Attio + Koji: Launch AI Interviews from Your CRM and Sync Every Insight Back to the Record

The short version

Connect Attio to Koji and your CRM becomes a research engine: select any contact, company, or list in Attio, launch a personalized AI interview in voice or text, and have Koji write the resulting themes, quotes, and structured scores straight back onto the record. No exporting CSVs, no copy-pasting notes, no separate research tool your CRM never hears about again. The result is an Attio workspace where every account carries living voice-of-customer data — and a research workflow that runs continuously in the background.

This guide shows you how the integration works, what data flows where, and how to set it up.

Why connect Attio and Koji?

Attio is the modern, data-model-flexible CRM that startups and revenue teams use to track every relationship. But a CRM only knows what your team logs — it captures deal stages and emails, not why a customer churned, what job they hired you for, or how they''d rank your roadmap. That "why" lives in conversations most teams never have time to run at scale.

Koji closes that gap. Koji is an AI-native research platform that runs moderated interviews — in voice or text, participant''s choice — without a human moderator. Its AI interviewer asks intelligent, context-aware follow-up questions in real time, then analyzes each transcript automatically into themes, verbatim quotes, and quantified answers. Wire Attio and Koji together and you get:

• Targeted research from real segments. Fire interviews at exactly the accounts that matter — recently churned, expansion-ready, stuck in onboarding — using the lists you already maintain in Attio.
• Personalized interviews. Koji reads each record''s attributes (name, plan, industry, lifecycle stage) and personalizes the conversation, so respondents feel known and answer more openly.
• Insights that live where you work. Every finding flows back onto the Attio record, so account owners see the voice of the customer without opening another tool.

What data flows between Attio and Koji

The integration is bidirectional:

Attio → Koji (trigger + personalize)

• Contacts and companies become interview participants
• Record attributes personalize each interview
• Attio list membership or attribute changes act as triggers ("interview every account that enters the At-Risk list")

Koji → Attio (sync insights back)

• Interview status (invited, started, completed) written to the record
• Top themes and a short AI summary appended to the contact or company
• Structured question scores (for example an NPS scale value or a single_choice reason-for-churn) saved as attributes you can filter and report on
• A link to the full transcript and the study report

Because Koji''s structured questions return clean, typed values, the data landing in Attio is not just free text — it is filterable, chartable fields you can build views and reports around.

Setup

You can connect the two platforms in three ways, from no-code to fully custom:

Option 1 — Zapier or Make (no code)

The fastest path. Use a trigger like New record in Attio list to kick off a Koji study, and a Koji interview completed trigger to update the Attio record. See the Zapier automation guide for the exact recipe; the same pattern works in Make.

1. In your automation tool, connect both Attio and Koji.
2. Trigger: record added to an Attio list (for example "Churned — last 30 days").
3. Action: create a Koji interview and send the invite, passing the contact''s email and attributes for personalization.
4. Second Zap/scenario: on Koji interview completed, update the Attio record with the summary, themes, and structured scores.

Option 2 — Webhooks (custom, real-time)

For engineering teams that want direct control, Koji emits webhooks on every interview lifecycle event. Point them at a small handler that writes to the Attio API. This gives you sub-second sync and full control over which attributes get updated.

Option 3 — MCP + API (AI-driven)

Koji ships a Model Context Protocol integration with 15 tools, so you can drive the entire workflow from Claude: "Pull the At-Risk companies from Attio, launch a churn-reason interview for each, and write the top theme back to the record." Under the hood this uses Koji''s developer API and Attio''s API together. See how AI interviewers work for the interviewer side, and the CRM research integration guide for importing participants at scale.

A concrete workflow: automated churn research

Here is a pattern teams put live in an afternoon:

1. In Attio, maintain a list called Churned — 30 days.
2. When a company enters that list, Koji automatically sends its primary contact a short voice or text exit interview built on the Customer Discovery methodology.
3. The interview opens with an open-ended "walk me through why you decided to cancel," then asks a single_choice primary-reason question and a scale question on likelihood to return.
4. Koji''s AI probes each answer with real follow-up questions and, thanks to its quality gate, only counts sessions that clear a coherence and effort threshold — so junk responses never pollute your data or consume credits.
5. On completion, Koji writes the churn reason, sentiment, likelihood-to-return score, and top themes back onto the Attio company record, and drops the full report link on the account.

The outcome: a self-updating churn-reason breakdown you can pivot on in Attio, sourced from real conversations instead of a one-click cancel survey.

Why Koji, not a survey tool bolted onto your CRM

Most "CRM feedback" is a one-question NPS blast. Koji replaces that with a real conversation. Where a survey captures a number, Koji''s AI interviewer captures the number and the reasoning behind it — asking "what would have changed that score?" the way a skilled researcher would. Compared with traditional tools like SurveyMonkey or Typeform wired into a CRM, Koji delivers depth (adaptive follow-ups), structure (6 typed question types), and automation (auto-analysis and write-back) in one loop. It is the modern, AI-native way to keep your Attio data honest about what customers actually think.

What it costs to run

Koji uses a simple credit model, so automated research from Attio stays predictable. A text interview costs 1 credit and a voice interview 3, and Koji''s quality gate means low-effort sessions never consume credits at all. Every new account starts on a free tier with 10 credits (no card), and paid plans begin at €29/month (Insights, 29 credits) and €79/month (Interviews, 79 credits), with pre-paid credit packs when you need more. Because interviews only bill when a real conversation clears the quality bar, you can point Koji at a large Attio list without worrying that abandoned or junk sessions run up the tab.

A quick setup checklist

• Confirm the Attio attributes you want to write back to (for example Churn Reason, Last NPS, Interview Summary) exist on the object
• Decide your trigger: a specific list, a lifecycle-stage change, or a manual "research this account" action
• Map Koji''s structured question IDs to the matching Attio attributes so scores land in typed fields
• Test with a single record end to end before enabling the automation for a whole segment
• Put the report link somewhere prominent on the record so account owners actually see it

Koji''s MCP integration can even run this checklist conversationally — ask Claude to create the study, wire the fields, and dry-run one record before you flip it on.

Related Resources

• Structured Questions Guide — the 6 question types that become filterable Attio attributes
• CRM Research Integration Guide — import participants and personalize every interview
• Zapier Research Automation — no-code recipes to trigger interviews and sync results
• Webhook Setup — real-time, custom sync for engineering teams
• How AI Interviewers Work — the technology behind Koji''s follow-up questions
• Generating Research Reports — turn interviews into shareable reports