SpeedRun Ep12 Conference Attendee Email Lead Parsing Tales from the Sky Lounge
Episode Summary
A developer experiments with using Claude AI to automate conference lead generation by parsing a PDF attendee list and creating a complete HubSpot marketing campaign. Claude autonomously researches the conference, segments attendees, writes email sequences, and designs lead scoring workflows—but hits output limits when generating the final CSV, revealing practical constraints in AI-driven business development automation.
Pull Quotes
“I thought how lazy can we be and make this thing do what we want it to do? So rather than just share the email with Claude, I said, 'Yeah, here's the attachment, a pre-conference registration list. Parse this out and create a campaign at HubSpot with his company with all these attendees enriched from sources that you can find. Just figure it out.'”
“He figured out the conference just from the title of the PDF. He figured out where it is, what day. Oh man, that is so awesome. Then key driver safety infrastructure revenue, liability, attendee profile. Oh, he broke it down. Look at that. He knows who's going there. That is cool. That is so cool. I didn't even ask for that.”
“So it's going to do lead scoring based on just a PDF of attendees. Oh man. Oh my god. Look at this. Key workflows. Okay. Pre, post, during the conference, lead scoring. Okay. Awesome. And then here's the properties. Okay. Compliance. Of course, you want CAN SPAM and GDPR, right?”
On Claude generating marketing strategy documents instead of the requested CSV file
“Very impressive. Very cool to look at, but this isn't what I wanted, right? Parsed the list, started outputting it, and got pretty far, but then hit a wall. Paused because it reads is max length from a message.”
“You got to know what it can and can't do. So we have a new limit. I'm not displeased. I think we're going to explore here in a little bit. You know, this is the typical thing that you run into in data science. You can only do so much on command line before you run out of RAM. So what do you do?”
Transcript
Okay, this one's kind of neat. I was given an email from one of my tennis buddies and he's got this new little company that does wraps for parking lot pillars with advertising on them. So it's kind of fun. He's got some cool videos that make you cringe if you're a car person. Anyway, so I thought how lazy can we be and make this thing do what we want it to do? So rather than just share the email with Claude, I said, "Yeah, here's the attachment, a pre-conference registration list." He's going to some parking lot conference. And I said, "Parse this out and create a campaign at HubSpot with his company with all these attendees enriched from sources that you can find. Just figure it out."
And Claude said, "Hey, I have some questions. What is this? What are their customers like? And what exactly do you mean by enrichment? What format would you like the output in? Where do you want me to put it?" And then observations on the PDF. I guess he paid to be an exhibitor or something. And then this is his company and we want to drive awareness and enrich the data. Anything that you can think of to help the marketing campaign. I have some more questions. What specific type of advertising? What's your primary call to action? That's kind of cool. What's your campaign timeline? And then why? And here's what I'm going to do with that.
And then here's the answers. There's conference. Oh my god, hair on fire. I got to get it done now. And then you have a booth and you have segments. Do you have key value proposition? And oh no. So Claude's going to go figure all that stuff out. Oh no, he wants the answer. Okay. So before he even gave me the answer, he wrote a little document and then allowed me to answer. Let's see what this document is. Oh my god, look at this. He figured out the conference just from the title of the PDF. He figured out where it is, what day. Oh man, that is so awesome. Then key driver safety infrastructure revenue, liability, attendee profile. Oh, he broke it down. Look at that. He knows who's going there. That is cool. That is so cool. I didn't even ask for that. Here's a typical export. If you care about the shape of the data, which maybe you do, enrichment strategy, it's going to go all this stuff. Fortunately, we have this as a tool available. Tech stack. Okay. Pre-conference email. Okay. He wrote an email campaign for you. Post-conference, another three emails. Very good. Very good. You probably want to tweak those out. I'm glossing over them here. Primary segments. Wow. It's going to segment all your people. And then why? Oh my god. Custom lead scoring. Oh man. So it's going to do lead scoring based on just a PDF of attendees. Oh man. Oh my god. Look at this. Key workflows. Okay. Pre, post, during the conference, lead scoring. Okay. Awesome. And then here's the properties. Okay. Compliance. Of course, you want CAN SPAM and GDPR, right? And then here's your KPIs, AB test, AB testing. Ain't that cool? Key talking points, objection handler. Holy crap, man.
Pre-conference, do this. Conference week, do that. After, do that. Next week, do that. Next steps, import the list to it. Still hasn't got to HubSpot yet. I think I got a little sidetracked doing the thing that it knew how to do, which is writing English, which is what it's very good at. And then I kind of answered this and said, "We don't have a booth, blah, blah, blah." And then it said, "Okay, here's your comprehensive email marketing strategy, blah blah blah." Very cool. Very, very cool. But I still don't have the CSV to give him or have anything in HubSpot. I ran into a limit on that particular chat. So I had to go take that chat, shut it down and filled up the memory, which this is a pretty extensive list. So I gave it the same list and then I gave it the plan pasted in from the last chat and I said please prepare the marketing list. And it said okay and there it goes.
And then it didn't—this is not a CSV—but it said here, here's how you think about it and here's some. It went and cherry-picked for me, which I guess is useful, right? But it's not what I asked for, right? Okay. Very impressive. Very cool to look at, but this isn't what I wanted, right? Okay. So let's put that down for a second and then I said okay, please output the leads. Give me a CSV. Parsed the list, started outputting it, and got pretty far, but then hit a wall. Paused because it reads is max length from a message. And then if you hit continue, it may or may not continue, but I don't think it did. So I got a lot of them and then we kind of stopped there and I said, "Hey, what do you want me to do? Keep going or what do you want to do?" So that was a little disappointing. There is a limit there, I feel like. You know, anytime you get to a limit, it's kind of good. And that's for the next one there.
Want to go deeper?
Work With MeRelated Episodes

SpeedRun Ep19 Marketing Plans with Virginie Glaenzer - Tales from the Sky Lounge
Todd and Virginie demonstrate live how to build a complete marketing plan using AI tools like ChatGPT and Claude. They identify lack of clarity as the biggest barrier to AI adoption among executives, design a five-step 'AI Value Blueprint' service offering targeting CMOs, CROs, and COOs in B2B SaaS, and develop persona-specific messaging and automation workflows—all within one hour using conversational AI.

SpeedRun Ep 09 Resource Use for ICP + GTM Plan -Tales from the Sky Lounge
A fractional CTO demonstrates using Claude AI with deep research mode to generate a comprehensive business development plan. By uploading marketing materials, case studies, and service offerings, the AI produces detailed ideal customer profiles, multi-channel outreach strategies, budget allocations, and personalized contact sequences—illustrating how experienced professionals can leverage AI to automate strategic planning while maintaining the judgment to guide and refine outputs.

How Two Veterans Are Solving Business Development with TractionBD | Yates & Bedol, Ep 43
John Yates, a tech attorney with 44 years of experience, and Greg Bedol, a serial entrepreneur and consultant, discuss their AI-powered business development platform Traction BD. They explain how AI is transforming professional services business development from a centuries-old mentor-based model into an automated system that helps lawyers, accountants, and consultants maintain relationships and convert prospects into clients through persistent, personalized outreach.