Cross-Pollinator Auto-Build System Demo

Project We Meeting - March 2026
Larry, Jon, Stacey, Greg

TASK COMMITMENTS BY PERSON

Larry:

• Will send list of businesses found near shops with clickable URLs to test logo/image scraping accuracy (Timeline: before end of week)

(Larry): "I'll send you stuff. What I'll do is we'll get you the list of businesses around you that we find, which will then have a URL that you can click on and you can go see what does it bring up for them."

• Will build page with three different graphics selection and comment functionality (Timeline: by Friday)

(Larry): "Then hopefully by Friday, we'll have the page where if you go to the page, it'll then have the ability to change three different graphics and have the places where you can do the comments piece, you can see how that works."

• Will hand system over to development team to build 12 different card templates

(Larry): "My job is to build it and go, here's the proof of works. And then I hand to the team and go, here's the things you need to add in or to clean up. Or in this case, here's the, you know, 12 different templates you need to go build"

• Will return to building John's messaging system matched to customer grading and probability scoring (Timeline: starting today)

(Larry): "Starting today, I'm starting on that. Because because of where we've gone, that's like that's the next big thing."

• Will send business category word list for review DONE

(Larry): "And if you guys are interested, we can send you this piece for you. Just look through it and go, oh, that makes sense or no, let's not, you know, I wouldn't use that word"

Jon & Greg:

• Will test business list URLs and report logo/image scraping success rates

(Larry): "You can see how well that works and tell us that's good, that's bad, here's what you're missing, where the heck did you get these, You know, this person's been out of business for twenty years."

• Will provide feedback on business category assignments and offer templates

(Larry): "Just so we see what's the reliability of the way in which we're finding stuff. Yeah. You know, and what's the best way to find stuff."

OVERVIEW OF DISCUSSION

Cross-Pollinator Auto-Build System: Larry demonstrated automated system that scrapes business websites to build cross-pollinator cards with zero manual design work. System achieves 60% logo detection rate, automatically pulls three hero images, determines business type from website content, and sizes all elements to fit card layout using pixel-perfect calculations (e.g., Blue Moon Bikes font sized to 29 pixels with -0.1 pixel spacing to hit 319px width against 321px target).

Frictionless Partner Onboarding Flow: Shop owners send unique URL to local businesses. Partner clicks link, system pre-populates their website URL from shop's referral, auto-detects logo and images, identifies business category, presents category-appropriate offer templates. Partner clicks through in under 5 minutes without typing anything unless scraping fails. System generates card in under 1 second but displays fake "building" animation to create perception of work being done.

Business Category Intelligence: System categorizes businesses (restaurants, fitness, health/beauty, auto, lodging, home services, education) by scanning website content for category-specific keywords. Each category triggers different offer template options. Larry included two experimental categories: funeral homes (for memorial events) and faith organizations (churches/mosques for community events) - recognizing cross-pollinators aren't always coupons, sometimes they're event passes or access.

Four-Box Feedback System: Partners review auto-generated card with four response options: "What sucks" (needs fixing), "I like it, but could you maybe..." (suggestions), "Pretty good, but I was hoping for..." (minor tweaks), "Nothing needs change, share it with [Shop Name]" (approval). This captures feedback spectrum without forcing structured forms or phone calls.

Logo Scraping Challenge and 80% Goal: Current 60% logo detection rate with goal of reaching 80%. Sites that block scraping or lack proper logo files get placeholder "Your Logo Here" image. System pulls three images so partners can choose if first auto-selected image is wrong (demonstrated with pizza place that changed website, making previously good image now look terrible). Partners can upload logo manually if scraping fails.

Email Spoofing for Shop-Branded Outreach: System will send partner invitations appearing to come from shop owner's email address (not Larry's or generic system email) unless shop has configured DNS A-class records to prevent email spoofing. Larry acknowledged this is "hacking" but necessary for authentic outreach. Shops with anti-spoofing DNS settings will need to send invitations manually.

No Large Language Models in System: Larry emphasized system uses deterministic code logic, not AI generation. Assets are scraped and assembled by rules, not generated by prompts. This eliminates risk of AI "making up something really stupid" or hallucinating content. System cannot create what doesn't exist - it only finds and formats what's real.

Businesses Without Websites Edge Case: Greg raised concern about partners lacking web presence. Larry's position: handle exceptions outside main tool to avoid confusing 95% of users with unnecessary options. Some shops have used Facebook/Instagram as fallback sources, but system won't force everyone to see social media input fields. Manual handling for <5% edge cases preserves simplicity for majority.

Community Bedrock Philosophy: Larry identified three community pillars: auto shops, barbershops/nail salons, and local taverns - the only places where people gather and talk about topics unrelated to why they're there. Unlike hospitals (talking about sickness) or vets (talking about pets), these businesses facilitate community connection beyond their service. Cross-pollinators leverage this natural community hub role.

Banana Ball as Business Model Validation: Discussion of Savannah Bananas baseball proving Larry's 40-year argument that customer happiness metrics matter more than quarterly reports, ARO, or KPI obsession. Team sold out across six teams with 2 million people in lottery for 60,000 seats, yet founder doesn't track detailed financials - only asks "did that make customers happy?" Larry predicted 3 years ago they'd surpass Major League Baseball within 10 years.

Greg's Community Radio Venture: Greg launched 302.fm (Delaware area code) after leaving WDEL talk radio's $2,000/month Sunday 10am slot between financial show and religious programming. Now building local business interview shows (half-hour BBQ show, metal music show "Rats in the Cellar"), getting board invitations from YMCA and networking with nonprofits. Demonstrates local radio's continued community value despite "local radio is dead" narrative.

Local Radio as Emergency Infrastructure: Larry's observation: people claim local radio is dead until there's an emergency, then everyone realizes its importance. Discussion of ham radio's underground popularity and low-frequency signals reaching 5,000 yards for hyper-local messaging. Greg researching these alternatives as community communication backbones.

AREAS OF AGREEMENT

• Cross-pollinator system dramatically reduces time investment

(Larry): "Can I make it so that you can get someone from you send them an email to doing it in less than five minutes of your time?"
(Jon): "And it's out of our hands, really." [agreeing with automation reducing workload]

• Events and passes are valid cross-pollinator applications beyond coupons

(Larry): "You're not asking someone to donate, you're asking them to go somewhere... I'm just giving them a pass to something."
(Jon): "I think that's a great idea."

• Industry moving away from metrics obsession back to community connection

(Greg): "I think a lot of operators are tired of the metric, the KPI drills... we weren't raised that way in this industry. We were raised with, you know, can I fill it up? How about a case of soda?"
(Larry): "I've fought for forty years against corporation, against running a community business like a corporation. You're not a frickin corporation. Those are your friends. Those are not your customers."

• Local radio remains community bedrock despite digital shift

(Larry): "I will always say, I don't care what comes and goes. Local radio will always be the heart of a community."
(Greg): "There isn't I mean, I've had to do a lot of homework on radio and their ham radio is exceptionally popular" [validating Larry's point with evidence of continued local audio importance]

📊 CONFIDENCE-CALIBRATE: 88%

High confidence in commitment extraction. All task items clearly stated with supporting quotes. Specific timeline given for Friday delivery of graphics selection page. Minor uncertainty around exact scope of "12 different templates" mentioned (whether this refers to 12 business categories or 12 offer variations). Business category word list mentioned as optional review item but no firm commitment to send it - interpreted as conditional on shop owner interest.