Hi! I hope all is well with you and your loved ones, and you’re having a great summer so far! ☀️
I’ve been playing with AI Agents and wanted to share my experience, as well as a best practices template. I’m also going to paste the most popular templates so far this year below, for those who haven’t seen them yet or who are new subscribers!
Before we dive into the main content, please check out a quick word from our sponsors, who are kind enough to support the creation of this free-to-you MOPs content 💖:
Wondering if you’re keeping up with your peers in ops? Default just surveyed 300 RevOps teams on how they’re currently using AI. The top insight: almost everyone’s using it 👍 …almost no one’s seeing results. 🫠 Learn why that is (and other interesting insights!) here.
Teams are leaner...expectations are higher...and RevOps/MOPs folks are stuck in the middle. Instead of logging into HubSpot every day and manually checking reports, I’ve got a Zapier AI Agent doing that for me. Check out Zapier AI Agent templates and try them for free here.
Okay, now my guide on AI Agents! 👇
Tl;dr, the key takeaways I want you to focus on:
AI agents are structured systems that execute goals with tools, memory, and logic, not just chat interfaces.
Model Context Protocols (MCPs) define agent behavior and guardrails, improving consistency and reducing hallucination. Which is a BIG deal for use cases that might impact revenue if they go wrong.
Useful first agents include automated list import QA, pipeline review bots, lead scoring, campaign performance monitoring, job change alerts, and newsletter drafting. Focus on 1-3 areas, don’t try to do it all at once — you’ll get overwhelmed and burn out.
Prompts should include step-by-step instructions, structured inputs and outputs, and context-aware scoring or recommendations. Make sure you include an example output, it really helps AI understand how you want the deliverable to look/feel.
AI agents have limitations, no matter how much hype they get: they lack human judgment, depend on reliable data/APIs, and aren’t fully autonomous. This ultimately will be a war between the vendors and AI companies — for example, will HubSpot truly allow an AI company access to their entire software suite, or will they try to port their own AI? Lines are still being drawn in this battle, and the outcome will affect how powerful AI can truly become. They need data and access to train and be enabled to actually do things of value.
Be careful with how much you trust Agents, unsupervised…I’ve had people recommend I have Agents automatically fix CRM data for me. While I love the idea, I’d want to keep a close eye at first and set alerts in place, so it can’t go too far off-script and really screw up my data. Also be careful with privacy and PII.
My personal favorite tools include Zapier AI Agents, Clay, ChatGPT, Claude, and LangChain — I find them easy to use, powerful, and trusted by large brands…so I trust them too.
You can use ChatGPT or Claude to draft agent prompts or MCPs by describing the desired workflow or objective. I actually highly recommend this, it has helped me save time.
By the way, the guide is a living doc for the community…please send any feedback you have!
Mid-Year Template Roundup 🤠🐮
Folks let me know they really liked last year’s year-end template roundup, so I’m dropping the most popular this year (so far) here for easy access:
While I have you…what kind of guide/info are you looking for next? Let me know! 😁
What I’m up to/what I’m studying 💭
Right now, I’m on a search for my next horror video game to play, considering Alien Isolation (I know it’s old, but I’ve had it recommended to me many times!)…also highly anticipating Silent Hill f in September. 😍 For those of the Bachelor fandom, I’ve been enjoying the revamped Bachelor in Paradise more than I thought I would! Would love to hear your thoughts.
In terms of professional learnings, I’ve been focused on Agents, HubSpot reporting, and territory planning. 🤓 I’ve been leaning on Zapier and ChatGPT for Agent stuff, HubSpot Academy for HubSpot reporting, and just getting started on territory planning. LMK your favorite learning resources! 👂
Dear Sara ✍️
New to marketing operations? On a team of one at your company? Shy/introverted? Wish you could ask a question to an experienced marketing operations professional, without them knowing who you are? Here’s your chance! Submit an anonymous question to me here and I’ll answer a new question in every issue.
Here’s my answer to a question from last week:
Absolutely love your articles on reporting. My question for you is... how do you account for the passing of time? What I mean is, opps are created when meetings are set. Execs only like to look at stage 2 opps. And in reality we know it could take sales one day to move from a stage 1 to a stage 2 OR it could happen in a completely different quarter. How do you handle that? Are you looking at opps created and moved to stage 2 in quarter or over a cumulative amount of time? It just adds compexity bc we know conversions happen at a different clip even based on the inbound type. So yeah, any insights around timebound reporting would be amazing! TY!
Thank you! 🥹
To account for the passing of time, I like to look at sourcing vs. qualification separately.
For example, if an opportunity was created in Q2 but didn’t hit Stage 2 until Q3, it still counts as Q2 sourced pipeline — we just can’t commit it in Q2 forecasts unless it qualifies in time.
So I usually build two core reports to show the left side of the bowtie:
"Pipeline Sourced This Quarter" →
Oppty Created Date = This Quarter"Pipeline Qualified This Quarter" →
Stage 2 Date = This Quarter
Showing them side-by-side helps tell a more complete story.
From there, I layer in conversion lag data to better explain delays or slippage. I start tracking the average time between:
Meeting → Oppty Created
Oppty Created → Stage 2
Stage 2 → Closed-Won
This helps me highlight where things slow down, especially across inbound types. For example, paid demo opptys may qualify fast, while content- or ABM-driven opptys typically take longer to mature.
If you know conversion speed varies by source or campaign, break your reporting down like this:
Opptys Created by Source + % that reach Stage 2
Avg Days to Stage 2 by Source
That gives you more nuance than a flat stage count.
I tend to follow this model:
Winning by Design Bowtie Model
The left side shows your acquisition funnel — lead to oppty to customer.
The right side shows post-sale value — onboarding, retention, and expansion.
That way, you’re not just showing what got sourced or closed, you’re also showing when value happens, how long it takes, and what channels are most efficient across the entire journey.
I have to give Bowtie Model and info credit to Winning by Design, I took one of their courses and learned it from them!
Thanks for writing in! 😄
If you have a moment, will you let me know if you enjoyed this edition of the newsletter? Thank you!
What Did You Think Of Today's Newsletter?
How can you support this newsletter? 🤔
1. Share this newsletter with friends, they can sign up here.
2. Post about the newsletter on social media, with a link to register!
3. Respond with your thoughts and suggestions on the newsletter!
4. If you are interested in purchasing one of my recommended tools, use the affiliate link in this newsletter to purchase it. It helps boost my content creation efforts, so I can do more cool free stuff for this community!
5. Run marketing at a brand? Respond to this email to inquire about sponsoring this newsletter in the future.
Thanks for reading,
❤️ Sara