Intercom Signup Enrichment: Making In-App Messages Worth Opening

Intercom is only as sharp as the data behind it

Intercom is where a lot of PLG products have their most direct conversation with users: a chat bubble, a product tour, a triggered message asking about a feature someone just tried. The tool is good at delivery. It has almost no opinion about who deserves which message, because a raw signup only gives it an email address and maybe a name.

That gap shows up fast. Every new user gets the same onboarding tour. A VP at a 500-person company and a student testing the free plan see the same "book a demo" nudge. Support tickets from your best accounts sit in the same queue as one-off questions from someone who signed up once and never came back. None of that is Intercom's fault. It's doing exactly what it was told, which is not much, because nobody told it who these people are.

Signup enrichment fixes the input problem, not the tool. Groful enriches new users with company, role, and ICP context right after signup, so by the time Intercom sees a contact, it already knows enough to act differently. If you haven't set up the enrichment layer yet, start with the PLG signup enrichment guide before wiring it into Intercom specifically.

What Intercom actually needs to know

Intercom doesn't need a research dossier on every contact. It needs a small set of fields that change what happens next: which tour plays, whether a message fires at all, and whether a conversation gets routed to a human or stays automated.

Who the person is

• Job title, seniority, and function, so messaging can speak to a role instead of a generic "user."
• Whether the email is a work domain or a personal one, and how confident the match is.
• A resolved professional profile URL when one exists, useful context for a rep glancing at a conversation before replying.

A support agent replying to "how do I add a teammate" writes a different answer to a solo founder than to an ops lead at a company that already has twelve seats. Role context makes that difference visible without asking the user to explain their job in the chat window.

Who the company is

• Company name, domain, and employee count band.
• Industry, so vertical-specific tours and playbooks can trigger.
• Whether the company already has other users in your product.

This matters most for personal-email signups, which show up constantly in self-serve PLG. A Gmail address with no company data attached gets treated as low priority by default, even when the person behind it works at a company you'd sign a contract with tomorrow. Groful's guide on enriching personal email signups covers how to recover that context instead of writing those users off.

Fit and routing tier

• ICP fit: high, medium, low, or unknown.
• A short reason string, not just a score. "Series B, 80 employees, growth team of 4" tells a rep more than "82."
• A recommended lane: self-serve, lifecycle, sales-assist, or review.

This is the field that turns Intercom from a messaging tool into a routing tool. Without it, every automation has to guess.

What the user has actually done

• Signup date and source.
• Activation milestones: first project created, integration connected, teammate invited.
• Usage-limit warnings or plan-relevant events.
• Pricing page visits or a demo request submitted elsewhere in your stack.

Fit alone isn't enough to justify interrupting someone with a message. A high-fit account that hasn't done anything yet needs help getting started, not a sales pitch. A high-fit account that just invited three teammates and hit a plan limit is a different conversation entirely.

Where each signal lives in Intercom

Intercom splits data into user attributes, company attributes, and events. Enrichment should map cleanly onto that structure instead of cramming everything into one object.

User attributes carry the person-level signals: job title, seniority, department, enriched email type, user ICP fit, and a recommended next action. These drive per-contact message targeting and give agents context in the inbox sidebar.

Company attributes carry the account-level signals: resolved domain, employee count band, industry, company ICP fit, active user count, and expansion flag. These drive account-wide rules, like pausing generic nurture messages once a company crosses a usage threshold.

Events carry the moments that should trigger something: user_enriched, icp_fit_changed, teammate_invited, usage_limit_reached, demo_requested. Events are what your Custom Bots and Series campaigns key off, not static attributes that only update occasionally.

Keep the attribute list small on purpose. Every field you add is one more thing that can go stale, and a stale ICP fit score is worse than none, because people will trust a number that no longer reflects reality.

Build message rules around lanes, not a single score

A single fit score tempts teams into one rule: message the top 20%, ignore the rest. That produces two failure modes. High-fit accounts that haven't activated get sales messaging when they need product help. Lower-fit accounts that are genuinely stuck get nothing, and some of them churn over something a two-line message would have fixed.

Lane 1: activation help. Any fit tier, early in onboarding, no meaningful usage yet. Trigger a checklist-style Custom Bot or a short in-app tour focused on the first real value moment, not a generic feature list. This lane should be the biggest one by volume.

Lane 2: sales-assist handoff. High ICP fit, meaningful product intent such as a teammate invite, a limit hit, or a pricing visit. Route the conversation to a human instead of a bot, and give the rep the ICP reason and the triggering event in the sidebar so they don't open with "so, tell me about your company."

Lane 3: expansion signal. Existing customer, new user from the same company, or a high-fit teammate discovered through enrichment. Notify the account owner rather than messaging the new user directly. The teammate discovery guide has more on identifying which colleagues are worth flagging.

Lane 4: low-touch. Low fit, low confidence, or ambiguous company match. Keep messaging generic and infrequent. Don't disable it entirely, since some of these will turn out to be miscategorized, but don't spend rep time here either.

Mistakes that show up once volume grows

Messaging every signup the same way. This is the default state before enrichment, and it's the reason teams add enrichment in the first place. If nothing changes after wiring it up, the rules probably still route everyone into one bucket.

Firing a bot message before enrichment finishes. If Intercom sees a new user before Groful has resolved company and fit data, the first message goes out blind. Add a short delay, or trigger the welcome sequence off the user_enriched event instead of user_created.

Sending low-confidence matches to a human queue. A rep who gets three wrong company guesses in a row stops trusting the system, and then ignores real high-fit accounts too. Confidence needs to gate the sales-assist lane, not just inform it.

Letting expansion messages go to the new user instead of the account owner. A cold "hey, looks like you're on a team plan already" message to someone who just signed up is confusing. The signal is more useful as a note to the person who owns the relationship.

A rollout that doesn't break existing automations

Week 1. Map the four lanes to your current Intercom setup. Identify which existing bots and Series campaigns overlap, so you're not running two competing welcome flows.

Week 2. Sync a minimal attribute set: job title, company domain, ICP fit, ICP reason, recommended lane. Confirm the values look right on a handful of real test accounts before turning on any automation.

Week 3. Turn on the activation-help lane first, since it's the safest to get slightly wrong. Watch open rates and reply sentiment for a week before touching anything else.

Week 4. Turn on sales-assist routing and expansion notifications. Review every routed conversation with the team handling it and adjust the fit threshold if reps disagree with what's landing in their queue.

Checklist before calling it done

• Attributes are split correctly between user-level and company-level.
• Bots trigger off user_enriched, not raw signup, so messaging doesn't fire blind.
• Low-confidence matches route to review, not to sales-assist.
• Expansion signals notify the account owner, not the newly signed-up user.
• Each lane has an owner who checks it weekly for the first month.
• Reporting can tie Intercom engagement back to ICP fit, not just open and reply rates.

Give Intercom something worth acting on

Intercom can carry a lot of the weight in a PLG motion, from first-session guidance to sales handoff to expansion nudges, but only when it knows who it's talking to. Groful enriches signups with company, role, and ICP context, scores fit, and surfaces the product signals that tell you when a conversation actually matters. Explore the PLG signup enrichment solution, see how growth managers use it day to day, or get in touch to map your first Intercom routing rules.