Beyond the Hype: How to Actually Unlock AI Agents for Customer Engagement

There is a lot of hype – and a lot of fear – about what AI agents mean for the future of business. Hype-mongering headlines — like “ 2025: Agentic And Physical AI — A Multitrillion Dollar Economy Emerges” — abound. So do fear-mongering headlines like “ Tech is now betting on AI ‘agents.’ Will they boost stocks or take your job?”.

So which is it? Are AI agents friend or foe?

Spoiler: here at Twilio we see AI agents as a means to unlock value for your customers, your business, and your employees. The key to doing it right is to stay focused on the customer experience.

This next generation of AI agents goes far beyond the previous generation of brittle chatbots that are costly to maintain and result in clunky customer experiences. Built on top of powerful LLMs, AI agents can engage in human-like dialogue and perform actions on behalf of a business or customer, taking us far beyond basic FAQ response. Even better, when AI agents can engage with customers on the channels they prefer and leverage contextual data, you get a rich and dynamic customer experience that businesses haven’t been able to imagine or scale previously.

Creating excellent customer experiences is one of the most important things a business can do, so it’s important to be thoughtful before incorporating AI agents into the customer journey. Building Twilio AI Assistants – an opinionated framework to build and host conversational AI Assistants for your customer-facing use cases – has given us a front row seat with startups and Fortune 500 companies into what matters most for brands building AI agents.

Let’s talk about how to actually craft AI agents that work for you, helping you to better serve your customers while driving new value to your business at a scale we haven’t seen before.

How to get started: AI agents for customer-facing use cases

Where is there friction in your customer journey today?

Before diving into the mechanics of configuring an AI agent, your first stop is to assess your current customer journey and be brutally honest about what’s not going well.

What’s hard for your customers to do today? Do they have to call support in order to make a change to their order? Is your top of funnel strong, but your conversion rates are falling short because customers hit onboarding hurdles?

What’s hard for your employees to do today? Is your contact center overloaded with demand, resulting in both long hold times for customers and burnout for your employees? Are you offering enough channels for customers to contact you? Are you surfacing the right information to customers at the right time, or leaving customers on their own to navigate your product and processes? Are you responding to every lead and prospect with a thoughtful response, or leaving some leads unresponded to?

Typically metrics like low NPS scores, long average hold times, leaky funnels, and low conversion rates are symptoms of a broken customer journey. Once you’ve identified the metrics you want to improve, and the problem areas causing those metrics to falter, then you can assess whether or not inserting an AI agent could streamline those parts of the customer journey.

Where are you sending 1-way communications today?

Another great indicator for potential AI agent opportunities is to assess your 1-way communications. 1-way alerts and notifications are a great way to stay connected with customers, but they’re limited. Providing proactive alerts and notifications helps keep customers informed, but oftentimes companies never supply a means for the customer to seamlessly engage with your brand after receiving a notification. AI agents can now unlock 2-way dialogue.

In the below example, the customer can easily converse with an airlines’ AI agent to upgrade their flight. By configuring a capable AI agent to respond to what was previously a 1-way alert, the business created a better experience for the customer, mitigated potential call center volume, and unlocked revenue that may never have come to pass if it was a more cumbersome experience.

Consider where you can drive value with 2-way communications that deliver a richer customer experience AND business outcomes. This is now scalable with AI agents.

Key considerations for customer-facing AI agents

AI agents are only as good as their design and configuration. At Twilio, we believe agents shouldn’t be haphazardly inserted, but instead thoughtfully embedded into a complete journey. Once you’ve identified the use cases which could use AI agent touchpoints, it’s time to map out the complete customer journey. Below are key considerations to ensure AI agents become value-drivers to your business…not detractors.

Escalation paths

Not everything should be solved by an AI agent (at least when this article went to press!). You should take a hybrid approach that allows customers to engage with AI agents and human employees. Also note that triggering an escalation from an AI agent to a human agent doesn’t necessarily mean something has gone awry. Below are some categories to consider when designing intelligent routing for escalation paths from an AI agent to a human agent.

• Sensitive workloads: you may have specific use cases in your business that you always want handled by a human, e.g., prospects who have highly regulated use cases that should always be routed to a specific team
• Low sentiment: if you incorporate sentiment analysis for AI agent conversations you should consider setting a sentiment threshold that triggers an escalation to a human.
• Customer segmentation and intelligent routing: depending on how you segment your customer base, you may want to always route specific customer segments (e.g., Gold status members) to bypass AI agents and speak directly to workers with specific skills.

Twilio's TaskRouter is a skills-based routing system that allows you to control routing from your code to fit your unique business needs. You can also check out this guide for how to transition a conversation with your AI Assistant to a Flex contact center agent. Finally, check out Flex Agent Copilot to help guide your human agents to better understand your customers and simplify and expedite the resolution process.

Don’t miss out on your data - it’s your differentiator

Before you deploy an AI agent, consider the bi-directional flow of data through that agent.

AI agents who know nothing about the customers they’re talking to will deliver a clunky customer experience and take much longer to resolve an issue. How can an AI agent answer a customer’s question about their order if they don’t know who that customer is?

Customer Memory is a feature in AI Assistants that allows agents to become customer-aware. By integrating with Twilio Segment, the Assistant can access and store relevant customer data, enabling it to personalize interactions based on previous engagements and customer profiles. Two key engines power this capability:

• AI Personalization Engine: This engine pulls relevant customer data from Segment, allowing the Assistant to tailor its responses. Simply put, this allows the AI agent to know who they’re talking to.
• AI Perception Engine: This engine captures relevant data from ongoing conversations and updates the customer profile in Segment, evolving the Assistant's understanding of the customer over time.

You wouldn’t have your employees walk into customer conversations blind, so don’t let your AI agents do it either. And as AI agents increasingly handle more conversations with customers, make sure you’re securely capturing data from those conversations so you can deliver your customers increasingly more personalized and delightful experiences moving forward.

Embracing the future of AI agents

Every brand worth its salt wants to drive more meaningful engagements with their customers. And the ones that will last also want to do more with less. AI agents finally unlock the ability for businesses to engage with their customers in a meaningful and personalized way that doesn’t require the linear scaling of human resources.

Now is the time to start thinking about how to insert AI agents into your customer journey that will drive value for your business and your customers. Start with low-risk use cases, using AI tools to automate repetitive tasks, gain a better understanding of your customer data, and try building your first AI agent using a framework like Twilio's AI Assistant platform.

Kat McCormick Sweeney leads business development for the ETI team. She gets pumped about solving customer problems with her team. When she’s not thinking about customer engagement (it happens), she’s exploring the wonders of Northern California with her family, has her nose in a good book, or is watching Barefoot Contessa reruns. Connect with Kat at kmsweeney [at] twilio.com.