Before you build another feature, use this workflow

Feature requests are easy to collect and hard to judge.

One customer asks for something. Another asks for something else. Soon, your backlog is full of ideas, but you still don’t know what to build.

Here’s a simple, automated workflow to make prioritizing feedback easier.

What the workflow does

This workflow gives every customer request a simple process:

• First, the request is collected through a form.
• Then it is saved.
• Then, AI summarizes it and tags it.
• Next, you check if similar requests already exist.
• Finally, you decide what action to take next.

The request then ends in one of these places:

• Follow-up question
• Backlog
• Customer interview
• Founder follow-up
• Sales follow-up
• Small build

This keeps you from building unnecessary things.

Tools

You only need a few tools for this workflow:

• Jotform: collect feature requests
• Airtable: store and review requests
• Zapier: connect the tools
• ChatGPT/OpenAI in Zapier: summarize and tag requests
• Gmail: send follow-up emails
• Slack: send alerts for important requests

Keep the system simple. One form. One Airtable base. One Zap. One weekly review.

Step 1 — Build the public request form in Jotform

Start by creating a form in Jotform.
Go to Jotform:

• Click Create Form → Start From Scratch
• Call the form: Feature Request Form

Add these fields (or something similar):

• What feature do you want?
  • Use short text.
  • Make it required.
  • Add helper text: “Example: I want scheduled PDF reports.”
• What problem are you trying to solve?
  • Use long text.
  • Make it required.
  • Add helper text: “Example: I send client reports every Monday, and I build them by hand right now.”
• How urgent is this?
  • Use a dropdown.
  • Make it required.
  • Add these options:
    • Nice to have
    • Important
    • Blocking my work
    • We may cancel if this is not solved
• What are you doing today instead?
  • Use a dropdown.
  • Make it required.
  • Add these options:
    • Doing it manually
    • Using a spreadsheet
    • Using another tool
    • Paying someone to do it
    • Not solving it today
    • Other
• Tell us more about the workaround
  • Use long text.
  • Make it optional.
  • Add helper text: “Example: I spend 45 minutes every week copying data into a document.”
• What is your email address?
  • Use an email field.
  • Make it required

The customer can now send a request with enough detail.

Step 2 — Create the Airtable base

Next, create the place where the requests will be saved.

Go to Airtable:

• Create a base called ‘Feature Requests’.
• Create a table called ‘Requests’.

This table has three kinds of fields.

First, add the fields that come from the customer form:

• Customer email
• Feature request
• Problem
• Urgency
• Workaround
• Workaround details

Next, add the fields that AI will fill:

• AI summary
• Category
• Request group
• Revenue signal
• Decision
• Decision reason

Last, add the fields that you will fill yourself:

• Customer plan
• Account value
• Customer ID
• Revenue score

Use simple field types:

• Use email for Customer email.
• Use long text for Feature request, Problem, Workaround details, AI summary, and Decision reason.
• Use single select for Urgency, Workaround, Category, Revenue signal, Decision, and Customer plan.
• Use short text for Request group and Customer ID.
• Use currency or number for Account value.
• Use number for Revenue score.

For Category, add these choices: Reporting, Integrations, Automation, Permissions, Billing, Import, Export, AI, Admin, Other

For Revenue signal, add these choices: Low, Medium, High, Expansion, Churn risk, Unclear

For Decision, add these choices: Backlog, Ask follow-up, Customer interview, Founder follow-up, Sales follow-up, Small build

Every request now has one place to live.

Step 3 — Create the Zapier trigger

Go to Zapier.

Click: Create → Zap

Set the trigger:

• App: Jotform
• Event: New Submission
• Form: Feature Request Form

Click: Test trigger

If Zapier does not find a request, submit a test request in Jotform.

Step 4 — Add the AI cleanup step

Next, add AI to the Zap.

• In Zapier, click the plus button.
• Choose ChatGPT by OpenAI.
• Choose a prompt or structured data action.

Paste this prompt (or something similar):

You are helping a small SaaS founder review a feature request.
Use only the information provided. Don’t make anything up._
Return:
*   AI summary: One short sentence that explains the request.   
*   Category: The closest request category.
*   Request group: A short lowercase label with hyphens. Example: scheduled-pdf-reports  
*   Revenue signal: The possible revenue impact.  
*   Decision: The best next step.  
*   Decision reason: A short sentence explaining the decision.
    

Rules:
*   Expansion = may lead to an upgrade, more seats, a higher plan, or a paid add-on.
*   Churn risk = customer may cancel or cannot keep using the product without it.    
*   High = clear pain, urgent pain, paid workaround, or high-value account._    
*   Medium = real problem, but money impact is unclear.  
*   _Low = nice-to-have.    
*   Unclear = not enough information.
Request details:
*   Feature: [map Jotform feature request]    
*   Problem: [map Jotform problem]    
*   Urgency: [map Jotform urgency]   
*   Workaround: [map Jotform workaround]   
*   Workaround details: [map Jotform workaround details]_    
*   Customer plan: leave blank for now    
*   Account value: leave blank for now_ 

Click: Test step

Step 5 — Create the Airtable record

Add another Zapier step.

Choose: Airtable → Create Record

Select:

• Base: Feature Requests
• Table: Requests

Map the fields from Jotform. Then, map the fields from the AI step ( Leave Customer plan, Account value, Customer ID, and Revenue score blank for now. You will fill them in during your weekly review.)

Click: Test step

Then open Airtable and check the new row. Each new request should now be saved with the customer details and the AI cleanup.

Step 6 — Create Airtable views for each next step

Now, create a few Airtable views.

Instead of one giant list, create one view for each action. These views should help you answer one question: “What should I do next?”

Create these four:
View 1: Founder Review
This is for requests that may matter to revenue or churn.
Create a new view called: Founder Review
Filters:

• Decision is Founder follow-up
• OR Decision is Sales follow-up
• OR Revenue signal is Expansion
• OR Revenue signal is Churn risk

View 2: Backlog
This is for requests you are not acting on yet.

• Create a new view called: Backlog
• Filter: Decision is Backlog

View 3: Customer Interviews
This is for requests where you need to talk to the customer.

• Create a new view called: Customer Interviews
• Filter: Decision is Customer interview

View 4: Small Build Candidates
This is for requests that may be worth building soon.

• Create a new view called: Small Build Candidates
• Filter: Decision is Small build

These views make the review faster and easier.

Step 7 — Group similar requests

In Airtable, open the Requests table:

• Click: Group
• Select: Request group

Airtable will now group similar requests. Review these groups weekly. Look for:

• Requests that come up again and again
• Paid users asking for the same thing
• High-value accounts asking for the same thing
• Manual workarounds
• Customers who may cancel
• Requests that could lead to an upgrade

AI may not group everything perfectly. For example, these may all mean the same thing:

• Scheduled-reports
• Scheduled-pdf-reports
• Weekly-client-reports

If they are the same request, give them one shared name during your weekly review.

Step 8 — Score the request yourself

Do this part by hand at first. Score at least the first 20 to 30 requests yourself.

Open the request in Airtable. Review:

• Customer plan
• Account value
• Urgency
• Workaround
• Request group
• Revenue signal

Use this scoring (or something similar):

• Paid user: +1
• High-value account: +2
• Blocking their work: +2
• May cancel if not solved: +3
• Manual workaround: +1
• Paying for another tool or person: +2
• Three or more similar requests: +2
• Could become a higher-plan feature: +2

Once you have the final score, use it to choose the next action. Here’s an example:

• 0 to 2: Backlog
• 3 to 4: Follow-up question
• 5 to 6: Customer interview
• 7 to 8: Founder follow-up
• 9 to 10: Smaller build or test pricing

Update these Airtable fields:

• Revenue score
• Decision
• Decision reason

You now have a simple way to compare requests.

Step 9 — Send the right reply

Use the Decision field to choose the reply.

• If Decision is Ask follow-up, send this (or something similar): Thanks for sharing this. Quick question: what are you trying to do that you cannot do today?
• If Decision is Customer interview, send this (or something similar): Thanks for the request. This sounds like a real workflow problem. Would you be open to a quick 20-minute call so I can understand how you handle this today?
• If the decision is Founder follow-up, send this (or something similar): Thanks for sending this. If we solved this well, would it help you upgrade, add more users, or replace another tool?
• If Decision is Backlog, this (or something similar): Thanks for sharing this. We’ve saved your request and will review it with similar requests from other customers. I can’t promise it will be built, but it helps us make better product decisions.

Users now get the right response based on the value of the request.

Step 10 — Add a Slack alert for important requests

In Zapier, add a filter or path.

• Use this rule: Continue only if Revenue signal is Expansion or Churn risk.
• Then add this step: Slack → Send Channel Message

Send this message (or something similar):

Feature request needs review:
*  Customer: [Customer email]
*  Plan: [Customer plan]
*  Account value: [Account value]
*  Request: [Feature request]
*  Problem: [Problem]
*  AI summary: [AI summary]
*  Revenue signal: [Revenue signal]
*  Decision: [Decision]

Next step: Reply personally before adding this to the roadmap.

Now, important requests do not get buried.

Step 11 — Test money before building

Before you build a big feature, email the users who asked for it.

Send this (or something similar):

Hi [First Name],

Thanks again for sharing more about this.

We are considering a small version of this feature.

Before we build it, I want to understand how important it is to you.

If we solved this well, would it help you do one of these?
*   Upgrade to a higher plan    
*   Add more users    
*   Replace another tool    
*   Pay for it as an add-on    
*   Stay with us instead of switching
    

No pressure either way.

I’m asking so we make the right product decision.

Thanks,

[Your Name]

This way, you learn whether the request is connected to revenue.

Step 12 — Build the smallest useful version

If the signal is strong, build an MVP of the feature first.