What holds up
The job is painfully clear.
Slow replies cost local service businesses real jobs. The owner understands the proposed outcome without learning a new category or vocabulary.
Public sample delivery · Leadkeeper
The finished package, before you trust the promise.
Leadkeeper is a fictional product. The prototype, evaluation, test plan, and scope decisions are built to the same depth as a paid Promptlaiy delivery.
01 · The input
The original messy idea
“Small home-service businesses lose leads because the owner is on a job and replies too late. I want an app that collects website and text inquiries, writes a response, and reminds the owner to follow up. Maybe it could become a full CRM later.”
Workflow we chose
A new lead arrives, a useful reply is drafted, and the owner reviews and approves it from one screen.
Why this oneIt isolates the riskiest promise: whether a drafted reply actually helps an owner respond faster.
02 · The clickable prototype
Try the core workflow yourself.
Select a lead, inspect the context, edit the drafted response, and approve it. The sample uses realistic mocked data and saves nothing.
✓ Three realistic leads
✓ Editable draft reply
✓ Approval state
✓ Responsive layout
Open Leadkeeper
03 · The blunt evaluation
Useful workflow. Unproven reason to switch.
What could kill it
Another inbox may make the problem worse.
If owners must remember to open Leadkeeper, it adds a habit instead of removing one. Delivery through a channel they already watch may matter more than the dashboard.
Biggest assumption
A draft is faster than a saved template.
The concept only earns its complexity if the context-aware reply feels meaningfully better than the text snippets owners already keep on their phones.
Bottom line
Do not build the CRM. First prove that five busy owners will review and send these drafts during a real workweek—and that they reply faster than they do today.
04 · The validation plan
Three tests. Each can prove the idea wrong.
-
Test 1
Watch five owners use the prototype
Recruit five plumbers, cleaners, landscapers, or garage-door operators who personally answer leads.
- Ask
- Handle two sample leads without coaching, then explain what felt missing.
- Pass
- 4 of 5 approve or lightly edit a reply in under two minutes.
- Fail
- They ask for CRM features before the reply itself feels valuable.
-
Test 2
Run a concierge workweek
For three owners, manually turn their incoming leads into drafts and text them an approval link.
- Measure
- Median reply time and percentage of drafts sent.
- Pass
- Reply time drops by at least half and 70% of drafts get used.
- Fail
- Owners ignore the drafts or rewrite most of them from scratch.
-
Test 3
Ask for a real commitment
Offer the concierge version for $49 for the next month. Do not ask whether they “like” the idea.
- Ask
- Pay now, introduce another owner, or schedule setup with calendar access.
- Pass
- 2 of 5 make one of those commitments without a discount.
- Fail
- Praise is high but nobody gives money, time, or reputation.
Included in this delivery
Enough to run the tests
- One responsive clickable workflow
- Realistic mocked leads and message drafts
- Blunt written evaluation
- Three-test validation plan
- Source package and setup instructions
- One revision and a 60-day preview
Intentionally not built
Because it would not answer the question
- Production login or team accounts
- Website, SMS, or email integrations
- AI model or automated message sending
- Billing, subscriptions, or admin tools
- Full CRM, analytics, or contact history
- Security review or production deployment
Your idea gets the same treatment
Apply first. Confirm the scope. Pay only if the workflow is useful.
If we cannot define a useful one-workflow prototype, we decline the project instead of taking payment.