Everyone in the sales AI space is racing to develop their own AI agents. But ask any seller what AI tool they use every day and 9/10 of them will tell you something like ChatGPT, Claude, Gemini, or Microsoft Copilot - and that isn’t likely to change anytime soon. In fact, what is already changing is that they won’t need to go anywhere else.
That’s because their day-to-day agent will work directly with expert agents to go beyond giving answers prompt by prompt to create their own prompts and skip straight to what sellers need to go do to reach their goals. We see this emerging with new innovations such as MCP (Model Context Protocol) and A2A (Agent To Agent Protocol) that equip AI agents to securely share data and context autonomously, giving sellers all the data and context they need from an “Expert Agent” (like People.ai) inside the tool of their choice.
It’s inspiring to us because our goal has always been to get sellers out of systems and out to customers, and now the day is here where GTM teams will spend a lot less time in tabs, forms and spreadsheets. But GTM teams will need to rethink what sellers will want in their daily experience in a very fundamental way, one that adapts to how we are all already changing how we interact with information.
General agents are winning the attention game
Here's what's happening right now: ChatGPT, Claude, Gemini, and Microsoft Copilot have become the AI interfaces people open first when they log into their laptops (we’ll call them “General Agents”). Your sellers are already using them.
Soon, these agents won't just complete tasks and regurgitate publicly-available information. They'll be able to connect with Expert Agents who have decades of data about how sales teams work so they can also:
- Determine what tasks need to be done to meet your objectives
- Work with specialized Expert Agents to get those tasks completed
- Operate across your entire tech stack on behalf of your team
This isn't theoretical. It's happening now.
As an example, with our MCP integration, we prompted Claude to build an updated forecast on behalf of a sales person. Claude worked directly with our SalesAI product to build an informed, accurate plan in a way that was inspiring to all of us. You can see the demo here:
What MCP and A2A mean for your revenue team
MCP (Model Context Protocol), created by Anthropic, standardizes the interaction between large language model (LLM)-based agents and external tools or APIs.
A2A (Agent-to-Agent Protocol), created by Google, lets these agents communicate and delegate tasks across different platforms.
Together, they create something powerful: agents that can work as users, not just tools.
At People.ai, we’ve been working on new MCP and A2A capabilities that will enable agent to agent communication, and what I'm seeing tells me the future is arriving much faster than most sales leaders realize. The implications for how your teams work—and where they spend their time—are profound.
If you're like most revenue leaders, your sellers are still buried in forecast calls and spreadsheets when they should be in front of customers. That's about to change.
Your First Agent will be able to ask Expert Agents to pull forecast data, analyze pipeline health, and recommend actions—all without human intervention. It will move from handling tasks and information that you request to fulfill a goal to working out which ones are needed on its own. And once it can, your GTM teams won’t need to have a bunch of tabs open to applications - they’ll just work through their agent of choice.
Why fragmented experiences won't win
Every vendor is racing to build agentic capabilities into their own platforms. They want to own your user experience.
But here's the reality:
- Your team already has preferred AI agents
- Context gets lost when jumping between systems
- Users won't log into multiple specialized copilots
The agents your people already trust will become their primary interface. These agents will reach into your CRM, forecasting tools, and other systems to do the work.
The challenge: Context and governance
General Agents are getting remarkably good, but they lack crucial context:
- How your data sources connect
- What each user should see and access
- What's relevant to specific roles and tasks
- Your company's unique processes and definitions
- Plus, all of the function-specific data that expert tools like People.ai provide
This is where Expert Agents come in. They prepare and govern data for General Agents, ensuring the right information reaches the right people at the right time.
After ten years and billions of transactions, we've learned these are complex problems that require purpose-built solutions.
The evolution: From human-led to agent-shaped
Today: Inefficient human-led processes with AI assistance
Tomorrow: Human-managed and agent-led tasks and processes
Soon: Agents actively shaping work and information flow, freeing humans to do strategic and valuable work with other humans.
This matches our core belief: people should work with people, while AI handles everything else.
What this means for sales teams
Sales forecasting is the perfect example of where this is heading. Today, it's human-led with manual processes, endless spreadsheets, and weekly forecast calls that pull your best sellers away from customers.
Soon, agents acting on behalf of sellers and sales leaders will:
- Automatically gather and analyze pipeline data
- Identify risks and opportunities in real-time
- Generate accurate forecasts without human intervention
- Surface insights that actually matter for decision-making
Your sellers get their time back. Your customers get more attention. Your forecasts get more accurate.
Our path forward
We're announcing our MCP A2A capabilities soon, and you'll see us integrating these protocols across our solutions. Sales Forecasting will be our first showcase of how agents can give your revenue team time back to focus on what matters most: customers.
This isn't about replacing your sellers. It's about giving them superpowers by removing the administrative burden that keeps them from selling.
The future where people work with people while AI does the rest is coming in months, not years.
Ready to see it in action? Request a demo.