AI Agents for retail see how AI Agents can support your store and frontline teams
Introducing Commerce Companion
Sunrise built Commerce Companion to support frontline retail employees with the kinds of questions that come up constantly in stores. Rather than forcing associates to search through multiple systems or interrupt coworkers for help, Commerce Companion brings together a collection of intelligent agents that can respond in natural language with information pulled from Dynamics 365 and internal knowledge sources.
The goal is straightforward: help employees get fast, accurate answers in the flow of work. Commerce Companion can answer questions like the price of an item, a customer’s purchase history, available inventory, or when the next transfer order is expected. It can also surface internal procedures and store documentation, giving employees another channel for training, onboarding, and daily decision-making.
Because Commerce Companion is built on an orchestration model, each subagent has a clear purpose and connects to the right data source for the task at hand. When deployed with Copilot Studio, it can be delivered directly in Microsoft Teams, which makes it easier for employees to access without changing how they already work.
Customer Stories: AI Agents in Action
Jaipur Living and Ordie
One Sunrise customer who is seeing success with AI Agents is Jaipur Living, a designer and distributor of hand-knotted rugs that sells across eCommerce, retail, wholesale, and B2B channels. After completing its initial Dynamics 365 Finance and Supply Chain Management project, Jaipur Living looked for new ways to build on its Microsoft investment and improve customer service operations.
The result was a sales order agent the team nicknamed Ordie. Ordie helps answer common order-related questions such as order status, tracking information, order holds, cancellations, and open amounts. Instead of relying on customer service teams to respond manually to every request, sales reps can use Ordie in Microsoft Teams to get quick answers while they are on the go or preparing for customer conversations.
According to the Jaipur Living team, one of the biggest benefits is speed. Order-status questions can take up a large share of customer service time, and Ordie gives team members direct access to real-time Dynamics 365 data from a mobile-friendly interface. The company also highlighted how quickly the solution was implemented and how easy it has been to refine based on feedback. That combination of faster answers, easier access to data, and quicker iteration is exactly what many retailers are looking for as they explore AI agents.
Licensed Sports Apparel Company
The second customer story focused on a global licensed sports apparel company that has continued to expand its Microsoft footprint as its business has grown. In this case, the organization deployed agentic solutions to support finance and accounts payable teams.
One solution was an Invoice Capture agent designed to step in when invoice processing errors occur. Instead of making AP staff investigate every issue manually, the agent interprets the error, researches the problem, gathers vendor details from Dynamics 365, and drafts outreach to request a corrected invoice. The draft can then be reviewed and refined by the AP team before it is sent.
The company also implemented a multi-agent accounts payable assistant that helps answer common vendor and invoice questions. Each subagent has a specific responsibility, and together they provide real-time responses on invoice status, pending invoices, vendor contacts, and more. One especially useful feature is a Deep Link agent that gives employees a direct URL to the relevant Dynamics 365 record, cutting down on clicks and search time.
These examples show that AI agents are not limited to customer-facing retail scenarios. They can also improve the efficiency of back-office teams by reducing manual work, speeding up routine processes, and keeping employees focused on higher-value tasks.
Recommendations for Getting Started with AI Agents
Here are four practical recommendations for organizations that want to build a stronger foundation for AI:
Build an accessible data foundation
Clean, structured, and accessible data is essential. This is often the biggest barrier to entry for organizations interested in AI. When data is well-organized and business processes are already understood inside the system, it becomes much easier to deploy agentic solutions quickly and effectively.
Identify structured, repetitive work
A good place to start is with work that happens every day, follows a predictable pattern, and creates measurable value when improved. These are the types of tasks where AI can make an immediate difference, whether than means saving time, reducing manual errors, or improving response speed.
Identify the right task types
AI is especially useful for summarizing, drafting, and reasoning over existing information. It is less suited for work that depends on rigid calculations, strict compliance logic, or hard-coded rules. In those cases, the best approach is often to keep the calculation or rules-based logic outside the AI layer, then bring the result back for the agent to interpret or explain.
Identify areas where AI can be used without causing harm
Any injection of AI into business processes must be done thoughtfully. Organizations should focus on areas where AI can support employees, reduce friction, and improve workflow without introducing risk. That starts with understanding where people lose time across their day-to-day tools and identifying opportunities to make work easier and more consistent.
Guidance for Deploying your First AI Agent
For organizations preparing to launch their first AI agent, here are four practical tips:
- Identify areas of opportunity: Start by looking department by department, or across cross-functional workflows, for repetitive processes with high impact. Once you find them, try to quantify the opportunity so teams can clearly understand why the use case is worth pursuing.
- Start small: While it’s tempting to go big with your first AI agent, it’s better to not chase a home run right away. Start with a focused use case, prove the value, gain buy-in, and expand from there. A series of smaller wins can build confidence and momentum across the organization.
- Learn and experiment with new tools: Retailers should stay curious and continue learning, but they do not need to rebuild everything every time a new model or tool is released. Technology partners can help guide these decisions. Once a working solution is in place, organizations can revisit and optimize it on a thoughtful cadence rather than constantly replacing it.
- Don’t over-solution your AI agent: An AI agent does not need to solve every scenario perfectly to be valuable. A solution that handles 80% of the work and leaves the remaining 20% for human review can still deliver meaningful results. Trying to cover every edge case from the start can slow progress and reduce adoption. The better approach is to keep the end goal in mind, focus on practical impact, and make the solution easy for employees to use in the tools they already rely on.
Get started with Commerce Companion today
Commerce Companion helps you deliver AI-powered customer service and back-office automation at the speed of today’s retail operations.
Choose the right team
