts centres are overloaded with huge call volumes and call center executives are overworked. This makes those executives error prone and results in unsatisfied, frustrated customers.

What different reasons are for executives to be over worked?

  • Repeat queries and mundane tasks
  • Little or no time for solving the complex queries
  • Handling agitated customers

Classical Contact Center

There are two Layers of contact center executives, First Layer handlers all the queries coming in and in case there is some complex query, it’s passed on to second Layer. Now this is where Customer gets frustrated for repeating entire conversation and also time duration it takes to resolve the same. We have seen above there is large chunk of customers who feel frustrated for customer service giving bad reviews to great product.

There is one more issue with existing Contact centers, due to multi-layers of communication, Conversation Context being lost between Layers. Along with loss of Conversation Context, it takes lot of time to reach to right executives to get the final answer.

Since current Contact Center technology doesn’t allow to gather conversational inputs from Customers – all it can take inputs like key strokes from Phone Numerical key pad.

Since it’s conversational and largely human dependent, one more inherent issue is, likelihood of missing vital information and there by not resolving the query.

With the help of Chat Bots and Machine Learning technology, above issues can be resolved to great extent. There can be few more channels to engage customers with quality output. Since it’s going to be Machine learning and Artificial Intelligence – Customers sentiments can be understood, time can be saved and passed onto Live agent only for complex queries along with context of conversation. More than that, Customers should be able to resume their conversation later point in time from altogether new communication channel and place. This saves lot of time and results in happy customers.

Modern Contact Centers

Modern contact centers are complete only if they are taking care of following things –

  1. Multi lingual Chatbots – Supporting multiple languages like English, Spanish etc. This will also cover large Customer base.
  2. Transition from One Channel to another with no loss of Conversation Context. E.g. If Customer initiates his conversation on Chat and leaves it halfway, and later point in time – he initiates it over phone channel, earlier Conversation should be picked up along with user inputs provided then. Also customer should be given choice of continuing earlier conversation or start fresh.
  3. In case of Complex or new queries – Customers should be transferred to Live Agent with little or no time. Live Agent should also be given Conversation context so that Customer doesn’t have to repeat the same. Again Conversation Context should be retained and Customer need not repeat entire inputs all over again.
  4. Natural language inputs (along with slangs) should be taken with no hassle and acted up on with same accuracy. Customer should be able to directly jump onto his query and start conversation with Bots – saving time.
  5. User awareness programs with well-designed (sample) conversation flow so that vital information is not missed by the users.
  6. Last but not the least – Analytics of Different queries raised by Customers, time of conversation and resolved / unresolved queries and sentiments of customers
  7. All Queries taken with Agents which could not be answered by Bots – can be segregated and bots can be trained automatically.

The Technology Stack needed to accomplish above solution –

  1. Dialogflow from Google or Lex from Amazon
  2. Web app for Fulfilment (e.g. .Net)
  3. Database for Conversation Context storage (e.g. SQL server, Oracle)
  4. Amazon Connect for Phone Channel

Let me ask you few questions before concluding this blog –

  1. What is “First Call Resolution” rate of your Call Center? – Ideal should be 74%, although it varies from industry to industry and it’s in the range of 41% to 94%.
  2. What is the “ Time in Queue” of your Call Center? – Customers should start getting answers from BOTs immediately and so “Avg. Time in Queue” will be negligible or zero?
  3. What is “ Abandonment Rate” of your Call Center? – If my “Avg. Time in Queue” is zero – I am sure, Abandonment Rate will drastically go down.

Ideally 50% of “First Call Resolution” calls should be handled by BOTs and should be resolved then and there.