Nos offres
Pourquoi Alan ?
Les bénéfices d’Alan pour votre entreprise
Pour nos membres
Votre partenaire santé partout, tout le temps.
Les dirigeants d’entreprise, les indépendants, les salariés
et leur famille ne sont plus jamais seuls pour
leur santé.
Ressources
À propos d’Alan
Rejoignez notre équipe
Ré-imaginez la santé avec des talents internationaux ambitieux et bienveillants.
À propos d'Alan
How We Automated User Interview Scheduling Using AI & No-Code Tools
Tired of manual user interview scheduling? Learn how we automated this process at Alan using AI and no-code tools, reducing scheduling time by 90%. Discover our step-by-step approach to identifying relevant users, sending personalized invitations, and maintaining data privacy.
- The Challenge
- The Solution: User Research on Autopilot
- The Impact: More Insights, Less Busywork
- Key Takeaways for PMs
The Challenge
As a Product Manager, conducting regular user interviews is crucial for building user-centric products. However, identifying the right users and scheduling interviews can be time-consuming and repetitive. With competing priorities and little time, these small frictions can quickly become an excuse to skip this critical step.
The Solution: User Research on Autopilot
I built an automated workflow that identifies users who have successfully or painfully used a specific part of our product, reaches out to them, and handles interview scheduling - all while maintaining data privacy and security.
Step 1: User Identification with AI-Generated SQL
First, I needed to identify users who potentially struggled with our coverage table. I used AI to help write a SQL query that finds users who:
- Consulted their coverage table
- Submitted a coverage-related ticket to our customer support within 5 minutes
This pattern suggests they couldn't find what they were looking for, making them perfect candidates for user interviews.
Here is the prompt I used to craft the query:
I used the prompt with a custom-AI agent that has the context of our data tables, using the great platform Dust.
💡Tip: Any direction you can give to the AI will reduce hallucination! For example, giving the table names can greatly help.
Step 2: Building the Automation Flow
Then, I created a weekly automation flow on activepieces (a self-hosted Zapier alternative, powering automations while preserving security) that:
- Runs every Monday at 2 PM
- Executes the AI-generated SQL query to find 15 relevant Members
- For each Member:
- Safely decrypts their data (privacy first!)
- Sends a personalized Intercom message with a Calendly link 💡Tip: You can use Claude.ai to generate a professional & branded HTML template!
- Logs the outreach in Google Sheets
- Notifies me via Slack when complete
💡Tip: You can pre-populate the invitee fields in the Calendly scheduling page to remove friction and maximize your conversion rate!
The Impact: More Insights, Less Busywork
The automation slashed my scheduling time by 90% - turning an hour of weekly busywork into a quick 5-minute check.
By reaching out to members right after they experienced friction with our coverage features, our response rates jumped significantly. Combined with friction-free scheduling, I’m now conducting 2-3 user interviews on a weekly basis, gathering meaningful insights.
The best part? The system maintains the human touch while handling the technological parts. Each message is personalized, secure, and users feel valued – not like they're part of an automated system.
Finally, experimenting, editing, and scaling is made easy with this approach. It takes me only a few minutes to A/B test copies or to duplicate the flow for another use case.
Key Takeaways for PMs
- Leverage AI wisely: Use AI for tasks like query writing while maintaining human oversight
- Automate thoughtfully: Focus on automating repetitive tasks. My rule: if you’re going to do a task more than 5 times, you can likely automate it
- Privacy first: Always build with security and privacy in mind
As PMs, embracing AI and automation allows us to focus on what truly matters: understanding our users and building better products.
