> ## Documentation Index
> Fetch the complete documentation index at: https://docs.highailabs.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Onboarding AI Emails
> AI-powered welcome email sequence after sign-up -- personalized to each user's preferences, behavior, and cannabis interests to drive activation and Pro conversion.
**Status:** Brainstorm (NEW)
**Phase:** Phase 3 (Personalization) | **Tier:** All Users
## The Problem
A new user who exclusively smokes high-THC indicas for sleep and a user who microdoses sativa edibles for creativity both get the same onboarding experience today. The same generic welcome. The same feature tour. The same nudges. Nothing acknowledges that these are fundamentally different people with fundamentally different reasons for downloading the app.
Generic onboarding is a retention killer. Industry data consistently shows that personalized onboarding sequences achieve 2-3x higher activation rates than templated alternatives. The first 14 days after signup determine whether a user becomes a subscriber or a churn statistic. Every touchpoint in that window needs to feel like the app already understands the user.
AI-generated onboarding emails solve this by treating each new user as an audience of one. The emails are not templates with a name variable swapped in. They contain genuinely personalized content -- strain recommendations based on stated preferences, Professor High insights tailored to their experience level, feature highlights relevant to what they told us they care about.
***
## The Email Sequence
Five emails over 14 days, each timed to a behavioral trigger or calendar milestone:
**Subject line:** Personalized based on signup preferences (e.g., "Welcome to High IQ -- let's find your perfect indica" or "Your cannabis intelligence journey starts now").
**Content:**
* Personalized greeting that references what they told us during onboarding (strain type preference, experience level, what they want from the app)
* One AI-generated strain recommendation based on their stated preferences, with a brief explanation of why this strain fits them
* Quick-start CTA: "Import your first order" or "Browse strains in your favorite category"
* Optional: A short AI-generated welcome clip from Professor High (15-second audio or animated GIF) that mentions them by name and their preferred strain type
**Why Day 0:** Strike while the intent is hot. The user just downloaded the app for a reason. The welcome email should reinforce that reason and give them one clear action.
**Subject line:** "Import your first order and see what High IQ can really do"
**Content:**
* Brief explanation of what happens when they import an order: AI research report, terpene analysis, High Family classification, effects forecast
* Step-by-step visual guide (3 steps with screenshots or illustrations)
* A preview of what a research report looks like -- using a strain related to their stated preferences
* "Your first order analysis is free" callout (reinforces the free tier value)
**Conditional logic:** If the user already imported an order on Day 0, skip this email entirely. Replace with a "Your first report is ready" notification email instead.
**Why Day 1:** The first order import is the key activation event. Users who import within 48 hours have significantly higher retention.
**Subject line:** "3 strains you'd love based on what you told us"
**Content:**
* Three AI-generated strain recommendations, each with a 2-3 sentence explanation of why it matches the user's profile
* Recommendations factor in: stated strain type preference, experience level, desired effects (if collected during onboarding), and any order history (if available)
* Each recommendation includes the strain's High Family badge, dominant terpene, and one standout fact
* CTA: "Explore these strains in the app" with deep links to each strain profile
**Why Day 3:** By Day 3, the initial excitement has faded but the user has not yet formed habits. A genuinely useful recommendation re-engages them and demonstrates the intelligence layer.
**Subject line:** "Meet Professor High -- your personal cannabis expert"
**Content:**
* Introduction to Professor High as the AI chat assistant: what it can do, what kinds of questions to ask, example conversations
* A personalized insight about the user's favorite strain type (e.g., "You mentioned you love sativas -- here's something most people don't know about how limonene and terpinolene interact in sativa-dominant strains to create that energizing effect...")
* Example prompts the user can try: "What's the difference between myrcene and limonene?", "Which strains help with focus?", "Tell me about my last order"
* CTA: "Ask Professor High anything" with deep link to chat
**Why Day 7:** One week in, the user has explored the basics. Professor High is the next-level feature that transforms the app from a reference tool into a personalized advisor.
**Subject line:** "Your first 2 weeks with High IQ -- here's what you've discovered"
**Content:**
* Personalized summary of their activity: strains viewed, orders imported, reports generated, labels scanned, badges earned
* If active: Celebrate their engagement, highlight their most interesting strain, mention features they have not tried yet
* If inactive: Gentle re-engagement with a compelling reason to return (new strains added, a trending strain in their preferred category, a Professor High fun fact)
* Nudge toward features they have not explored (e.g., "You haven't tried the label scanner yet -- point your camera at any cannabis label and get instant terpene data")
* For users on Free tier: Soft Pro upgrade nudge showing what Pro users see that they do not (AI reports, analytics, Professor High, rankings)
**Why Day 14:** Two weeks is the retention cliff. Users who are active at Day 14 have a high probability of becoming long-term users. This email is the last structured touchpoint before the user either converts to a habit or drifts away.
***
## Personalization Signals
The emails are only as good as the data they personalize against. These are the signals available at each stage:
* Preferred strain type (sativa / indica / hybrid / no preference)
* Experience level (new to cannabis / casual / experienced / medical)
* Primary interest (recreation / medical / curiosity / social)
* What they want from the app (track orders / discover strains / learn about cannabis / all of the above)
* AI tone preference (if set during onboarding)
* Orders imported (count, strain types, dispensaries)
* Strains viewed (which profiles they browsed)
* Reports generated (topics, strains analyzed)
* Labels scanned (strains, terpene profiles found)
* Features used (which tabs and screens they visited)
* Session frequency (how often they open the app)
* No order imported by Day 1 = send the import nudge email
* No strains viewed by Day 3 = make recommendations more discovery-oriented
* No features used beyond browse by Day 7 = highlight specific features in the Professor High email
* No activity at all by Day 14 = shift the recap email to re-engagement mode
***
## What Makes These Emails Different
These are NOT templated emails with `{{firstName}}` and `{{strainType}}` variables. The AI-generated content in each email is unique per user. Two users who both prefer sativas will receive different strain recommendations, different Professor High insights, and different feature highlights based on their individual behavior and the current strain database state.
The difference between a templated email and an AI-personalized email:
| Aspect | Templated | AI-Personalized |
| --------------------- | ---------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Greeting** | "Hi John, welcome to High IQ!" | "Welcome, John -- you mentioned you're into high-THC indicas for sleep, so let's start there." |
| **Recommendation** | "Check out our most popular strains" | "Based on your indica preference and interest in sleep, Granddaddy Purple's myrcene-heavy profile (0.42%) makes it a strong first pick. Here's why..." |
| **Feature highlight** | "Did you know you can scan labels?" | "Since you're focused on medical use, the label scanner will be especially useful -- scan any dispensary product to see exact terpene percentages and research-backed effects." |
| **Re-engagement** | "We miss you! Come back and check out new features." | "3 new indica strains were added this week that match your terpene preferences -- including one with the highest myrcene content in our database." |
***
## Technical Approach
### Architecture
| Component | Technology | Notes |
| ----------------------- | --------------------------- | ---------------------------------------------------------------------------------- |
| **Email service** | Resend or Postmark | Transactional email with HTML templates; Resend preferred for developer experience |
| **Sequence scheduling** | Trigger.dev scheduled tasks | One task per email in the sequence, triggered by user signup event |
| **Content generation** | Claude (via AI SDK) | Generates personalized paragraphs, strain recommendations, and insights |
| **User data** | Convex queries | Read user profile, preferences, and behavioral data at generation time |
| **Strain data** | Hono API | Fetch strain recommendations and profiles for email content |
| **Template rendering** | React Email or MJML | Responsive HTML email templates with dynamic content slots |
| **Unsubscribe** | Email service native | One-click unsubscribe per CAN-SPAM / GDPR; granular per email type |
### Sequence Logic
```
User signs up
├── Day 0: Welcome email (immediate)
├── Day 1: Check if order imported
│ ├── Yes → "Your report is ready" email
│ └── No → "Import your first order" nudge
├── Day 3: Strain Discovery email (always sent)
├── Day 7: Professor High Introduction email (always sent)
└── Day 14: Weekly Recap email
├── Active user → Celebration + unexplored features + soft Pro nudge
└── Inactive user → Re-engagement + compelling reason to return
```
### Trigger.dev Task Structure
Each email in the sequence is a separate Trigger.dev task, allowing independent retries, monitoring, and A/B testing:
| Task | Trigger | Delay |
| --------------------------------- | ----------------- | ---------------------------- |
| `onboarding-email-welcome` | User signup event | Immediate |
| `onboarding-email-order-nudge` | Scheduled check | Day 1 (skip if order exists) |
| `onboarding-email-discovery` | Scheduled | Day 3 |
| `onboarding-email-professor-high` | Scheduled | Day 7 |
| `onboarding-email-recap` | Scheduled | Day 14 |
Each task reads the latest user data at execution time, not at scheduling time. This ensures the Day 14 recap reflects 14 days of actual behavior, not a stale snapshot from Day 0.
***
## Tier Impact
| Tier | Experience |
| --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| All users | Full 5-email onboarding sequence. Every email is personalized. The sequence is a conversion tool -- it demonstrates the app's intelligence to drive Free-to-Pro upgrades. |
Onboarding emails are deliberately not gated behind Pro. They serve two purposes: (1) activate new users by showing them personalized value immediately, and (2) demonstrate the depth of AI personalization that Pro unlocks. The Day 14 recap email includes a soft Pro nudge showing the features they are missing.
***
## Dependencies
* [x] User signup flow -- **built and live**
* [x] Onboarding preferences collection -- **built and live**
* [x] Strain recommendation engine -- **built and live**
* [x] Trigger.dev infrastructure -- **built and live**
* [x] AI generation pipeline (Claude) -- **built and live**
* [ ] Email service integration (Resend or Postmark)
* [ ] React Email templates (5 email designs)
* [ ] Trigger.dev onboarding email tasks (5 tasks)
* [ ] AI content generation prompts for each email type
* [ ] Unsubscribe handling and preference management
* [ ] Email analytics tracking (open rates, click rates, conversion)
***
## Open Questions
1. **Email service choice** -- Resend has the best developer experience and React Email integration. Postmark has the best deliverability reputation. SendGrid has the most features but the worst DX. Which matters most for a small user base that will grow?
2. **AI-generated images in emails** -- Should emails include AI-generated visuals (strain art previews, Professor High avatar)? Images increase engagement but add generation cost and email rendering complexity.
3. **A/B testing** -- Should the sequence support A/B testing from Day 1 (different subject lines, different email timing, different content strategies)? Or ship v1 as a single path and iterate based on aggregate metrics?
4. **Unsubscribe granularity** -- Should users be able to unsubscribe from individual email types (e.g., keep the recap but stop the nudges), or is it all-or-nothing?
5. **Post-Day 14 emails** -- Should there be a continued email cadence after the onboarding sequence (monthly recaps, new feature announcements, personalized re-engagement)? Or does the onboarding sequence hand off entirely to push notifications?
6. **Welcome video/audio** -- The Day 0 email could include a short AI-generated Professor High welcome clip. Is the added complexity worth the engagement boost, or should this wait for v2?
***
## Related Features
* [AI Preferences](/planning/personalization/ai-preferences) -- Tone preference (if set during onboarding) shapes the voice of all onboarding emails
* [Smart Suggestions](/planning/features/smart-suggestions) -- The strain recommendations in the Day 3 email use the same recommendation engine
* [AI Podcasts](/planning/ai-media/ai-podcasts) -- The Day 0 welcome clip concept is a miniature version of the podcast format
* [Referral System](/planning/growth/referral-system) -- Onboarding emails could include the user's referral code and encourage early sharing