What is agent assist?

Agent assist is when machine learning and artificial intelligence technology are used to help provide customer service representatives with relevant information in real time for more effective, self-service customer support. For example, agent assist tools might surface common responses to frequently asked questions, and provide real-time guidance across multiple systems to find answers and insights. Their role as a copilot is to improve human agent productivity, optimize overall customer experience and reduce operational costs in a single agent workspace.

While prescripted chatbots can be useful for handling frequently asked questions (FAQs), more complex virtual contact center agents can extract insights from customer data and suggest possible solutions for human agents from the company website or app. Today, most call centers use some form of agent assist technology to address customer inquiries, simplify workflows and automate certain customer service functions.

Agent assist technology evolved from computer-based systems for automatic call routing and intelligent virtual assistants that can perform simple tasks such as setting timers or retrieving weather information. But sometimes, customer needs are complex and have multiple components. Chatbots and virtual agents often cannot address these alone, creating a need for a technology that uses the benefits of human agents with the knowledge base of AI-powered tools.

Agent assist uses conversational AI technology such as natural language understanding (NLU) and natural language processing (NLP) to analyze the information and context around customer issues. It also applies speech recognition to transcribe customer calls and provide summaries of customer interactions.

For example, AI can analyze call transcriptions and identify keywords and phrases to retrieve relevant information related to the customer’s account or a specific document they’re referencing, such as billing statements or receipts. This can help human agents streamline their conversations with customers and provide issue resolution with more efficient response times and give them metrics to use for future interactions.

Separately, AI agents have become a popular technology tool, though they differ from agent assist. An AI agent refers to a system or program that completes tasks autonomously or on behalf of users. It uses machine learning algorithms to create baselines and identify deviations based on the initial input.

In addition to streamlining workflows and empowering human agents, agent assist technology can also result in more sensitive customer care and provide actionable insights tailored to individual users. For example, AI can provide feedback on customer satisfaction by performing sentiment analysis and suggest when the human agent should pause and either offer empathy or acknowledge the customer’s frustration.

Some companies use customer satisfaction scores (CSAT) to quantify customer experience. Automation is 1 factor that can improve an organization’s CSAT score. If the conversations veer off track or the AI’s automated suggestions are incorrect, human agents can take over.

Agent assist technology can help lower rates of burnout and improve the agent experience as well. Automating repetitive tasks, such as marking an issue as closed when it detects that a customer has left the chat or scheduling follow-up action items based on the customer conversation can have a significant impact on reducing the manual workload.

Agent assist has been shown to reduce issue resolution time by 26%. It can boost customer satisfaction with answers by about 150%.

Here are some examples of industries where agent assist technology can be used.

• Human resources 
• IT help desk
• Customer service
• Financial services 
• Healthcare

In the hiring process, agent assist can advise HR staff on salary recommendations during contract negotiation calls. During the onboarding process for enterprise employees, agent assist technology can help HR staff focus on more valuable work by automating responses to HR policy questions. It can also consider the differences in policies based on each role and based on where the employee is located in the world.

Case study: East and North Hertfordshire NHS Trust uses agent assist to support their HR staff in handling employee queries related to management training, policies and regulations, schedule and pay.

IT help desks can use agent assist to help them keep their ticketing system organized. As help desk specialists assist customers with troubleshooting, agent assist can identify common software issues and suggest step-by-step instructions on how to diagnose the problem. It can also retrieve relevant information from application documentation and video guides.

Case study: In a paper by IBM researchers (link resides outside ibm.com), agent assist might provide quick solutions related to application documentation, ticket management systems and knowledge transfer video recordings. It has been used in 650 projects within IBM.

Customer service call centers can use agent assist to address customer complaints. This feature can help human representatives understand the customer complaint and pull up relevant purchase history, habits, preferences and previous interactions to suggest personalized pricing offers or discounts to help with business retention. Call centers can also use generative AI to field customer calls and direct them to the appropriate human agent or department.

Case study: Brazil’s Bradesco bank uses agent assist to answer 283,000 customer service questions a month across 62 different products. Its agent assist has a 95% accuracy rate, response time of seconds, with just 5% requiring calls for further assistance.

With agent assist, banks can free up client advisors from memorizing the details of different products, services and banking offers so they can focus on building client relationships. For example, mortgage service providers can use agent assist to suggest policies and answer complex queries related to each specific policy.

Case study: Crédit Mutuel uses agent assist to sort and answer half of the 350,000 daily emails received by the bank’s client advisors. The technology helped advisors find answers up to 60% faster.

Healthcare administrators can use agent assist to help them with routine tasks such as entering data, scheduling appointments over the phone, verifying insurance information, reviewing hospital visit guidelines that patients should follow and invoicing for various healthcare services. It can also help insurance providers answer patient questions about healthcare plans, coverage policies for certain services and resources for specific conditions. For patients interested in clinical trials, agent assist can allow clinical coordinators to answer questions about eligibility, requirements and timeline of the study.

Case study: Humana uses agent assist trained in healthcare terminology to help healthcare providers get answers related to patient insurance coverage across various data points. It can field 7,000 voice calls from 120 providers per business day.