Airtable + Koji: Turn a Base Into a Customer Research Engine

Answer first: Airtable is where thousands of teams already keep their customers, leads, churned accounts, beta lists, and feature requests. Koji is the AI-native research platform that turns any of those rows into a real conversation. The integration runs both directions and needs no native connector. Into Koji: a new or updated record in an Airtable view (a fresh signup, a closed-lost deal, a support escalation) triggers a Koji AI-moderated interview to that person automatically. Back into Airtable: the moment Koji finishes analyzing the interview, it writes the top themes, overall sentiment, a quality score, and a transcript link back onto the exact record it came from. Airtable exposes automations, incoming webhooks, and scripting; Koji exposes a headless API, HMAC-signed webhooks, and a Zapier app. Wire them through Zapier, Make, or native Airtable webhooks and a working v1 takes about 30 minutes.

If you have ever built an Airtable view called "trial users, day 14, still not activated" and wished you could just ask them what happened, this is the integration that turns that view into a stack of analyzed transcripts by the end of the week.

Why connect Airtable and Koji

Airtable is a superb system of record. What it cannot do is tell you why a record looks the way it does. A row can show you that a deal was lost, a subscription cancelled, or a feature request was logged — but the motivation behind it lives in the customer's head, not in a cell. Teams usually paper over that gap with a quarterly survey blast or a handful of manually scheduled calls that sample a rounding error of the list.

Koji removes the scaling ceiling. Its AI moderator runs voice or text interviews around the clock — no scheduling, no human moderator, no calendar Tetris — and asks adaptive follow-up questions instead of running a fixed script. That means one triggered interview reaches the "why" a static form never gets to, and because Airtable already segments your audience into precise views, you are never interviewing "customers" in the abstract. You are interviewing the 84 specific people in one specific view. Platforms like Koji make the depth of a moderated interview available at survey scale, and Airtable is the trigger that decides exactly who gets one.

What flows in each direction

Airtable → Koji (trigger interviews from a record)

• A record enters a filtered view or matches an automation condition (status changed to "Churned", a signup row is created, a CSAT score drops below 3).
• Airtable forwards the record to Koji, which sends a personalized interview link by email, SMS, or in-product embed.
• The person completes a voice or text interview whenever it suits them.

Koji → Airtable (write insights back onto the record)

• When Koji fires its analysis_ready event, a forwarder updates the originating Airtable record with:
  • Top themes (Koji's automatic thematic coding, clustered across the interview)
  • Overall sentiment (positive / neutral / negative)
  • Quality score (Koji only counts conversations scoring 3+ as research-grade)
  • Transcript link and interview mode (voice or text)
• Downstream Airtable automations, interfaces, and dashboards update instantly.

Three ways to connect (pick one)

There is no one-click Airtable app inside Koji, and you do not need one. Choose the path that matches your comfort level:

1. Zapier (about 30 minutes, no code). Use the "Airtable — New/Updated Record in View" trigger to call Koji's Start Interview action, and a "Koji Interview Completed" trigger to run an "Airtable — Update Record" step. This works on any Koji plan, including Insights.
2. Make.com (about 45 minutes, no code, more branching). The same shape as Zapier but with visual routers — handy if different views should launch different studies.
3. Native Airtable automations + Koji webhooks (about an hour). Use Airtable's "When record matches conditions" trigger with a "Send webhook" or scripting action to hit Koji's API, and Airtable's "When webhook received" trigger to catch Koji's analysis_ready event. This keeps the whole loop inside Airtable and Koji, no middleware.

Step by step (Zapier path)

1. Build the trigger view in Airtable. Filter to exactly the cohort you want to hear from — for example, Status = Churned and Cancelled Date is within the last 7 days.
2. Create a study in Koji. Design your interview using Koji's six structured question types (see below), then grab the study ID.
3. Zap 1 — Airtable to Koji. Trigger on new record in the view; action calls Koji's Start Interview endpoint, passing the customer's email as respondent_email (or your Airtable record ID as respondent_id so the write-back lands on the exact row).
4. Zap 2 — Koji to Airtable. Trigger on Koji "Interview Completed"; action updates the matching Airtable record with themes, sentiment, quality score, and transcript URL.
5. Test with one record, confirm the round trip, then let it run.

Design questions so answers come back clean

The quality of what lands back in Airtable depends on how you design the study. Koji supports six structured question types, and each one maps to a tidy, filterable value in your base:

• open_ended — the qualitative heart of the interview; returns AI-clustered themes and quotes.
• scale — NPS, CSAT, or 1–10 satisfaction; returns a number you can roll up or chart.
• single_choice — pick one; returns one clean category value.
• multiple_choice — pick many; returns a list.
• ranking — order options by preference; returns a ranked array.
• yes_no — a crisp binary field.

Mixing a couple of structured questions (a scale plus a single choice) with open-ended questions gives you the best of both worlds: chartable columns in Airtable and the conversational "why" behind them. Because every structured question carries a stable ID, the answers aggregate consistently across every interview in the study — see the structured questions guide for the full breakdown.

Identity resolution

If your Airtable identifier is an email, pass it to Koji as respondent_email and it echoes back on the webhook. If you key records on Airtable's own record ID or a customer UUID, pass it as Koji's respondent_id, which returns unchanged so the insight attaches to the exact row — even for pseudonymous or account-level records. For sensitive topics, run the study in anonymous mode and skip the write-back entirely.

What plan do I need?

Zapier automations and CSV participant import work on any Koji plan, including Insights (€29/month, 29 credits). The headless API and HMAC-signed webhooks — required for the native Airtable-automation path — are on the Interviews plan (€79/month, 79 credits) and Enterprise. Credits are consumed only by research-grade conversations: a text interview costs 1 credit, a voice interview 3, and a report refresh 5, and Koji's quality gate means low-effort or spam conversations (scoring below 3) never burn a credit. New accounts get 10 free credits to prototype the whole loop before committing.

High-value use cases

• Churn win-back. A "Churned this week" view auto-launches an exit interview; the reasons land back on each account so CS can triage.
• Lead qualification depth. A "Demo requested" view triggers a short discovery interview before the sales call, and the themes ride along on the record.
• Feature request validation. When a request table row crosses a vote threshold, interview the requesters to learn the real job behind the ask.
• Beta feedback at scale. A beta list in Airtable becomes dozens of analyzed interviews without a single scheduled call.
• Post-purchase / onboarding. New-customer rows trigger a day-14 check-in that writes activation blockers straight to the record.

Traditional tools like SurveyMonkey or Typeform can push a form link from Airtable too — but they stop at the first answer and hand you a pile of raw text to code by hand. Koji's AI probes each answer in real time and returns clustered themes automatically, so what flows back into Airtable is insight, not homework.

Related Resources

• Structured Questions Guide: The 6 Question Types in Koji
• Google Sheets + Koji: Launch AI Interviews from a Spreadsheet
• Connect Koji to Zapier: Automate Research Workflows
• Connect Koji to Make (Make.com): No-Code Research Automations
• Sync Koji Research Insights to Notion
• Research Automation: Build Real-Time Pipelines with Webhooks