Improve Service with AI Agents
Learning Objectives
After completing this unit, you’ll be able to:
- List sample order support actions.
- Summarize how topics, actions, and instructions help an AI agent meet customer needs.
- Explain what happens behind the scenes when a customer interacts with an AI agent.
Help Customers with Order Support Actions
Ursa Major Solar wants to add AI agents to the self-service portal on its website. Just as service reps can take action from their Service Consoles, AI agents can take actions based on customer interactions. Unlike chatbots and Copilots, AI agents review real-time data and perform the next logical step. To keep the AI agents on track, you can set up guardrails and conditions. If necessary, the AI agent can also hand off a case to a service rep with a summary of the interaction, customer details, and recommendations on the next steps. Although Ursa Major is starting with their website, AI agents can help out from anywhere, including via SMS, WhatsApp, or an online store. AI agents can perform several actions.
- Status: AI agents can report on the status of customer orders by accessing information in the order summary including when their order shipped, what items are in the order, payments status, any changes made to the order, and more.
- Cancel: AI agents can cancel orders at any point in the fulfillment cycle. Guardrails can be created to limit certain cancellations or refer the case to a service rep in specified situations (for example,if it’s above a certain dollar amount or at a certain stage of the fulfillment cycle).
- Modify: AI agents can change existing orders, such as updating a shipping address or changing payment details.
- Return: AI agents can perform simple or complex returns, from finding the order to creating return merchandise authorizations.
- Exchange: AI agents can perform basic exchanges, such as returning an item for one of a different color or size. They can also process uneven exchanges, such as where a customer exchanges a red shirt for two shirts and the AI agent then changes the payment amount.
In addition to performing these actions, AI agents can turn a customer interaction into a revenue-driving experience by recommending similar or complementary products. For example, the AI agent can offer a customer returning a solar panel a different model that’s more efficient or fits better on their roof. It can also recommend adding a solar water heater to the order.
Guide Customers Through Product Exchanges
Ursa Major receives a notification that an entire line of solar panels is being recalled because of a manufacturing defect. ‌The Ursa Major team sent out a notification to all affected customers with steps to follow to exchange these panels for similar ones. They also posted those instructions on the Ursa Major website’s self-service portal. Even so, the team knows that some customers will need additional help. This is a good opportunity to set up an AI agent to handle the easy returns, leaving its service reps to provide a personal touch if needed.
The service team asks their Salesforce admin, Maria Jimenez, about the next steps. Because Ursa Major uses Agentforce for Commerce Cloud, it’s already set up. The team asks Maria exactly how the AI agent works behind the scenes. She explains that the AI agent uses its internal Atlas Reasoning Engine to understand natural language. After it determines what a customer wants, it selects predefined topics that tell it how to handle the customer interaction. The topics contain actions the agent can perform and instructions which tell the agent how to make decisions. Together, they give the agent the context and boundaries needed to get the job done. Then, she shows them a video of a sample customer interaction.
As a next step, Maria asks the service team for a list of actions the AI agent needs to take to help with the recall. The list is as follows.
- Retrieve product details, like customer name and address, and product to be returned.
- Verify that the recall covers the product to be returned.
- Start the existing exchange process flow, including scheduling appointments for a field service technician to uninstall the recalled solar panels and install new ones.
- Update all relevant records.
Maria creates a Recall Exchange topic that contains all the actions listed plus the required context. She uses standard actions for things like looking up customer information and updating records. She also creates custom actions to ensure that the product is eligible for the recall. Where necessary, she connects to existing flows, like setting up field service appointments. After she’s completed and tested the topic, she activates it and puts it on Ursa Major’s website.
Example of an AI Agent Processing an Exchange
The AI agent doesn’t have to wait long for its first task. A customer named Cyrus receives the email about the product recall and goes to Ursa Major’s website to ask the AI agent for help. Click to see what’s happening behind the scenes.
If the customer requests it or if the AI agent needs help resolving the issue, the AI agent creates a case and transfers the customer to a service rep. The service rep sees the exchange between the AI agent and the customer and has the proper context to get started. That wasn’t necessary in this case: Cyrus and the AI agent have sorted everything out and a service technician is on the way to uninstall the old panels and install the new ones.
In this module, you discovered that combining Service Cloud with Commerce Cloud gives you a unified view of your customers and allows service reps to help out more effectively. Then, you stepped through a sample scenario where an AI agent helped a customer ‌return a recalled solar panel. Are you ready to integrate Service Cloud, Commerce Cloud, and Agentforce to enhance your customers’ satisfaction and simplify your service operations like Ursa Major Solar?
Note: AI supported the writers who created this content.