Beyond Prompting — BMAD Demo Pack (anonymized)

Input dataset for the BMAD demo. All names, companies, markets, domains and dollar figures are fictional. Contents: 1. Slack thread · 2. Founder email · 3. RFP · 4. Discovery call notes · 5. README (cast, contradictions, dry-run checklist).

1. slack-thread.md

Slack: #client-staylocal

Lantern Works workspace · private channel · Tuesday, April 21

Maya (BA) · 11:51 AM Pulled together everything StayLocal has sent so far: RFP v3, Grant's email, the notes from the intro call and the link to their Webflow mockup. Working on the requirements now, aiming to get something over to Grant and Leah by end of week.

Dan (Sales) · 12:20 PM man, that's a bunch! jeez!!! So question… if they accept this and move to Phase 2 (discovery), what's really left? There's a lot to digest here and it's not easy to read IMO

Maya (BA) · 12:34 PM Once they sign off on the requirements, that's the discovery phase done, right? Or at least what they paid for?

Dan (Sales) · 12:41 PM confused by your question… requirements and discovery are separate. Requirements first, then full discovery.

Maya (BA) · 12:47 PM Is that a separate charge or covered by the current $3k?

Dan (Sales) · 1:05 PM separate charge… supposed to be $6k for full discovery, then the build on top of that. Also, Grant keeps saying "direct booking site"? Not understanding that, are we doing bookings or not?? did I miss something?

Maya (BA) · 1:22 PM Also worth exploring: pull listings from their property-management software's API (photos, amenities, calendar, rates, booking rules) instead of making hosts retype everything.

Priya (PM) · 4:54 PM This looks good to me 👍 Let's not promise the API piece until we're clear on what Phase 1 actually covers. Maya, can you send me the open questions before anything goes to the client?

2. founder-email.eml

From: Grant Calloway <grant@staylocal.example>
To: Dan <dan@lanternworks.example>
Cc: Leah Calloway <leah@staylocal.example>
Subject: Here's our RFP, can you give us a plan?
Date: Wed, 01 Apr 2026 19:50:00 -0500
Message-ID: <rfp-v3.0401@staylocal.example>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
X-Attachment-Name: StayLocal_RFP_v3.docx (see rfp.md)

Dan,

Thank you again for making time for us, Leah and I really enjoyed the call and honestly it's the first time we've talked to anybody where we came away feeling like they got it. Here's the RFP, it's v3 so it's gone through a few rounds between the two of us but I'm sure there are holes in it, that's kind of why we're coming to you.

Quick background in case it helps whoever on your team picks this up. We built our cabin up at Lake Pines almost three years ago and we self-manage it, we're on Airbnb and Vrbo and we've had our own direct booking site for about two years. The problem is nobody can find it. Last spring break a guest asked if we could do any better on price, so I priced out the exact same stay, same dates, same cabin: $2,300 on Airbnb, $1,900 on Vrbo, $1,700 on our own site. And we can't even tell guests that through the Airbnb messaging. Leah said if people just knew that, hosts would make more and guests would save more and everybody wins. That's StayLocal.

What we want first is the host side. Get hosts in, get them listed, get them verified so guests can trust them. Our goal is 1,000 hosts in Lake Pines and then we take the same model market by market, and long term we want to plug into the property management software everybody already pays for so hosts come in automatically, think 100,000 hosts. We want to capture every host that comes to StayLocal, their login and contact info, single-property owners like us list free, and the money comes from the pro hosts with 3+ properties and eventually the big management companies. We also want to do a founding host thing where the first 500 or so get 90 days of everything free.

I'll be the first to tell you we are not technical. I built the mockup in Webflow myself over about five weeks and I put Airtable in the RFP because that's what the YouTube videos said to use, but honestly it could be anything, we just want it to be able to grow with us and not have to rebuild it in a year.

On budget, we're not ignorant, we know there's going to be some upfront cost to get this tested. We were thinking somewhere in the $15–20k range all in for the first launch, maybe a little more if it really makes sense, but we'd want to see the plan before we go much past that. Timing wise we'd love to have hosts signing up before the summer season, ideally before spring break next year so the founding host push lines up with when people are booking, but we're also not interested in skipping steps, we have one chance to make a first impression with these hosts.

So I guess the ask is can you take the RFP and everything we talked about and give us a plan? What gets built first, what it costs, what we're missing. Whatever you need from us, logos, the Webflow link, more detail on the host signup flow, just say the word.

Thanks Dan, talk soon.

Grant

Grant Calloway
Co-founder, StayLocal

3. rfp.md

Request for Proposal (RFP)

StayLocal – MVP Platform Build (v3)

File received from the client as StayLocal_RFP_v3.docx and converted to Markdown. The wording is unchanged, and the bracketed placeholders are the client's own.

1. Project Overview

Company: StayLocal (staylocal.example)

Objective: Launch a lean MVP platform connecting travelers directly with property owners while enabling monetization and trust via a Verified Host system.

• Capture host property submissions
• Capture guest inquiries
• Store and manage structured data
• Enable host subscription payments
• Implement Verified Host system

Note: This is a validation-stage MVP, not a full marketplace.

2. Project Goals

• Organize host and guest data
• Enable property submission and approval workflow
• Support host-paid tiers via Stripe
• Implement Verified Host approval
• Build scalable backend structure

3. System Architecture

Webflow → Forms → Stripe → Zapier/Make → Airtable → (Optional) Webflow CMS

Alternative architectures must be justified.

4. Scope of Work

• Airtable database design (Customers, Properties, Inquiries, Verified Hosts, Markets)
• Webflow integration (forms + CMS readiness)
• Automation workflows (host, guest, verification, payments)
• Stripe subscription setup
• Approval workflows
• Optional CMS sync

5. Verified Host System

• Submission form
• Airtable tracking
• Admin approval
• Badge display
• Paid-tier only

6. Payment Requirements

• Stripe integration
• Subscription billing
• No guest booking payments in MVP
• Future extensibility required

7. Out of Scope

• No booking payments
• No login system
• No messaging
• No mobile app
• No advanced analytics

8. Deliverables

• Airtable base
• Automation workflows
• Stripe setup
• Webflow integrations
• Verified Host workflow
• Documentation

9. Proposal Requirements

Approach, Timeline, Cost Breakdown, Experience, Post-launch support

10. Technical Expectations

• Clean, scalable database
• Reliable automations
• Maintainable by non-developers

11. Guiding Principles

• Simplicity over complexity
• Speed to launch
• Scalable foundation
• Monetization + trust

12. Submission Instructions

• PDF proposal
• Include scope, timeline, cost
• Deadline: [Insert Date]
• Contact: [Insert Email]

13. Final Note

Focus is a functional, monetized MVP — not a full platform. This is just our initial launch!

4. discovery-call-notes.md

StayLocal: Free Consultation, call notes

Date: Thursday, April 2 · about 30 min · video call Attendees: Grant Calloway (co-founder, StayLocal), Leah Calloway (co-founder, StayLocal), Dan (Sales, Lantern Works) Source: Notes taken automatically during the call and lightly cleaned up. Quotes are close paraphrases of what the founders said.

Summary

The founders want a direct-booking discovery site for short-term rentals. Guests would find and book host properties directly instead of paying platform fees. It starts in one market, Lake Pines, with a target of 1,000 hosts before going national. The MVP is host first: hosts pay subscriptions, and guests pay no booking fees. Revenue comes from pro host tiers (3+ properties) and enterprise property managers, so a Verified Host process is needed to build trust. Dan walked through the three-step engagement and its pricing: $3k for a 2-hour discovery, $6k for a 2-day full discovery / requirements, and about $12k for a 2-week MVP build.

Details

Background and the problem

• The founders built a short-term-rental cabin at Lake Pines almost three years ago and self-manage it. They list on Airbnb and Vrbo and have run their own direct-booking site for about two years.
• Direct bookings earn more because there are no platform fees, but "it's nearly impossible to get traffic to your site… it's like a needle in a haystack." Guests can't search for direct sites.
• Airbnb monitors guest messaging, so hosts can't tell guests "you'd save money booking direct."

The price gap (the founding "aha")

• Grant priced the same stay (same dates, same cabin) on all three channels: Airbnb $2,300, Vrbo $1,900, direct site $1,700. All three pull from the same dynamic-pricing tool, and the platforms add their own fees on top.
• Leah: if there were a way for people to know that, "hosts would make more money, guests would save money, and everybody wins."
• The guest pitch: "Come find this house in Lake Pines this weekend. Save money. Just book direct." For example, "I can save $600 if I just book direct with you."

Goals: short and long term

• Short term: treat this as a discovery / testing phase. Build a working model in Lake Pines with a small media campaign and get to at least 1,000 hosts, then start sending direct-booking traffic through the site.
• Long term: 100,000 hosts. Partner with the big property-management software (PMS) platforms through API integration, so a PMS customer is "automatically integrated" into StayLocal. Grant estimates one large PMS has a couple hundred thousand customers, and "if we converted 20%, that's 20,000 hosts just like that."
• Expand market by market to other heavy STR markets (e.g. Coral Bay, Cedar Ridge) once Lake Pines works.
• "This is starting out more of a host experience, but it's going to quickly transition into a guest experience."

Host experience and membership model

• Grant has built "the bones" of the host experience in Webflow ("graphically it's pretty crude"). It has a landing page, find cabins, list your property and pricing. The guest experience has not been designed yet.
• Membership is "geared entirely towards hosts." "We want to capture everybody that comes to StayLocal: login credentials, information, contact."
• Browsing is free for guests. Single-property owners list for free.
• Revenue starts with more sophisticated hosts: pro hosts / small property managers with 3+ properties ("a ton of people up there own more than one home, like 30"), and later enterprise customers, meaning large property-management companies with hundreds of listings nationwide. "That's a completely different user experience."
• Founding Host program: the first 500–700 hosts get 90 days of "the whole thing" free.

Verified Host

• "You've got to make sure the host is verified. Do they have insurance? A utility bill that says you actually own the place?" Many direct-booking sites skip this, and "we know we need to have a verified host process."
• Dan's paraphrase: "you build trust as these packages go up." Founders: "Yes."
• Grant describes the basic tier user as "families that own properties… I'm a verified host, I'm doing all the things I'm supposed to do, just like my wife and I."

Technology

• The RFP lists Webflow → Stripe → Zapier/Make → Airtable. Dan said he hadn't worked with Airtable but "it could be anything." Leah: "Whatever it needs to be. We just don't know this piece of it at all."
• The founders' current direct site runs on a PMS backend that handles pricing, calendar and messaging.
• Grant: "Four or five weeks ago I didn't know anything about Webflow… I can paint the picture, I can't get the pictures to move once I paint them."

Process, pricing and timing

• Three steps: (1) 2-hour discovery, $3k; (2) full 2-day discovery / requirements doc, $6k; (3) 2-week MVP build, "usually about $12k," which could change if big unaddressed requirements come out of full discovery. After launch, new work is billed per project or on retainer.
• The founders said the pricing "doesn't scare me at all… that's almost exactly what we were expecting," and that they are ready to move forward.
• Grant: "We have one chance to make a first impression with most of these hosts and travelers, so we're not interested in skipping steps or moving too fast."
• Dan: build the host side first, because "if you get traffic, you don't have any houses."

Next steps (from the call)

• Dan: look into the founders' PMS platform for functionality and integration options.
• Dan: send the agreement and the client document-intake list within 1–2 days.
• Grant: send the Webflow sandbox link (shared during the call).

Addendum: Phase 1 goals as written in the signed statement of work

Lantern Works wrote this, and the founders signed it after the call. It restates the client's goals for the $3k Phase 1.

• Objective: an "MVP host listing site," a visualization the client can use to start selling host memberships on a platform that connects travelers with property owners around Lake Pines first. It should be easy to expand and to copy for other markets.
• Phase 1 covers the host signup stage only. "The initial stage of the project will not include any traveler/guest requirements."
• Deliverables:
• Property submission approval and workflow: a visual walkthrough of how a property owner lists a property.
• Scalability: choose a tech stack that can grow as the client adds features, guest workflows and new markets.
• Verified Host system: an admin plan for submissions, tracking, approvals and badges.
• Subscriptions: "This will ultimately be a subscriber type site for home owners." Recommend the workflows and tech requirements.
• Project brief: core requirements and the proposed technical path.
• Client to provide: RFP v3, a business document describing how a host signs up for an account and registers a property, the Webflow theme link, brand guidelines, and 2–3 inspiration sites plus "non-negotiable" features.
• Next phase: a 2-day requirements document ("full discovery") to finalize total cost and timeline for implementation.

5. README.md

BMAD Demo Pack: StayLocal (anonymized)

This is the input dataset for the live demo in Beyond Prompting (AI Coding Summit NYC, a 20-minute talk with a 10-minute demo, pre-recorded late October). In the demo, Grok Bot pulls client-sourced Slack, email and documents and feeds them through BMAD v6.12.3: Mary (analyst) → product brief → bmad-prd → validation → party mode. The run builds a project plan, an issues list and a PRD from scratch.

Every person, company, market, domain and dollar figure in these files is fictional. The contradictions are left in on purpose: they come from the real engagement, and they are what BMAD should surface.

Files

The real-to-fictional mapping is in mapping-PRIVATE.md. Never show it on screen, never feed it to BMAD, and never upload it.

Cast

• StayLocal: a direct-booking discovery site for short-term rentals. Hosts pay subscriptions; guests pay no booking fees.
• Lake Pines: the launch market, a cabin-rental lake town. Future markets: Coral Bay, Cedar Ridge.
• Price gap: the same stay costs Airbnb $2,300, Vrbo $1,900, or $1,700 direct.
• Host goals: 1,000 hosts in Lake Pines, then 100,000 nationally through PMS integrations. Single-property hosts list free; pro tier is 3+ properties; enterprise tier is property-management companies. Founding Host program: the first 500–700 hosts get 90 days free.
• Money: Phase 1 is $3k, full discovery is $6k, and the MVP build is about $12k. Founder's ceiling is "$15–20k all in… maybe a little more."
• Timeline (fuzzy): "before the summer season, ideally before spring break next year," but also "not interested in skipping steps." The agency floats a 2-week build.

Contradictions BMAD should catch

Use this list to score the run. Mary or validation should flag at least 6 of the 9 as open questions or issues.

1. No login vs. host accounts. RFP §7 says "No login system." The call says "capture… login credentials," and the SOW mentions a host "signing up for an account."
2. No-code stack vs. scalable backend. The RFP prescribes Webflow → Zapier/Make → Airtable, "maintainable by non-developers." The same RFP asks for a "scalable backend structure," the SOW asks for a stack that scales, and the goals call for 100k hosts and PMS API integrations. On top of that, "Alternative architectures must be justified" sits against "it could be anything."
3. Free hosts vs. paid-only Verified Host. The RFP says Verified Host is "Paid-tier only." On the call, single-property hosts list free, yet the basic-tier host says "I'm a verified host," and verification is framed as essential for trust. The Founding Host 90-days-free offer muddies this further.
4. Guest scope. The RFP says "Capture guest inquiries," automates guest workflows, and has an Inquiries table. The SOW says Phase 1 has "no traveler/guest requirements." The call says the product will "quickly transition into a guest experience."
5. "Direct booking site" vs. no bookings. The pitch is "just book direct," but the RFP excludes booking payments, and Dan is asking whether they are doing bookings at all.
6. Phase and price confusion. Does signing off on the requirements end discovery? Is full discovery ($6k) separate from the $3k? Where does the ~$12k build start?
7. Budget vs. ambition. A $15–20k ceiling against multi-tier subscriptions, verification, multi-market support and PMS integrations.
8. Timeline vs. quality. A 2-week build and "before spring break" against "we have one chance at a first impression, we won't skip steps."
9. Single market vs. multi-market data. Launch is Lake Pines only, but the RFP already includes a Markets table and the SOW says "readily copied for other markets." There is also an MVP-vs-platform tension: "not a full marketplace" against "100,000 hosts."

Dry-run checklist

Environment (do this the day before)

• [ ] Node 20.12+: node -v (bmad-method 6.12.3 declares engines.node >=20.12.0)
• [ ] Python 3.10+: python3 --version
• [ ] uv installed: uv --version (install from docs.astral.sh/uv if missing)
• [ ] Fresh empty project folder, e.g. ~/demo/staylocal, and cd into it
• [ ] npx bmad-method@6.12.3 install. Pin the version; npm latest is already 6.12.4. Choose the BMM module and your IDE/agent tool.
• [ ] Confirm the skills exist: bmad-agent-analyst (Mary), bmad-product-brief, bmad-prd, bmad-validate-prd, bmad-party-mode
• [ ] Copy the four input files into ./inputs/. Do not copy mapping-PRIVATE.md or this README.
• [ ] Do one full rehearsal end to end, then rm -rf the output folder (_bmad-output/ or whatever the installer configured) so the recording starts from zero
• [ ] Turn off notifications, close Slack and mail, and hide the browser bookmarks bar

Feed order

1. founder-email.eml: the ask, in the founder's voice
2. rfp.md: the client's structured wants
3. discovery-call-notes.md: the rich context, plus the SOW goals
4. slack-thread.md: the internal mess (phase and price confusion, the API idea)

Opening prompt to Grok Bot: "Pull everything StayLocal has sent us (email, RFP, call notes) and our internal Slack thread, and hand it to BMAD. Start with Mary."

Timing per beat (10:00 demo budget)

If you run long, shorten the PRD beat first (time-lapse) and never cut Mary or validation; those are where the contradictions show. Hard stop at 10:00.

Screen and font

• [ ] Record at 1920×1080. Terminal/editor font 18–20 pt (at least 16 pt for dense Markdown), with a high-contrast theme
• [ ] Zoom the editor to about 125–150% when showing generated docs; collapse sidebars
• [ ] Keep the terminal to about 100 columns wide so lines don't wrap mid-word in the recording

What to capture

• [ ] Full-screen recording of every beat, raw, plus a separate clean take of the payoff
• [ ] Stills: Mary's contradiction list, the validation issues list, one party-mode exchange, the final PRD table of contents
• [ ] Save all generated artifacts (brief, PRD, validation report, party transcript) into demo-outputs/ for slide screenshots
• [ ] Record the wall-clock time per beat for the edit
• [ ] Before publishing, grep every captured artifact for real identifiers; it must come back clean