A recap of insights from our June 9, 2025 webinar "How to Actually Use AI in RevOps," featuring Kevin Heraly, VP of RevOps at Demand Science and Apoorva Verma, COO and Co-Founder at Rattle.
If you've attended any kind of RevOps gathering lately, you've probably heard a wide range of opinions about AI for RevOps. You may have heard that it’s taking over jobs, or that it doesn’t work well yet, or you’re being promised world domination. You might have heard things that made you feel like you’re behind where others are. And maybe you aren’t sure where to start.
That’s the purpose of this webinar series “HowRevOps Leaders Are Actually Using AI Today.” We want to provide a place where RevOps professionals can share what they’re doing, and hear what others are doing in detail.
During this session with Kevin Heraly, VP of RevOps at DemandScience, we heard his point-of-view: the most powerful implementations he uses are narrow, specific use cases that solve real problems his business faces.
The session was divided into two parts: a candid fireside chat about AI realities in RevOps, followed by Kevin walking through three specific use cases where he has successfully implemented Rattle AI at DemandScience.
Part 1: The Fireside Chat - Getting Real About AI
What's Working vs. What's Just Noise
Kevin opened with a refreshingly honest take: "What works best right now is when the use case is specific and somewhat narrow. When we're clear on what we're trying to accomplish and using our own data, we are getting great results!"
You've probably been bombarded with AI solutions promising to transform everything overnight. One webinar attendee captured the frustration with: "Today we are trying to sift through the AI pitches that seem designed to make RevOps feel guilty/not good at your job."
But here's where Kevin's experience is especially interesting. Instead of feeling overwhelmed by the hype, he has focused on use cases he can build out now, that deliver specific, measurable value to his team. No guilt in sight!
The Sweet Spot: Specific Use Cases + Human Creativity
As AI continues evolving rapidly, RevOps leaders are finding success when they can define specific use cases and then use AI tools like Rattle to explore outputs and refine them to reach their desired outcome.
While broader solutions certainly have their place in the evolving landscape, the current sweet spot appears to be implementations that combine human strategic thinking and creativity with AI capabilities. You conceive the use case, then work with AI to get the results.
Pro Tip: Rattle offers templates to help customers get started, but the real magic happens when you use your creativity to explore narrow use cases that support your specific business needs.
The Foundation: Trust + Narrow Focus
Kevin shared two ingredients for successful AI implementation with today’s toolsets and AI capabilities:
Building Trust from Day One
Nothing kills AI adoption faster than unreliable outputs. Kevin's team focuses on ensuring their AI use cases are only sharing results when criteria are truly met. No more "no results found" responses or keyword-based outputs that flood you with false positives.
He accomplishes this with prompt engineering phrases like "only return results if..." combined with human oversight during the testing phase.
Keep It Narrow, Keep It Valuable
Instead of trying to automate every process at once, Kevin identifies very specific use cases that, if solved, would move the business forward. Then he uses Rattle's AI to develop and test prompts against his business's own GTM data - recorded calls, logged emails, CRM records, and Slack channels.
Once he's confident in the output, he creates automations that feed intelligence directly into Slack where his team can take immediate action.
Part 2: Three Live Use Cases at Demand Science
Kevin walked through three specific Rattle AI implementations his team uses daily, complete with prompts, outputs, and measurable results.
1. Macroeconomic Risk Detection
The Challenge: Leadership needed visibility into conversation signals around tariffs, hiring freezes, and budget pauses—without manually combing through hundreds of call transcripts each week.
Kevin's Solution: Using Rattle's AI toolset, Kevin built a custom prompt that "only returns output if" true economic-impact context is found. When the AI detects a qualifying snippet, it pushes a formatted alert into a dedicated Slack channel—complete with the customer quote, a brief synopsis, and recommended next steps.
Why It Works:
- Context over keywords: By focusing on meaning rather than string matching, false positives plummet
- Trust-building filters: The "only return output if..." guardrails ensure every notification is actionable
- Real-time leadership insights: Executives no longer learn about external risks weeks after the fact
2. Early Channel & Partner Engagement
The Challenge: DemandScience's channel team was often looped in too late, missing prime opportunities to co-sell.
Kevin's Solution: Kevin extended the same AI-listening approach to detect genuine channel-fit signals in live calls, emails, and chat. Once the AI confirms relevance, it delivers a SPICED breakdown to the channel Slack channel, along with a "next best step" template for the AE.
Why It Works:
- Proactive alerts: Rather than waiting for sales reps to remember tagging partners, the AI flags them immediately
- Human-in-the-loop: Summaries empower the channel team to decide when and how to engage, preserving trust
- Measurable lift: Kevin reports "a couple more partner-sourced or partner-influenced leads every month" since launch
3. Automated Sales-Methodology Tracking
The Challenge: Capturing consistent SPICED or MEDDPICC qualifiers in Salesforce without overburdening reps.
Kevin's Solution: Kevin's team built AI prompts that extract qualification criteria from unstructured data (from record calls, logged emails, Slack conversations) and append them—formatted as rich-text HTML—into Salesforce opportunity fields. Crucially, the prompt reads the existing field value and only appends new information, preventing duplication or data loss.
Why It Works:
- Easy consumption: Bold headings and bullet points make deal reviews lightning-fast
- Minimal rep lift: Reps simply continue their calls; the AI handles the notes
- Clean history: Appending rather than overwriting preserves chronology and context
The Path Forward
The RevOps leaders winning with AI today are embracing their creativity and strategic strengths to. They're identifying specific pain points, building focused solutions, and iterating based on results.
Kevin's advice captures it perfectly: "Done is way better than perfect. Pick the low-hanging fruit, pick something that's low effort but high impact that'll start building trust right away."
AI is already transforming how RevOps teams operate. The distinction lies in whether you'll be among the leaders building practical solutions that deliver measurable value, or waiting for the perfect comprehensive solution.
Ready to get started? The best AI implementation is the one you actually use. Pick one specific use case that solves a real problem for your team, build trust through reliable outputs, then expand from there. We’d love to help you figure it out!
Want to explore how AI can work for your RevOps team? Connect with us to see how Rattle's AI capabilities can help you surface the insights hiding in your unstructured data while maintaining the focused, practical approach that drives real results.
We'll be hosting webinars like this every 4-6 weeks. Follow Rattle on LinkedIn to see when our next event is happening. https://www.linkedin.com/company/gorattle/
.png)