Human-and-AI Collaboration Models: How Virtual Agents and Humans Work Together for Better CX

Why AI alone doesn’t deliver CX success

Across the board, CX teams face growing customer demand. According to McKinsey, 57% percent of CX leaders say they expect call volumes to increase over the next 1-2 years ^[1].

Companies need to scale to maintain their service goals. No wonder that Gartner predicts that 80% of customer service and support organizations are turning to generative AI technology in some form ^[2].

Yet, CX isn’t just about answering more questions using virtual agents and AI-assisted customer experience. Many interactions are increasingly complex, requiring human skills like empathy. Balancing human and AI collaboration is the only way to achieve speed, efficiency and optimal customer experience.

Get it right and the benefits are clear. Get it wrong and the experience suffers. Customers and human agents get frustrated, and you don’t see promised efficiency or CX gains.

What’s needed is a hybrid approach that blends the best of humans and AI for customer service. This blog explores how to design smart, collaborative workflows that empower both humans and virtual agents to deliver better outcomes together.

Successfully implementing AI-assisted customer service

CX automation is increasing, but designing hybrid workflows is complex and strategic.
Demonstrating this, ContactBabel research found that 74% of US contact centers that offer web chat use chatbots or other virtual agents ^[3]. But nearly a third are neutral or negative when it comes to how satisfied they are with the technology. Customers are the same. They want fast answers, but also empathy and understanding.

Overcoming AI CX risks

CX leaders want to harness customer experience automation to free up human staff. At the same time, they want to carry on delivering the right experience to customers on every channel. There’s no point reducing incoming call volumes if customers struggle to get satisfactory answers from AI agents.

In fact, failing to integrate AI and humans successfully brings additional risks:

• Customer frustration as they can’t escalate from AI to human channels. Or, if they can handoff, they have to repeat themselves to the human agent.
• Agent frustration as they have to deal with angry customers, without knowing the full context
• Inefficiency as conversations are duplicated across humans and AI
• Potential confusion if AI and humans give inconsistent responses to queries

Why Hybrid? Because CX Isn’t One-Size-Fits-All

Customers get in contact with a wide range of needs. Often what they are trying to do will change during a single interaction. They might want to find out basic information through a chatbot, then talk to a human to get more details, for example. Trying to apply a single approach – all automated customer experience, or all humans – simply doesn’t work.

Escalation and handoffs go both ways. Interactions might move between humans and AI multiple times, depending on customer needs and the job to be done.

It’s not about choosing between people and bots. It’s more about creating a hybrid approach that uses both intelligently.

Hybrid use case: Insurance

Take someone making an auto insurance claim:

• The initial call from a potentially upset or shaken up driver needs to be handled by an empathetic human agent. They triage the call and deliver reassurance
• AI bots then help the driver fill in forms, seamlessly automating processes
• The conversation then shifts back to human agents to go through next steps or to cover any outstanding questions
• If there’s a need to for tasks that require privacy, such as taking payment details, AI agents again come into play.

The result? A faster overall interaction, greater efficiency, customer satisfaction and employee engagement.

Hybrid makes that possible.

A New Model for Human-AI Collaboration

We’ve all seen the impact of silos on customer experience. Customers can’t switch between channels or departments without restarting the process from scratch. Frustrating, annoying, and unproductive.

It’s the same problem when AI and agents are siloed. There’s no holistic view, handoffs are poor and everyone suffers.

AI and humans = one team

What’s needed is to adopt a customer-focused, team approach. Rather than replace humans with AI, look at the roles and tasks involved in a typical customer interaction. Where can virtual agents handle routine tasks, and humans step in at the right moments with the right tools? This is the hybrid or human-in-the-loop approach.

Focus on the moments that matter and make handoffs seamless and transparent:

• Moving from finding answers to basic questions from AI to focusing on more specific and personalized information from a human agent.
• Escalating a problem from AI to agents, while providing them with full context to help solve the issue
• Handling security and identity verification using AI or IVR and then passing to an agent to begin the conversation
• Automating taking payments or filling in forms, with AI supporting the customer, ensuring privacy and confidentiality

The benefits of a hybrid approach

The hybrid approach is good for the business, customer experience and employee engagement:

• Businesses optimize their use of customer service resources, strengthening customer loyalty and retention
• Customers feel understood and valued
• Agents are less stressed and more fulfilled as they handle interactions where they can make a real difference

Collaboration Models That Work

Hybrid CX models should flex with customer needs and complexity levels and map to the overall journey.

While every journey is different, combining AI and human agents delivers real benefits through four common approaches:

1. Assistive AI

AI supports humans in real-time when they are speaking or interacting with a customer. Agents can access accurate answers from an AI-driven knowledgebase or AI can listen in on the conversation and feedback in real-time. It can suggest information, ensure that mandated areas, such as policy information, are covered, or flag to supervisors if sentiment levels are worsening and the agent needs support.

Best for: Reducing interaction times for high-value calls where human input is required or for training and supporting new agents.

Pros: Faster for customers, empowers agents, ensures consistency

Cons: Potentially resource-intensive, assistive AI must be seamless and not disrupt the conversation flow

2. Sequential Handoff

The customer starts by talking to AI, which takes them through routine, standardized processes such as paying a bill or resetting their password. If they need support or have other queries, AI hands over to a human agent, arming them with a complete history of the interaction to date.

Best for: Automating common support journeys where speed and ease of use are vital.

Pros: Greater efficiency, customer gets job done quickly

Cons: Doesn’t work well for more complex interactions

3. Loop-In Models (Human-in-the-loop)

AI handles the full flow but flags complex or sensitive cases for review or support from a human agent. For example, the customer could be applying to open a bank account and have to provide further information or clarification. Alternatively, it might cover processes where regulations mandate human involvement or simply where the customer wants the option to escalate.

Best for: Automating processes and complex journeys in regulated industries or for high-impact conversations.

Pros: Greater efficiency while ensuring compliance

Cons: Requires skilled, responsive agents available to step in instantly when required

4. Multi-Step Process

Customer journeys are made up of multiple individual steps, each of which may be handled best by humans or AI. Being able to switch between the two maximizes efficiency and saves time for the customer. This could be making an insurance claim or speaking to a bank agent after going through AI security and then handing back to the AI at the end of the call.

Best for: Journeys with clear steps, some of which can be easily automated.

Pros: Optimal use of resources, seamless experience for customer

Cons: Handoffs must be clear and instant with processes that match customer behavior

How to Design a Collaborative Workflow

When rolling out hybrid AI/human workflows, CX teams need to design it right, avoiding costly missteps while delivering results.

Following this design process maximizes the chances of success:

1. Start with the customer journey

Analyze your customer journeys. Which are generating the greatest number of interactions, taking most resources, or leading to complaints and churn? Look to see whether a combination of bots and humans will deliver greatest value, both to customers and to the business. Set clear metrics for measuring success.

2. Define clear handoff rules

When planning the hybrid journey focus on handoffs at specific parts of the journey. Monitor for triggers such as intent, sentiment, or data to switch between AI and humans. Apply your rules in an omnichannel way, matching AI and human agents across your different digital and voice channels.

3. Make your AI effective and trustworthy

Give AI agents the tools they need to answer queries. Alongside access to your knowledgebase, connect to CRM and other systems so they can pull in or update information. Put in place robust guardrails to prevent misuse, including escalation if required.

4. Empower agents with AI

Use AI to support agents with information, relevant knowledge and suggested next-best actions. This enables them to spend less time searching for answers, and more time building a relationship with the customer.

5. Track outcomes that matter

Don’t just focus on efficiency metrics such as resolution time. Instead, look at the wider picture, including agent satisfaction and customer sentiment. Hybrid journeys should improve results in all three areas.

6. Monitor and improve

Take a holistic view and monitor the performance of AI and human agents using the same tools. Supervisors should be able to see factors such as utilization across their extended team in order to allocate resources effectively. Use the same tools to drive improvements too. Analyze AI and human conversations to uncover where tweaks or extra training are required.

Showing the real-world results

Effective design delivers real results. One energy company interviewed by McKinsey reduced the volume of billing calls by around 20 percent and cut customer authentication time by 60 seconds by deploying an AI voice assistant at the front end of the customer journey ^[4].

Next Steps: Moving Toward Smarter CX Collaboration

When it comes to CX, there’s no reason to choose between people and AI virtual agents. The right blend, backed by well-mapped journeys, clear information, and shared visibility will deliver hybrid, collaborative CX that delights customers and increases efficiency.

Follow this checklist to deliver success:

1 Audit your customer journeys:

• Where do customers drop off or get frustrated?
• Where do agents spend time they don’t need to?
• What would a seamless experience look like, from both sides?
• Write up the job lists for the journey and use this to plan your workflow

2 Prototype hybrid journeys:

• Create outline workflows by involving the team and tracking user journeys
• Build a prototype based on defined needs
• Test and refine it, initially with a small but representative sample

3 Optimize over time:

• Customer needs and behaviors will change, so continually monitor performance
• Look at where you can optimize, based on customer and agent feedback
• Don’t be afraid to retire or replace entire workflows if needs change

Is Hybrid AI the Right Direction for Your Contact Center?

Enghouse Interactive has developed and deployed contact center technology for over 40 years, including AI since the early 2000s. Our focus is on practical customer service solutions that enhance customer and employee experiences and deliver measurable ROI for the business.

Explore more about how the right AI deployment can enhance your service delivery, with this Practical AI for Smarter CX eGuide.

Download the eGuide

Further Reading

• Video: Modernize Your Contact Center with Virtual Agent Integration
• Blog: How IT and CX Leaders Align on Modern Customer Experience Strategy
• Video: Evolve QA with AI: Stay Current and Keep Humans in Control
• Blog: 9 Signs Your Contact Center Tech Is Holding You Back — And How to Fix It

About the Author

Steve Nattress VP of Product Management
Enghouse Interactive
LinkedIn

Steve’s unique background in Customer Support and AI development positions him perfectly to drive product and AI innovation at Enghouse. Steve is a creative and strategic leader, passionate about the potential of AI to transform customer experience, while staying sharply attuned to the critical need for secure and transparent industry practices.

References

1. McKinsey and Company: The contact center crossroads: Finding the right mix of humans and AI
2. Gartner: Three Technologies That Will Transform Customer Service and Support By 2028
3. ContactBabel: AI for Digital Customer Contact
4. McKinsey and Company: The contact center crossroads: Finding the right mix of humans and AI