Move from Intercom to Ochre in a weekend.
A 25-agent team saves roughly $48,000 a yearmoving from Advanced + Fin to Ochre on equivalent capability. Same answers, ~200× cheaper per resolution because BYOK strips the per-resolution markup. Migration is concierge: bulk export, conversation IDs preserved, messenger swap, 30-day parallel run.
Intercom
Conversations, contacts, companies, articles, tags
Ochre
Aliases, RAG, customer context, messenger, BYOK AI
Weekend cutover map
Honest checklist of every entity.
We accept Intercom’s bulk export ZIP (JSONL), API pulls, and older CSV exports. Parser: src/lib/migration/parsers/intercom.ts. It handles the conversation-parts shape, maps source/admin author types correctly, and skips internal notes from the public timeline.
Four steps. We do most of them.
Source
Intercom workspace
Mapping
Bulk/API export
Destination
Ochre workspace
Connect to your Intercom workspace
Settings → Apps & Integrations → Developer Hub. Create an OAuth app or an access token with read scopes for conversations, contacts, companies, articles, and tags. Paste it into Ochre’s migration tool. We never request write scopes during the pull.
We pull conversations via the Intercom API
All conversations + parts, contacts, companies, custom attributes, tags, and admins are pulled with the original Intercom conversation IDs preserved as searchable aliases. Configurable window: last 30 days for a pilot, or the full history (5,000 conversation cap in v1, lifted on request).
AI auto-generates KB articles from your top intents
We cluster your inbound conversations by intent and draft KB articles for the top 10–30 questions Fin used to handle. A human reviews and ships. This usually closes the gap between Fin’s tuned answers and Ochre’s RAG layer in two afternoons.
Swap the messenger when you’re ready
We replace your Intercom('boot', {...}) snippet with Ochre’s equivalent — most apps need 5–10 lines changed. We mirror Intercom for 30 days so new conversations land in both. Flip the snippet when the team is comfortable.
What you actually save.
Drawn from /compare/intercom. Fin at $0.99 per resolution is one of the cleanest pricings in the industry — also one of the highest. The math compounds.
| 5 agents | $195/mo on Ochre | $792/mo on Intercom Advanced + Fin |
| 25 agents | $975/mo on Ochre | $3,960/mo on Intercom |
| Annual difference at 25 agents | +$48,000/year |
At 1,500 Fin resolutions/mo, Intercom bills $1,485 in AI. The same answers cost about $7 in OpenAI tokens through Ochre’s BYOK setup — ~200× cheaper per resolution. See /byok for the full token math.
We do this with you, not at you.
Concierge migration is bundled with the Scale tier. You get a Slack Connect channel with a founder for as long as the move takes. We run the API pull, review the company-attribute mapping, hand-hold the messenger snippet swap, and stay on call through the 30-day parallel run.
For Team-tier customers, the same pull is self-serve in the dashboard. If your workspace has 250K+ conversations or unusual custom attributes, ping us anyway — we’d rather scope it on a 20-minute call than have you run into a parser ceiling at 3am.
What about Intercom-specific X?
What about Series, tours, and proactive messaging?+
Ochre doesn’t ship a Series-equivalent in v1. Two paths: (1) keep Intercom on a Messenger-only / Engage plan just for Series and use Ochre for support inbox + AI — the combined bill is usually still cheaper than full-suite Intercom; or (2) move proactive messaging to Userflow, Chameleon, or Appcues. We help scope both during the migration call.
Will I lose Intercom messenger quality?+
Honestly, Intercom’s messenger is best-in-class. Ours is good but not the gold standard. If website chat is your primary support channel and your customers expect that polish, the trade is real. If your support is mostly email + Slack Connect with messenger as an option for logged-in users, Ochre’s messenger is plenty. We’d rather you stay on Intercom than oversell here.
What about Fin training and tuning?+
If Fin is tuned on your Help Center + custom training, that work doesn’t directly port. We re-ingest your articles into Ochre’s RAG layer — quality typically matches within 2 weeks of usage. Plan a tuning sprint and use the AI-generated KB drafts (step 3) as the starting point.