Agentic AI Needs More Than Automation: Mollie Bodensteiner’s RevOps Playbook for Execution

From AI Experimentation to Revenue Execution: Building Governed AI Workflows

AI is quickly becoming part of how revenue teams work. Sellers are experimenting with AI agents, RevOps teams are automating processes, and organizations are testing new ways to use buyer and pipeline data.

But as AI adoption grows, a new challenge is emerging: How do Revenue Leaders decide what should become part of the core revenue process, and what should remain an experiment?

On this episode of Revenue Architects, Steve and Jay speak with Mollie Bodensteiner, VP of Revenue Operations at ZoomInfo, about building RevOps around execution, using AI to improve forecasting and workflows, and creating the governance needed to scale AI responsibly.

Here are the key ideas from the conversation.

AI Adoption Starts With Experimentation

Mollie’s approach to AI starts with giving people room to experiment.

When employees get access to AI, they naturally begin creating their own workflows to solve everyday problems. One seller may build an agent for pre-demo research, another may create a post-demo workflow, and someone else may develop a useful prompt for customer follow-up.

That experimentation is valuable.

The problem starts when every individual workflow becomes part of the organization's operating model.

If ten sellers have ten different ways of doing the same thing, the next step for RevOps is to look across those experiments, identify what works, and bring the strongest ideas into a consistent, governed workflow.

The goal isn't to eliminate experimentation. It is to turn useful experimentation into repeatable execution.

Not Everything That Can Be Automated Should Be Automated

One of the recurring ideas in Mollie's approach to AI is simple:

A useful way to think about this is across three levels:

1. Human-led with AI support
AI can do much of the work while a person makes the final decision.
For example, AI can draft a follow-up email based on a customer conversation, but the salesperson reviews and sends it.

2. Fully autonomous
Some activities are predictable and low-risk enough to run without human intervention.
A meeting reminder, for example, doesn't necessarily require a salesperson to review it before it is sent.

3. Experimental
Other AI workflows may be useful but haven't yet earned a place in the core process.
These should be tested, measured, and refined before becoming standardized.
The important question isn't simply "Can AI do this?"
It is "What level of AI involvement makes sense for this part of the revenue process?"

AI Can Give Forecasting a Better View of Deal Health

The conversation also explored how AI can improve one of the most important responsibilities in RevOps: forecasting.

Traditional forecasting often relies heavily on stage-based probabilities and information entered by sales reps. Those inputs are important, but they don't always tell the complete story.

Mollie's approach looks at additional signals around authenticity and engagement.

Authenticity can include signals such as:

Engagement looks at how the buying group behaves over time:

These signals can help RevOps understand what is actually happening inside an opportunity rather than relying only on what has been entered into the CRM.

Context Matters More Than a Simple Engagement Score

An important nuance is that AI needs context.

Suppose a prospect suddenly stops responding. On the surface, that could look like declining engagement.

But perhaps the prospect mentioned during a call that they would be attending a major industry conference for two weeks.

AI can potentially identify that information from call recordings, emails, or other interactions and provide context around the delay.

That changes the conversation from:

"Why isn't this deal moving?"

to:

"We know why this deal is delayed, and the close date should be adjusted accordingly"

For Revenue Leaders, this kind of contextual intelligence can make forecasting more useful because it helps explain why the pipeline is changing.

RevOps Has to Build From the Field

Another strong theme from the conversation was the importance of understanding how sales teams actually work.

Mollie believes RevOps should build from the field rather than designing processes from a whiteboard.

That means:

This matters because RevOps decisions can directly affect how salespeople work, and ultimately their earning potential.

A process might look efficient on paper but create unnecessary friction in the real world.

The field can reveal those problems much faster.

For Mollie, this also comes down to credibility. If RevOps expects sellers to adopt a new process, the team needs to understand and experience that process themselves.

RevOps and Finance Should Work as Partners

Revenue execution also doesn't happen inside RevOps alone.

Mollie describes Finance as one of RevOps' most important partners because Finance understands the financial model behind the business.

One idea from the conversation stood out:

Not every dollar means the same thing to the business.

Understanding which revenue matters most can influence decisions around:

This requires RevOps to understand Finance's assumptions and models rather than simply receiving financial targets.

The same principle extends to Enablement and HR.

If RevOps sees a pipeline issue, HR sees high attrition, Finance is thinking about headcount and productivity, and Enablement is responsible for ramping new hires, those issues are connected.

Working together allows those functions to solve the larger revenue problem rather than addressing each piece independently.

Strategy Needs a Path to Execution

Mollie also challenges the traditional separation between go-to-market strategy and execution.

Strategy can provide the direction, market insights, and business intelligence. But without execution, it remains a plan.

The closer strategy, operations, systems, and enablement work together, the fewer handoffs there are between deciding what needs to happen and actually making it happen.

For revenue organizations, this means turning:

Insight → Decision → Process → Execution → Learning

into a continuous cycle.

Core AI Workflows Should Be Governed

As AI becomes part of revenue processes, governance becomes increasingly important.

If every seller has their own agents running in different places, the organization may eventually lose track of:

That is where RevOps has an important role.

Core AI workflows need ownership, monitoring, and clear rules. They should also have a rollback or "kill switch" when something goes wrong.

The more directly an AI workflow affects customers, sellers, or revenue, the more important that governance becomes.

AI Workflows Should Be Tested, Not Set and Forgotten

Even once an AI workflow is working, the job isn't finished.

Revenue processes evolve. Data changes. Messaging changes. Sales teams learn new approaches. Customers behave differently.

That means AI workflows need to evolve too.

Mollie recommends treating AI workflows like ongoing experiments. If a seller discovers a better approach, test it against the existing approach. If it performs better, consider bringing it into the core workflow.

This creates a continuous improvement cycle rather than treating AI as a one-time implementation.

Start Small Instead of Trying to Automate Everything

One of the most practical lessons from the episode is to avoid trying to solve the entire revenue lifecycle at once.

Instead:

The same applies when data quality isn't perfect.

You don't necessarily need to fix every data problem in the organization before launching an AI use case. Fix what is necessary for the specific workflow, learn from it, and continue improving.

The Goal Isn't More AI. It's Better Revenue Execution.

The conversation ultimately comes back to execution.

Mollie's advice to RevOps leaders is straightforward: have a plan and work your plan.

There will always be more requests, ideas, tools, and problems than a RevOps team can address at once. The challenge is identifying the few things that matter most and focusing on moving those forward.

That means:

AI gives Revenue Leaders new capabilities. But the real advantage comes from knowing where those capabilities belong, how they should be governed, and how to turn them into repeatable revenue workflows.

The next phase of RevOps isn't about giving every seller another AI agent.

It is about creating a connected operating model where data, buyer signals, human judgment, AI, and execution work together, with enough governance to make those capabilities reliable at scale.