How to Use AI Agents for Commerce

Think of the most labor intensive, time-consuming parts of your job. Now imagine that you had a trustworthy, capable assistant who could do the grunt work for you — quickly and accurately. That’s what autonomous AI agents can do. For commerce teams, this means merchandisers won’t need to spend hours sifting through data to uncover trends and insights to build relevant offers and product bundles. Service agents won’t be buried under endless queries about order status or delivery timeframes. Instead, AI agents can take the wheel to move business forward at scale, so your teams can focus on more strategic, high value work. Here’s how commerce teams can use AI agents to increase sales, productivity, and revenue.

You can think of Agentforce as a limitless extension of your workforce — available 24/7 and capable of massive scale. With Agentforce, your teams can quickly set up autonomous AI agents and work side-by-side with them to complete all kinds of tasks, from the mundane to the most complex. Using fillable forms and natural language, it’s simple for any business user to build an agent, assign them specific actions, and give them the right context to perform the job. Some agent actions come out-of-the box, and you can also build custom agents to fit your own unique business needs. Let’s take a look at the set up process and explore a few examples of how agents can help increase revenue, productivity, and customer satisfaction.

Setting up an AI agent with Agentforce is designed to be user-friendly and straightforward, even for users without extensive technical expertise. All you need to do is simply describe the project and any desired outcomes — no technical knowledge required! The process includes simple form fields and plain language instructions to ensure streamlined use. Here’s a look a t the steps involved.

1. Select or create an agent:

You can either choose a prebuilt agent or create a custom one using Agent Builder. For routine commerce tasks (ranging from creating a product bundle to managing inventory), select a pre-built agent. If you have a more specific business need that isn’t covered by a prebuilt agent, you can create a custom agent. The first step is to decide which type of agent you will use.

2. Define agent roles and data:

Next, assign and name your agent with its specific function using simple form fields. If you’re creating a custom agent, enter agent instructions in plain language based on data from back-end business systems, like Data 360, your CRM, ecommerce platform, and more. For example, if your sales cycle involves a series of approvals from different stakeholders, you can create actions to alert each team member when it’s their turn for review.

3. Set up guardrails:

Use plain language to give your agent rules to follow, ensuring compliance and secure use of customer data. For example, a merchandiser might add this guardrail to agent instructions: “Ensure product recommendations are always in line with inventory levels. Never, ever promote an item that’s out of stock or in limited supply. Don’t offer any promotion that goes below minimum margin levels.”

4. Add source materials:

Upload or select relevant data from your CRM, such as product documents, FAQs, and industry one-sheeters.

5. Test, deploy, and go live:

Verify all inputs and confirm that the agent outputs are accurate. (This is what we call having a “human in the loop.) Test the agent or activate it to go live.