Microsoft launches genAI sales agents, also focuses on finance and supply chain

Microsoft is planning a string of new AI agents to automate mundane tasks and improve worker productivity.

The company on Wednesday announced two AI-based sales agents that can autonomously follow up on sales leads. The company is also focusing on using agents to automate workflows across other sectors.

The sales agents, which are available in Microsoft 365 Copilot, connect to data in Microsoft 365 and Salesforce and Dynamics CRM systems to automatically generate and communicate with sales leads.

Salespeople spend hours verifying leads, writing to customers and then waiting for responses, said Bryan Goode, corporate vice president for business applications and platforms at Microsoft. “If you can take something that used to take hours and do it in minutes, you can spend more time selling,” he said.

One agent, called “sales chat,” can automatically create documents by drawing data from CRM systems, pitch decks, meetings or emails. The second agent, called “sales agent,” is an autonomous agent that works continuously converting contacts a company might have into leads and opportunities.

The sales agents are “finished agents” built on Microsoft capabilities and a variety of large language models (LLMs) that customers can start using right away. The LLMs used depend on customer needs, Goode said.

There is a provision for humans to review emails and documents going out, Goode said.

Microsoft’s Copilot Studio also allows customers to build, customize, and manage the fleet of agents.

The company sees opportunities for agentic AI in the finance, supply chain and other customer service areas, Goode said. “Those are the three functional areas in particular where we’re pretty, pretty heavily focused,” he said.

He cited, as an example, customers trying to navigate through complicated supply chains to reduce inventory and lower costs.

Microsoft also announced an AI Accelerator for Sales program, in which it will help customers migrate from old CRM systems to “AI oriented selling,” Goode said.

AI agents are creating an inflection point for traditional CRM systems, which could go the way of a mainframe, Goode said. “It’s going to be something that might be around, but probably has some dust on it,” he said.

The Microsoft announcement comes amid a flurry of agentic AI announcements this week. Amazon formed an agentic AI business group, according to Reuters.  And Salesforce at its TDX tradeshow announced AgentForce 2dx, which has new tools to allow AI agents to do better reasoning. 

AI agents can help worker productivity, but customers should approach them with caution, said Charles King, principal analyst at Pund-IT. As with early days of LLMs, AI agents are still in their infancy and face many challenges, he said.

“If the data collection is done legitimately and done with the idea of delivering truly valuable results, it’s great. But there are many examples of AI tools that are not reliable,” King said.

King pointed to OpenAI’s ChatGPT, which in its early days hallucinated and generated inconsistent results. Agents are expected to mature and generate more reliable results, but human intervention and guardrails against hallucination will be needed, King said.

Companies today use large LLMs for general-purpose querying, but personalized agents tailored for business processes are the future, King said.