Hey! 👋 

I’ve been working on rewriting my RevOps team charter for this whole new world of AI, and it’s reminding me so much of writing the case for MOPs investment. This is a topic that many of you DM me about on LinkedIn or ask in Dear Sara, so I thought I’d share my best tips and tricks. And before you say, “I’m an IC, this is irrelevant to me,” listen up! These skills are relevant to you, even if you’re only a team of 1. But storytime first. 📖

Early in one of my MOPs roles, I had a CMO who saw the function as governance and support. A necessary cost center. The team that kept the forms working and the lists clean and stayed out of the way while marketing did the real work. 😥 

It sucked. But hey, it wasn't malicious…it was just the mental model this CMO was living in, due to previous experiences.

But as many of you have experienced, that perception will cost us dearly if left to fester. It will cost us budget. It will cost us headcount. It will cost us a seat at the table on decisions that MOPs should have been part of from the start. And the most frustrating part -- the part I've seen play out at company after company -- is that most MOPs leaders respond to that dynamic by working harder, delivering more, and hoping that just increasing their output will convince other leaders of their worth.

Nobody notices. That's not how this works. Let me tell you why.

The mistake many MOPs leaders make 😬

I’m sorry to say this, but it is the truth: the behavior that keeps MOPs stuck in the governance-and-support box is almost always self-inflicted. Not intentionally -- no one wants to suffer, I know that. And it doesn’t mean you deserve it. BUT, it also doesn’t mean that we are helpless. This is what I see play out:

MOPs does the work. MOPs fixes the problem. MOPs moves on to the next ticket. Nothing gets tracked in a way that shows the impact. Nothing gets communicated beyond the people directly involved. The charter is informal at best, nonexistent at worst. The roadmap lives in someone's head. MOPs is not included in strategic conversations, so they end up getting buried in catch-up or clean-up tasks. You know what I mean…fixing attribution because no one planned it from the start, cleaning up botched list imports, doing things manually “because we’ll only need it this 1 time, we don’t need a process for it!” 🫠

And then when budget season comes or a reorg gets announced, the CMO/CEO/CRO/COO genuinely doesn't know what MOPs does -- because MOPs never told them. They’re caught up in super high level discussions and activities, so they aren’t close enough to see any of the work happening. They have to rely on hearsay, often from folks who candidly don’t want to admit that they needed help or created a mess. The result is the same: they fund what they understand, and they cut what they can't see.

The perception problem isn't the CMO's fault, though. Like, I wish they would pay closer attention, for sure, but…it's a communication problem that MOPs has to own.

I can hear you all thinking: “Alright, well that was depressing enough, Sara. What do I DO about it?????” 😒😡

More coming after the break, including a guide you can use at work… 📺

This section is sponsored by Default, a partner who supports this newsletter. 🧩

✨ Everyone is telling you to build AI agents. Nobody is telling you to fix your data first.

A human SDR works around a messy CRM every day without thinking about it. They know the account ownership is wrong. They know the enrichment data is six months old. They compensate with judgment and move on.

An AI agent doesn't do that. It reads the field, trusts the value, then acts on it -- confidently, at scale…and in the wrong direction. The data problem you've been tolerating becomes an agent problem you can't ignore.

The unexciting prerequisite: you have to fix the data layer before you launch an agent. Or at minimum, create clear enough definitions that you can prompt around the mess. Either way, the foundation can't be skipped.

Default keeps enrichment, routing, ownership, and your GTM stack in sync on one workflow canvas -- so instead of wrangling APIs and duct-taping point solutions together, you have a clean data layer your agents can trust.

🔗 See how it works: default.com

…back to our regularly scheduled programming. 📺

How to get yourself out of this quicksand

Once I understood the perception and visibility issues, I stopped trying to convince my CMO that MOPs was strategic -- and started making it impossible for her not to see it.

Make the work visible. The first move was getting everything into a tracking system (think: Asana, ClickUp, Linear, Jira) -- not for MOPs' benefit (although we do benefit!), but so there was an undeniable record of what the team was doing, how long it took, and what it produced. When you have that data, you can show what you’ve done and how it helps build the business. Without it, you're asking leadership to trust your word against their assumptions. And this is superpowered with AI, by the way — with MCPs, it’s so easy to pull quick reports and insights and push them to Slack or wherever you need to! More on that in the guide.

Do internal marketing. I KNOW, I KNOWWWWW. We’re in ops, we don’t want the spotlight. We’re a bunch of nerds, I know. But we HAVE to do our own internal marketing. In my case, we started sharing wins in the broader #marketing Slack channel -- not every ticket, but the work that had exciting impact. If another team lead the initiative and we supported it, we made sure that we were mentioned and thanked in the messages. We built showcases into the marketing all-hands agenda, live in front of the whole team. At least 1x per quarter, we had the full stage for at least 15 minutes, to showcase the impact we were making. We created a walking deck that new hires and new marketing leaders got taken through so they understood the function before they ever had a chance to mischaracterize it. The MOPs team that does great work in silence will always lose to the MOPs team that does decent work loudly. This still is CRITICAL to your career, I cannot emphasize this enough! 🗣️🔈

Uplevel the work itself. Visibility only works if there's something worth seeing. 🤷‍♀️ We made a deliberate decision to stop accepting every support request as equally important and start protecting capacity for strategic work. When the CMO started seeing MOPs outputs show up in pipeline conversations and campaign performance reviews, the mental model shifted on its own. This is another really uncomfortable piece, but you HAVE to carve out time to make these assets and do strategic work, even if it means dropping things for a bit. If you’re worried about getting in trouble, talk with your manager: explain that you’re drowning in tactical work and need to do a week sprint to catch up on these decks, assets, efforts etc so you don’t end up just becoming ticket takers.

Going from “no budget for ops” to “you got it! ✅”

Do NOT waste your energy trying to convince leaders of the strategic importance of MOPs if you do not have all of the key pieces we’ve talked about in place.

Listen: when a CMO believes MOPs is governance and support, they're usually right -- based on what they've been shown. The argument for strategic partnership can't be made by just saying it, it has to be demonstrated over time, through a track record that's been deliberately made visible.

Fix the foundation first, then make the ask.

cue funny TV host but that’s not all!

I’ve created this guide you can steal and use at work, to help you! Check it out here. ✍️ 

What I’m up to/what I’m studying 💭

I’m still working on weaving more and more AI into our GTM at Vector, including automated digests into Slack — but a big topic in the last few weeks has been my RevOps Bot.

The idea is to have a handy AI Slackbot that is trained and armed with documentation and can answer employee questions about our HubSpot instance or any of our practices or definitions across GTM at Vector. Things like:

• What are our deal stages and definitions?

• What is an MQL?

• What is our ICP?

• Which HubSpot list do we use for our newsletter?

• What is our master suppression list?

And to 1) log questions it doesn’t know the answer to, so we can prioritize creating documentation around those questions 2) create or edit existing documentation to reflect updates or add more context as the business changes.

Here’s how we set it up at a high level (deeper dive to come as we launch): The bot is built on two main APIs: Slack's API (so it lives natively in Slack, responds to @mentions, and can respond within threads) and Anthropic's API, which is the company behind Claude — the AI model doing the actual reasoning (Claude Code itself is doing a lot of heavy lifting to help build it for us).

When someone asks a question, Slack passes it to Claude, Claude decides whether to look in Notion or HubSpot, fetches the documentation or context, and sends the answer back to Slack.

All of that happens in a few seconds behind the scenes. We also gave it a sort of "conventions" layer — a running document of team-specific definitions (like how we define an open deal, which differs from HubSpot's default) that gets baked into every response automatically.

We’ve made a lot of progress! I have v1 in our Slack instance and am testing it now, hoping to launch to the team by end of this week or early next week. 🙂 

BTW: this is another way for us to avoid getting buried in repetitive tickets! If AI can take over a lot of the grunt work of answering questions and maintaining documentation, that’s a huge win for us and for our stakeholders! I’m hoping we can evolve it to even help onboard folks as the GTM teams grow…

Btw check out our bot’s “logo”:

P.S. If you want to follow along with our wider GTM strategy and tactics, follow the Vector blog! You can subscribe to our newsletter too! 👻

This section is sponsored by Knak, a partner who supports this newsletter. 💌

✨ AI is generating more email drafts than ever. The bottleneck just moved.

Most teams are investing in AI on the creation side without thinking about what comes next. Every ops person in the room already knows what comes next -- because they're the ones who inherit the pile.

The gap between "draft exists" and "on-brand, reviewed, approved, and launched" is where teams bleed time. Faster draft output means more in the pipeline, and the downstream pressure compounds. Brand guardrails, editing, collaboration, approvals, translations, MAP syncing -- that work still needs a home. An AI agent without a production layer behind it just moves the bottleneck.

Knak just announced MCP support, which means AI agents can now call Knak directly and get a launch-ready asset back. Your orchestration layer can now touch any part of the production process -- QA, module-level performance data flowing back into your stack, on-brand assets generated from wherever your team already works.

OpenAI figured this out early. Jeff Canada, their Head of B2B MOPs, uses Knak as the production layer behind their AI workflows -- and he was on stage at Adobe Summit last week with Knak's co-founder walking through exactly how they've built it.

🔗 See how it works: knak.com

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:

Yes! And honestly, this is more common than people admit -- a lot of MOPs leaders came up through marketing and figured out the operational side on the job. So first, don't panic. 😅

A few things I'd prioritize:

Learn the data layer first. Before anything else, understand how data flows through your stack -- from form fill to CRM record to sales queue. You don't need to be a data engineer, but you do need to know what touches what and in what order. Walk through a lead from creation to close manually in your own systems. Be your own “mystery shopper.” 😎

Get comfortable with the CRM. If you're a HubSpot shop, go through HubSpot Academy's MOPs and RevOps certifications. If you're Salesforce, Trailhead is free and pretty good. The MAP is usually the comfortable part for someone with a marketing background -- the CRM is where the gaps tend to show up.

Learn to speak revenue. MOPs leaders with a pure marketing background sometimes struggle to translate their work into pipeline and revenue language. Start reading your company's pipeline reports, sit in on sales forecast calls if you can, and get familiar with the metrics your CRO really cares about. This is the single biggest credibility unlock.

Find your community. MOPsPros, the MarketingOps.com Slack community, and RevOps Co-op are all great places to ask questions without judgment. The MOPs community is unusually generous with knowledge.

Resources I'd recommend:

• HubSpot Academy (free, practical, covers lifecycle and reporting well)

• Trailhead for Salesforce basics

• Winning by Design's Revenue Architecture course if you want to understand the full GTM motion

• The RevOps Co-op course library for sales and CS operations context

• My own guides on saramcnamara.com/guides -- the lifecycle, attribution, scoring, and campaign hierarchy ones are good starting points for someone building foundational knowledge

The most important thing: don't try to learn everything at once. Identify the two or three areas where you feel the most exposed in your current role and start there. The rest will come.

Good luck! 🫶

News of the week 🗞️

HubSpot moved to outcome-based pricing for two of its Breeze AI agents this month -- Customer Agent and Prospecting Agent now charge per resolved conversation and per qualified lead handed to sales.

Why you should care: this is the pricing model that will define the next era of AI tooling. "What do I pay when it doesn't work?" is now a legitimate question to ask every AI vendor. 👀

Gartner is predicting that more than 40% of agentic AI projects will be canceled by end of 2027 -- not because the demos disappoint, but because costs escalate, risks surface, and business cases never solidify. Scott Brinker's latest piece on the state of martech in 2026 is worth a full read if you've been quietly skeptical of the hype.

Why you should care: the orgs that survive the cancellation wave are the ones that built the governance and prioritization layer before they started building agents. If you've been doing that work, you're ahead of where most teams will wish they'd started.

HubSpot also launched MCP server support in April, meaning AI agents can now connect to and take actions inside HubSpot directly without custom integrations. It's early access, but the direction is clear.

Why you should care: this is the infrastructure shift that makes the AI use cases we keep talking about actually buildable without a full engineering team behind you. Worth getting familiar with before your next planning cycle.

Clay launched MCP support with Functions this month -- meaning reps can now run Clay enrichment workflows directly inside Claude or ChatGPT, while admins control which Functions are available and set per-user credit budgets so nobody burns through credits unchecked.

Why you should care: candidly, this makes MCP finally usable…you can provision a budget of credits per user, which means your AE can’t accidentally use your entire budget on accident.

Inflection.io acquired Keyplay on April 8th -- bringing account scoring and ICP intelligence directly into Inflection's AI-native MAP. Keyplay CEO Adam Schoenfeld joins as CMO. His take on why: standalone account intelligence is a feature, not a platform. When agents are building campaigns, they need to know who to target and why -- and that context can't live in a separate tool.

Why you should care: this is the consolidation pattern to watch across the martech stack right now. Point solutions that don't connect to an execution layer are becoming acquisition targets. If you have a standalone tool in your stack that doesn't tie directly to action, it's worth asking how long it stays standalone.

What else have you heard about recently? Reply back to this email to send me any other word on the street. 👂

Interesting tech of the week ⚙️

WisprFlow: an AI voice-to-text tool that seems way better than Siri.

My AEs are obsessed with Wispr for all communications, they claim to rarely type anymore! 🤯 Which has me testing Wispr as well. I have a lot of documentation to create for existing workflows and processes, and I’ve been using Wispr to speed that up by enabling me to do verbal brain dumps while on walks, etc. It’s pretty dope! It’ll even do formatting for you. And it’s definitely worth a look if your hands are tired of typing. 😅 (Reminder: these AI tools of the week are never sponsored!)

This section is sponsored by Zapier, a partner who supports this newsletter. ⚡

✨ Your AI agent framework isn't the problem. The integration layer underneath it is.

The models are smart. The orchestration is getting better. The demos look great. And then you try to make your agent do something real -- update a deal in Salesforce, send a follow-up in Slack, log a note in HubSpot, reschedule a calendar invite -- and suddenly you're not building an agent anymore. You're building an integration layer from scratch.

Every tool needs its own connector, its own auth flow, its own token management, its own retry logic, its own error handling. I've watched people spend a week just getting OAuth to behave across four tools before their agent could do a single useful thing. Meanwhile the exec sponsor is asking when they can see a demo. 😅

The Zapier SDK just dropped into open beta and it's worth your attention. A few lines of TypeScript and your agent can take actions across 8,000+ integrations. Zapier handles token refresh, retries, rate limiting, and auth. Your agent just calls the actions.

Their docs example says it well: a user asks to move their 2pm meeting to Thursday and notify the attendees. The agent finds the event in Google Calendar, reschedules it, and DMs every attendee in Slack. No OAuth nightmares. No token rotation fires.

If you've been stuck in integration purgatory while your AI strategy sits on a slide deck -- this is worth a look.

🔗 Check out the Zapier SDK: zapier.com

Meme of the week ⚙️

If you have a moment, will you let me know if you enjoyed this edition of the newsletter? Thank you!

Thanks for reading,

❤️ Sara

P.S. Special thank you to our sponsors for supporting the creation of this newsletter. 🫶