the 11 Claude workflows that ate my busywork
the most-read Looped In of 2026, back in your inbox
I wrote this one ahead of time. By the time it lands, I’ll be a few days into a long 4th of July weekend with the family. So, no new post this week.
Instead I’m replaying the most-read issue of Looped In this year. Partly because I want the week off. But also because a bunch of you subscribed after March and never saw it, and it’s the one post I still get replies about.
👋 Hi, it’s Kaylee Edmondson and welcome to Looped In, my newsletter exploring demand gen and growth frameworks in B2B SaaS. Subscribe to join 2k+ readers who get Looped In delivered to their inbox every Sunday.
This week’s replay 📼
Every Demand Gen Use Case I’m Running in Claude Right Now
This one outperformed everything else I’ve published this year.
The TL;DR is 11 workflows I run in Claude: discovery call research briefs, the weekly HubSpot to Salesforce data pulls, personalized ABM copy by industry and persona, and eight more. Plus the order I’d build them in if you’re starting from zero. The boring stuff comes first, and I stand by that.
If you read it back in March, read it again and actually build one this time. If you’re newer here, start with this post. It’s the fastest way to understand how I think about AI inside demand gen work.
I spent 4 hours last week training a client’s marketing team on Claude. Not the “here’s how to write a blog post with AI” kind of training. The kind where we built a shared marketing brain (very similar to the one I shared here last week), loaded it with ICP definitions, competitive battle cards, messaging guidelines, and editorial standards, and then showed the team how to use it to augment the parts of their job that are most repetitive and could be 80% augmented with this new brain.
By the end of the session, we had talked through additional potential use cases, and it felt like it was finally starting to click for a room full of people who had been using Claude to “help me rewrite this email.”
I’ve been playing and building in Claude Code and Claude Cowork for a few months now. I keep finding new use cases that save me time, help me think differently, iterate on concepts faster. And like I shared on LinkedIn earlier this week, I feel like I’m spending every waking moment possible in this new stack, yet still feel more behind in my craft than I ever have. So I’ll say this, if you’re building, exploring, testing in a new tool this week, you’re right where you’re supposed to be. There’s so much hype, especially on LinkedIn these days, and these tools are shipping new models, functionality, and features faster than ever. And that combined makes all of us feel like we’re falling behind. But as long as we keep building, and sharing what we’re learning, I think we’ll all turn out just fine.
So, let me open up a bit about what I’ve been testing and learning.
👋 Hi, it’s Kaylee Edmondson and welcome to Looped In, my newsletter exploring demand gen and growth frameworks in B2B SaaS. Subscribe to join 2k+ readers who get Looped In delivered to their inbox every Sunday.
Why Most Marketers Are Getting 10% of What Claude Can Do
Most demand gen teams I talk to are using AI for two things: writing first drafts of copy, and summarizing meetings. That’s fine. But you’re leaving a lot on the table.
The gap between “I use Claude” and “Claude runs half my workflows” really comes down to context.
I built what I call a “marketing brain.” It’s a set of markdown files that load automatically at the start of every Claude session: who I am, who my clients are, how I write, how I run campaigns, what tools I use, my ICP definitions, my messaging house. Claude reads all of it before I type a single word. That means when I ask it to do something, it already knows my business. It’s not starting from zero every time.
(I wrote about the marketing brain concept last week, so I won’t belabor it. If you missed it, go read that one first.)
Here are the 11 use cases I’m running (or building/finessing) right now. Some of these save me 30 minutes a week. A couple of them eliminated entire workstreams.
1. Discovery Call Research Briefs
Before every discovery call with a potential client, Claude pulls together a research brief. Company background, the prospect’s LinkedIn activity, any public talks or posts they’ve done, likely pain points based on role and company stage.
Last week I had a call with a Head of Demand Gen at a customer experience platform. Claude surfaced that she’d spoken publicly about whether marketing attribution is broken, that she was actively hiring a Demand Gen Manager and ABM Manager (suggesting the engine is early-stage), and flagged that she likely was the budget holder. So I should position DemandLoops as complementary to her hiring plan, not competitive with it.
90 seconds. That used to be 20-30 minutes of LinkedIn stalking and Googling.
2. Weekly Data Pulls from HubSpot + Salesforce
I’m embedded in a client right now where the HubSpot-to-Salesforce integration is... let’s just say it’s a project.
Every week, Claude pulls data from HubSpot, runs VLOOKUPs against Salesforce records, cleans up naming conventions, deduplicates contacts, and flags anything that looks off.
Not glamorous. But this used to eat 3-4 hours a week, and if you skip it, we had no idea what to go optimize for pipeline.
3. CRM Property Mapping Across 5 Systems
Same client. They run HubSpot, Salesforce, Vitally, NetSuite, and PandaDoc. Five systems. Trying to figure out which property maps to what across all of them in a spreadsheet made me want to quit consulting. (I’m being dramatic. But only slightly.)
I built a Claude assistant that takes the property lists from each system and creates a unified mapping doc. Markdown file for quick reference, structured spreadsheet for the full picture. Now when someone asks “where does this data live?” I can answer in seconds instead of opening five admin panels.
This is one of those use cases where Claude is serving as a bandaid solution. Eventually these systems will all be cleaned up, or replaced entirely, synced to the data warehouse, and integrated appropriately, but for now while we’re in the messy middle, post M&A (we’ve all been there), this Claude task is doing some heavy lifting.
4. Competitive Intel, Weekly
I have a competitive intel workflow that runs every week. Claude pulls from competitor websites, checks their ad libraries on Meta and LinkedIn, and flags what changed: messaging shifts, new product positioning, campaign themes, creative formats they’re testing.
The output is a structured report. What changed, what it probably means, whether we need to respond. Typically, quarterly competitive reviews are already stale by the time they ship. This approach keeps you within a week-ish of what competitors are doing.
5. Personalized ABM Ad Copy and Landing Pages
For one client’s ABM program, we’re building hundreds of individualized ads and landing pages. And I mean individualized. Not “Hi {Company Name}” personalization. Actually different messaging by industry, company size, and persona.
Claude generates the copy variations using our ICP definitions and messaging house as the foundation. We pipe account data through Clay for enrichment. The output is account-specific ad copy and landing page content.
6. Lead Scoring and Account Tiering
I’m building what I’m calling an enterprise appetite scoring matrix for a client. Six components for now: company size signals, tech stack indicators, buying intent, engagement depth, organizational complexity, and budget authority signals. I’ll also add in their GTM Alpha. Claude will weigh each one and assign a tier.
The part I find most useful is what I’ll call “synthetic attributes” for now. Data points that don’t actually exist in your CRM but can be inferred from combinations of other fields. For example: you might not have a “budget authority” field, but you can infer it from title seniority + company size + the presence of a procurement process. Claude is surprisingly good at this kind of inference when you give it a clear framework to work within.
7. Salesforce Flow Documentation
If you’ve ever inherited a Salesforce instance with 40+ automation flows and zero documentation, you know this pain.
Claude analyzes the flows, documents what each one does, flags redundancies, and identifies which ones are actually firing vs. sitting dormant. What would have been a two-week documentation project took about three hours. My output for the first run was far from perfect, but probably 60-70% there.
8. Campaign Consistency Checks
Messaging drift is real. Especially when you have three or four people writing copy across email, ads, landing pages, and social.
I built a workflow where Claude checks any new piece of copy against a client’s campaign strategy doc and messaging house before I launch it. Flags anything off-brand, off-message, or inconsistent with what they’ve already published. Takes about 10 seconds. Replaces what used to be a “can you review this” Slack thread that took a day to resolve.
9. Newsletter Topic Development
I use Claude to help me develop newsletter topics, but probably not in the way you’d expect. I don’t ask it to “give me 10 newsletter ideas.”
Instead, I have it pull from my meeting notes (via Granola), my Slack conversations, and current industry trends, then find the intersections. Where does my lived experience this week overlap with what the market is talking about?
This newsletter is a good example. Claude surfaced that the intersection of “I just trained a client team on AI workflows” and “the industry is obsessed with AI in marketing but nobody’s sharing specific use cases” was a strong topic. The pattern-matching was collaborative. The writing is mine.
10. Automated Content Maintenance
For clients with large content libraries, I’m building a series of Claude agents that pull existing site content, find what needs updating (outdated stats, broken cross-links, FAQ gaps), make the changes, and push them back for approval.
Nobody wants to do this work. I’m finding it sits undone for months. But it compounds. Outdated stats kill credibility. Broken links hurt SEO. FAQ pages that don’t reflect the current product confuse prospects. Claude is great for this grunt work.
11. Meeting Prep and Follow-Up
Every morning, Claude pulls my calendar, cross-references it with my meeting notes from prior conversations with the same people, and gives me a prep brief. After meetings, it processes the transcript and drafts follow-up emails, action items, and internal notes for my team.
The follow-up emails are where the time savings really add up. Claude knows my voice, knows the client context, and knows what was discussed. The drafts need light editing, not full rewrites. I was spending 15-20 minutes per follow-up before. Now it’s 2-3 minutes of editing.
How I’d Prioritize If You’re Starting From Scratch
If you’re looking at this list and wondering where to start, here’s the honest answer: start with the boring stuff.
Data pulls. Research briefs. Competitive monitoring. Copy QA. The work you do every single week that requires knowing your business well but follows a predictable pattern.
After that, move to the structured creative work. ABM copy, account scoring, content ideation. These still need your judgment. But Claude handles the 80% that’s pattern-matching, and you focus on the 20% that requires taste. Claude has very little natural taste.
Last priority: the one-off projects like documentation.
The through-line across everything I listed: the marketing brain is the multiplier. Every single use case works dramatically better because Claude already knows my clients, my ICPs, my voice, and my tech stack before I ask it to do anything. Without that context layer, you’re prompting from scratch every time.
Two follow-ups if you want to keep going: 3 AI-Native Demand Gen Plays You’re Not Running and My most-used Claude Skill.
I’ll be back next Sunday with fresh content. Hope you’ve enjoyed the long weekend. Demand gen will, once again, survive without you for a weekend.
See ya next week,
Kaylee ✌