Intercom Search Not Working? Here's Why (and What to Do About It)
Intercom Search Not Working? Here's Why (and What to Do About It)
Intercom's search is one of the more capable search tools in the customer support category — but it has well-documented limitations that frustrate support teams at scale. If you've found yourself losing conversations, unable to surface old tickets, or jumping between Intercom and Slack just to answer a customer question, this post explains why and what you can do about it.
The Most Common Intercom Search Failures
1. Conversations from 12+ months ago don't surface
Intercom's conversation search prioritizes recent activity. Older conversations, especially those from before your current plan tier, may not return in results at all. This is a known limitation: Intercom indexes recent conversations more aggressively, and deeper history requires exact-match filters or exporting to a data warehouse.
What this means in practice: A support rep asking "did we ever deal with this payment issue before?" will get incomplete results if the original conversation is from 14 months ago. The answer may exist in Intercom — but search won't find it.
2. Knowledge base articles and conversations aren't searched together
Intercom separates its Help Center (articles) from Conversations. When you search in the Inbox, you're searching conversations. When you search in Help Center, you're searching articles. There's no unified search that returns both.
For support teams, this means two separate lookups every time you're not sure whether the answer lives in a past conversation or a documented article.
3. Custom attributes and internal notes are hard to search
Intercom's search works well on conversation content and contact email/name. It works poorly on custom attributes you've added to contacts or companies, and inconsistently on internal notes within threads. If your team uses internal notes to log context that isn't customer-facing, that context is often invisible to search.
4. Full-text search requires exact-match or close proximity
Intercom's search is keyword-based, not semantic. If a customer described their issue as "the export button does nothing" and you search "data export failed," you may not find the conversation. Search requires knowing the words used, not just the concept.
5. Cross-workspace search doesn't exist
If your organization uses multiple Intercom workspaces (common for companies with separate product lines or regions), there is no unified search across them. Each workspace is siloed.
The Workarounds (and Their Limits)
Exact-match filters: Intercom does let you filter conversations by assignee, tag, status, time range, and custom attributes. Using filters alongside search narrows the result set. This helps when you know how the conversation was categorized but doesn't help when you're searching conceptually.
Saved searches and segments: You can save filter combinations as segments. Useful for recurring workflows, not for one-off discovery.
Intercom's AI features: Intercom Fin (their AI agent) and AI Copilot search your Help Center articles and can surface relevant answers. These are designed for customer-facing and agent-assist workflows, not for searching historical conversation data.
Export to data warehouse: For complete historical search, teams sometimes export Intercom data to BigQuery or Redshift and run SQL queries against it. This is accurate but requires a data engineer and isn't practical for frontline support reps who need an answer in 30 seconds.
The Deeper Problem: Context Lives Across Tools
Even if Intercom's search were perfect, it would only solve half the problem.
Support and customer success teams don't store all their context in Intercom. A customer's onboarding context is in a Notion doc. The technical explanation for why a bug happens is in a GitHub PR comment. The internal escalation decision is in a Slack thread. The account context for a key customer is in a Jira ticket.
When a support rep gets an escalation and needs to understand the full customer history before responding, they're searching Intercom for conversation history, checking Slack for internal decisions, opening Notion for account docs, and looking at Jira for reported bugs — four separate searches, all with different interfaces, returning incomplete pictures.
This is the real search problem. Not Intercom specifically, but the fact that support context is distributed and none of the individual tools search the others.
What Actually Helps
For Intercom-specific search, the main levers are:
- Use tags and custom attributes consistently so filters can substitute for search
- Use Intercom's conversation export for deep historical audits (Intercom → Reports → Export)
- Use Intercom Fin for Help Center search in agent assist mode
- For multi-workspace: accept the limitation or consolidate workspaces where possible
For the broader cross-tool search problem, a few teams use AI-powered knowledge search that connects Intercom to their other tools. AskOro connects to Intercom, Slack, Notion, Jira, GitHub, Google Drive, and Confluence — and makes all of them searchable from a single interface. When a support rep asks "what's the history on this payment issue with Enterprise customers?", AskOro searches across all connected sources simultaneously and returns a synthesized answer with sources.
The setup takes about 15 minutes per integration. Pricing is $49/month for the entire workspace, not per user.
Intercom Search Checklist
Before concluding that something is a search bug, check these:
- [ ] Is the conversation within the last 12 months? (Older history may need date-range filters)
- [ ] Are you searching the right section? (Inbox vs Help Center are separate)
- [ ] Did you try exact phrases the customer actually used?
- [ ] Is the conversation tagged or filtered in a way that excludes it from default search?
- [ ] Are you looking for something that lives in a different tool entirely?
If the checklist confirms you're hitting Intercom's limits, the long-term fix is making your support team's knowledge searchable across all the places it actually lives — not just in Intercom.
Intercom product features and limitations current as of August 2026.