How CRM with AI is Transforming Customer Service


  • The question for customer service leaders isn't if Artificial Intelligence will impact their business, but how and when.
  • Intelligent call routing, sentiment analysis, text and speech analytics, voice of the customer and chatbots are some of the AI-driven capabilities transforming customer service organizations.
  • Customer service leaders use AI technologies to improve the customer experience (CX), lower operating costs and better leverage contact center agents.
  • AI as part of customer service doesn't just benefit customers, it delivers an improved agent experience which increases agent productivity and decreases staff turnover.
Johnny Grow Revenue Growth Consulting

Artificial Intelligence (AI) is the simulation of human intelligence by machines, and includes data analysis, decision-making and self-improvement.

AI in customer service uses machine intelligence to personalize customer engagement, deliver faster services, lower cost to serve, increase customer satisfaction and scale customer service operations.

For example, AI aids agent productivity with guided service fulfillment, next-best-action recommendations, suggested knowledgebase articles, case resolution responses and personalized offers. It can deliver push-based coaching and suggest how to use customer data to deliver differentiated customer experiences.

It allows Customer Service Representatives to better solve for the customer. AI can automatically apply customer transaction or case history, and even customer personas, to deliver more personalized and contextual customer support.

Customer service leaders were early adopters in AI and gained both strategic advantages and impressive ROI. However, AI in customer service is no longer in the early days. We crossed that chasm and any customer service organization not leveraging the synergistic combination of CRM and AI is clearly substituting labor for technology automation while incurring increased costs and decreased agent and customer satisfaction.

CRM Service Cloud leader Salesforce surveyed service organizations to see how they are using this technology.

Customer Service AI Use Cases

See the research results showing how customer service organizations are using #AI to improve customer, agent and contact center outcomes.

Click to Tweet

Based on our own experience, we've expanded upon the Salesforce findings to go deeper into some customer service use cases where AI drives significant and sustained company and customer benefits.

Intelligent Call Routing

Ever navigate through a maze of vague and sometimes irrelevant phone prompts, enter the requested information, wait on hold for a while and then be greeted by an agent who requests the information previously entered? Yea, me too, and it immediately puts the customer experience in a bad place.

Two top customer service complaints are getting to the right agent and taking too much time to explain the issue to an agent not familiar with the situation. Intelligent call routing can solve both these problems.

AI infused call routing accelerates the customer journey by routing the customer to the right agent. It uses the information acquired to find an agent with the skills to resolve the issue.

Call routing integrates with CRM software to direct calls or tickets based on customer need and privileges such as Service Level Agreement, entitlements or loyalty status.

Intelligent call routing may use customer intelligence (such as customer insights or a 360-degree customer view), customer signals (such as clicks, views and hold times), and call classification (based on user history or current trends) for intent prediction. Calls are routed to agents based on skills, availability, channel, customer history or other factors that ensure each caller gets to the best destination. When customers get to the right agent the first time, there will be a sharp rise in Customer Efforts Score (CES) and Customer Satisfaction (CSAT) ratings.

Sentiment Analysis

Electronic channels conceal body language and facial expressions making it difficult for customer support representatives to gauge the customer's emotional state and the importance of an incident.

AI technology such as sentiment analysis can decipher customer sentiment by analyzing content and assessing customers' emotional states, and use this information to help reps classify, prioritize, route and escalate cases.

We worked with one product, Cogito integrated with Salesforce, to perform contact center in-call voice analysis based on speaking patterns such as voice amplifications, word choice, and other conversational dynamics.

The sentiment analysis identified each caller's emotional state and provided agent guidance with a color-coded meter that visually showed fluctuations in customer sentiment in real-time. If the agent reacted prematurely or off topic, or isn't understanding the customer, the meter would change from green to yellow or red. The tool can then suggest that the agent confirm the real problem, slow down their speech or allow the caller more time to speak.

This guidance was extremely helpful in maintaining an effective conversation and resolving the incident amicably. We found that agents effectively used the tool as a layer of emotional abstraction between themselves and the customer. It helped agents detach themselves from the customers emotion, to objectively view the emotional state without the temptation of being pulled into the customers emotions, and potentially becoming defensive or exacerbating the situation. With this tool, CSAT scores improved 19 percent after 60 days.

Sentiment analysis varies by channel, such as call, email or electronic submission of a case, but generally considers factors such as intensity, empathy, participation, interruption, tone and pace. Sentiment detection may classify calls or tickets as frantic, frustrated or excited. Frantic calls may be routed to a customer retention team, while excited calls are routed to an account manager.

Customer service organizations use this type of artificial intelligence to read, analyze and tag customer cases, route the case to the best agent, advise the agent of the customers emotional status and provide real-time guidance on how to best connect, converse, deescalate or engage for an optimal outcome.

Customer emotion is a powerful influence and sentiment analysis can decipher emotion to increase empathy, rapport and trust.

Voice of the Customer (VoC)

Innovation is one of only four sustainable competitive advantages. Whether your innovation is incremental improvement to existing products or services, or the creation of new products, customer markets or revenue streams, the best source for innovation is customer feedback.

Contact centers or customer service operations use a combination of Voice of the Customer programs and artificial intelligence to acquire, analyze, classify and tabulate customer feedback.

Sources of customer feedback include social networks, online review sites, community forums, blogs, Net Promoter Score (NPS) surveys and the company’s customer service case management system.

AI-based text analytics, or Natural Language Processing (NLP), can scour these sites, group emotional and other types of meaningful keywords, and quickly decipher customer feedback of the company or a product as positive, neutral or negative. They can extract insights such as what customers like, dislike or want. They can classify insights into categories such as price/value, product quality, fulfillment/delivery and customer support. Online product review sites are especially powerful in extracting product complaints and suggestions.

Feedback is a gift. Knowing what customers like or don't like about your company or products is an invaluable source information and innovation.


It seems there's no discussion about contact center AI use cases that doesn't include customer service chatbots. Maybe that's in part because they are especially well suited for the high volume of repetitive and mundane questions that bore agents.

According to IBM, 52 percent of customers hang up on customer support when the wait becomes too long and 50 percent of those customer support calls are left unresolved. Lingering and unresolved customer frustrations directly contribute to brand deterioration and customer churn.

Chatbots can resolve much of this pervasive problem by giving customers on-demand access to a self-service channel that is ideal for simple and commonly asked questions.

Chatbots, sometimes called virtual agents or virtual assistants, are software programs that conduct on-line chat conversations via text or speech-to-text.

Intelligent chatbots are conversational customer service platforms that use AI technologies such as sentiment analysis, Natural Language Processing (NLP) and predictive analytics to interpret text or voice-based customer questions, predict customer behavior and provide accurate responses such as complete answers, troubleshooting steps, knowledgebase articles or other content. When chatbots are unable to resolve the customer inquiry, they transfer the customer to a live agent.

Most Intelligent chatbots are integrated with CRM software so they can access customer history. They may also be integrated to ERP software, so they have visibility to product orders, fulfillment and payment history.

There are many use cases for chatbots with artificial intelligence.

  • Deliver 24 by 7 support and give customers on-demand customer service.
  • Give customers a service channel they prefer. There are many customers who prefer to avoid contact centers. Sometimes they dread clunky IVR routings or long waits, fear being endlessly passed around, are concerned they cannot understand agent dialects or accents, or they just prefer not to talk to a person. These people prefer a channel to resolve their own issues.
  • Give customers real-time service during peak periods. This also flattens peak demand periods by using this automated technology to provide relief to live agents.
  • Lower agent labor needed for simple and repetitive questions. A tremendous amount of labor is used to respond to simple questions that could be handled by a chatbot. Redirecting menial incidents to bots also improves the agent experience.
  • Escalate when necessary. Simple questions are resolved by the bot, but if the conversation becomes complicated the bot transfers the conversation to an agent.
  • Create the power of scale. Unlike customer service reps, bots can scale to accommodate any number of customer inquiries. They can be trained on new service transcripts to support new products or services, or even enable territorial expansion into new geographic markets with local language support.

Our experience has been that generalized bots don't work and annoy or alienate many customers. However, chatbots with artificial intelligence can successfully accommodate about half of all customer service inquiries and over time become the preferred channel for many customers.