Findings sat in the backlog with no status change for two weeks, these are the ones losing money silently.
At a glance
Open Vortex IQ-tagged Intercom tickets/conversations that have had no team-side activity in 14+ days. The silent-leak subset of the open backlog: findings that were filed, never picked up, and are quietly compounding revenue risk while the team focuses on live shopper messaging. On Intercom merchants this card is particularly important because the team’s daily rhythm prioritises live conversations over Back-office ticket triage; without this card, the abandoned bucket grows unnoticed.
| What it counts | Open Intercom tickets tagged vortex_iq whose last_admin_assignee_id change, status change, internal note, agent reply, or tag change is older than 14 days. Snoozed tickets are excluded; if a ticket was snoozed and the snooze expired without team action, the auto-reopen counts as zero movement and the abandonment clock continues from the original last-action timestamp. |
| What counts as movement | Any of: agent reply, internal note, status change, assignee change, tag change, applied macro. Customer reply does NOT count (the card focuses on team progress). Fin AI auto-reply does NOT count by default (treating Fin auto-reply as movement would mask abandonment). |
| Project / board scope | All Intercom workspaces the connector token reads. Multi-workspace merchants see a blended count by default. |
| Status filter | state:open only. Closed tickets drop out; snoozed tickets are excluded (intentionally deferred is not abandoned). |
| Issue type filter | All Vortex IQ-tagged tickets included regardless of Intercom Type (Customer / Back-office / Tracker). |
| Resolution counts | N/A; abandonment is about the open population only. |
| API endpoint | POST /tickets/search with body filtering on tags, state, and updated_at < now-14d. The 14-day staleness clock evaluates on every webhook event and on a 5-minute scheduled refresh. |
| Time window | RT |
| Alert trigger | >5 warn, >15 critical. The warn threshold catches early stagnation; the critical threshold reflects systemic execution failure on Intercom-CX-led teams. |
| Sentiment key | Threshold-based, {warn: 5, critical: 15}. |
| Time zone | Workspace timezone in Settings, Workspace, General; the 14-day cutoff aligns to it. |
| Roles | owner, operations |
Calculation
Calculated automatically from your Intercom data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A US SaaS-meets-DTC brand running Intercom Pro across both their B2C ecommerce site and their B2B account portal. Single workspace, two segmented Inboxes. Snapshot taken on 02 May 26 at 15:45 ET.| Inbox | Open Vortex IQ findings | Abandoned (>14d no movement) | Notes |
|---|---|---|---|
| B2C Storefront | 18 | 7 | Mostly post-release findings from 18 Apr; assigned to a senior agent on PTO. |
| B2B Portal | 8 | 2 | Two SSO callback findings; the engineering owner is in India timezone and rarely opens Intercom. |
| Total open | 26 | 9 | At warn (>5), below critical (>15). |
- Abandoned doubled vs 30D average. The headline open count (26) is healthy in absolute terms, but the queue is stagnating. The team is filing Vortex IQ tickets at a normal rate; the close rate dropped, and Fin’s auto-replies on some of these tickets may have masked the real activity (Fin auto-reply does NOT count as movement on Vortex IQ tickets, which is correct).
- The B2C Storefront cluster has a single root cause. Seven findings stagnant since the senior agent went on PTO. The fix is not “close the findings”, it is “reassign the queue today”. Open Intercom’s Reports, Conversations, Group by Assignee and confirm.
- The B2B Portal SSO findings are higher-revenue-risk per finding. Two SSO findings on a B2B portal can cost more revenue per finding than the seven B2C-storefront findings combined; an SSO regression breaks login for an entire seat-pool. Triage these first regardless of count.
- Configure timezone-aware assignment. The B2B Portal engineering owner in India timezone rarely opens Intercom during US business hours; tickets sit untouched until next morning India time. Use Intercom’s Workflows, Round-robin to fall back to a US-timezone agent if the primary assignee has not opened the ticket within 4 hours.
- Pair with
int_vortexiq_findings_openandint_fin_resolution_rate. A high Fin resolution rate alongside a rising abandoned count means Fin is auto-replying to findings (which it should not); audit the Fin rules. - Right action: triage meeting today, not a hire. Run a weekly 30-minute “abandoned review” with one ops lead. Mark anything with no real merchant impact as
closedand move on. Reassign the rest into next sprint. We see Intercom merchants close 30 to 50% of abandoned findings on this triage; the remainder being correctly prioritised is the win.
Sibling cards merchants should reference together
| Card | Why pair it with Abandoned Findings | What the combination tells you |
|---|---|---|
| VortexIQ Findings Open | Abandoned is a subset of open. The ratio is the steadier read. | Abandoned ÷ Open above 30% means a third of the queue has gone dormant. |
| Finding Resolution Rate (90d) | Resolution rate drops first; abandoned count rises second. | The textbook capacity-collapse sequence. |
| Snoozed Conversations | Findings agents intentionally deferred. | Snoozed > Abandoned = team is using snooze to mask rather than action; investigate. |
| Fin AI Resolution Rate | Fin should NOT be replying to Vortex IQ tickets. | High Fin handling + high abandoned = Fin is masking abandonment with generic replies. Fix the rule. |
| Tickets by Assignee | Where abandonment concentrates. | If 70% of abandoned findings sit on one assignee, the fix is reassignment. |
| Avg First-Response Time | Intercom’s headline metric. | FRT rising + abandoned rising = team straining; FRT flat + abandoned rising = team prioritising shoppers over findings. |
| Refund Rate / Conversion Rate | The downstream truth metrics. | Abandoned >10 for 4+ weeks predicts a 0.05 to 0.15 percentage-point conversion-rate dip. |
| Customer Service Sentiment | NPS-side outcome. | Sustained abandoned >10 predicts CSAT drop of 3 to 5 points within a quarter. |
Reconciling against the vendor’s own dashboard
Where to look in Intercom’s own dashboard:
Intercom Inbox, build a saved view: tags:vortex_iq AND state:open AND last_admin_action_older_than:14d. Pin the view for the team. Intercom’s filter language supports a “last admin action” filter on most account tiers.
Reports, Conversations for trend visibility on the same tag filter.
Intercom does not ship a built-in “abandoned” view; this card is closest to the saved-filter pattern.
Why our number may legitimately differ from a saved Intercom view:
| Reason | Direction | Why |
|---|---|---|
| Time zone | Boundary day off | Intercom Reports use the workspace timezone; the Inbox UI uses each agent’s profile timezone. The card uses workspace timezone for the 14-day cutoff. |
| Snoozed-ticket inclusion | Either | The card excludes state:snoozed (intentionally deferred is not abandoned). Some Intercom views include them. |
| Customer-reply movement definition | Either | The card excludes customer replies as movement; Intercom’s filter primitives include them. A ticket where a customer replied 7 days ago counts as abandoned in the card if the team has not responded; Intercom’s filter would flag it later. |
| Fin auto-reply movement | Ours stricter | Fin auto-reply does NOT count as movement on Vortex IQ tickets in this card; Intercom’s “last admin action” filter usually counts Fin replies as admin actions. |
| Workspace scope | Either | Card aggregates across workspaces; Intercom Inbox scopes to one. |
| Tag drift | Ours stricter | We require literal vortex_iq. Legacy tags drop out. |
| Webhook delay | Up to 60s stale | A movement event from the last minute may not have reached the card. |
| Card | Expected relationship | What causes the divergence |
|---|---|---|
shopify.refund_rate / bigcommerce.refund_rate / adobe_commerce.refund_rate | Inverse correlation. Higher abandoned rate predicts a refund rate rise over a 4 to 8 week trailing window. | Abandoned on returns/checkout findings drives refunds up; abandoned on cosmetic findings only does not. |
shopify.ecommerce_conversion_rate | Inverse correlation. | Abandoned >10 for 4+ weeks predicts conversion-rate dip. |
shopify.customer_service_sentiment | Inverse correlation; 2 to 4 week lag. | Sustained abandoned predicts CSAT drop. |
datadog.dd_health_score | Independent peer; correlates only on technical findings. | Both elevated = engineering firefighting + ops falling behind. |
Known limitations / merchant FAQs
Intercom shows my ticket was modified yesterday but the card says abandoned. What is wrong? Almost always one of these. (1) The “modification” was a Fin AI auto-reply; the card excludes Fin auto-replies as movement (treating Fin auto-reply as movement would mask abandonment). (2) The “modification” was a customer reply; the card excludes customer replies. (3) Webhook delay: an edit from the last 5 minutes may not have reached the card. Open the ticket and check the activity feed for an explicit human-team-member action in the last 14 days. What counts as movement? Any of: agent reply, internal note, status change, assignee change, tag change, applied macro. Customer replies do NOT count. Fin AI auto-replies do NOT count. Manual snooze counts (it is an explicit team decision); auto-reopen from snooze expiry does NOT count (no team action occurred). The 14-day window feels arbitrary. Can we tune it? Yes, in Settings, Connectors, Intercom, Abandonment threshold (days). Default is 14 days. CX-led merchants on Intercom often tune this down to 7 days because their daily rhythm catches stagnation faster than other contexts. My team uses Intercom for both ecommerce and B2B. The count looks alarming. The card aggregates across all Inboxes in the connected workspace. Open the per-Inbox stack panel (configure two integrations or use saved-view filtering) to see which Inbox drives the abandoned count. Most B2C+B2B merchants find the abandoned bucket concentrates in B2B because B2B findings need engineering owners who don’t open Intercom often. Velocity dropped, abandoned count rising. What changed? Standard playbook. (1) Open Tickets by Assignee and look for sudden concentration; a key person on PTO is the single most common cause. (2) Open Avg First-Response Time; if FRT rose sharply, the team is being pulled into live conversations. (3) Open Snoozed Conversations; if snoozed count rose, the team is using snooze to mask rather than action. (4) Cross-check with Datadog incidents for engineering-firefight context. The abandoned count appears suddenly higher overnight. Why? Because the 14-day clock keeps ticking. If five findings were last touched on the same day exactly 14 days ago, all five become abandoned at the same time the next day. Normal and self-corrects within 24 hours of the team picking the queue back up. Should I close abandoned findings or fix them? Both, depending on triage. Run a weekly 30-minute “abandoned review” with one ops lead. For each: if no real merchant impact, close. If real merchant impact, reassign and put in next sprint. We see Intercom merchants close 30 to 50% of abandoned findings on this triage; the remainder being correctly prioritised is the win. Snoozed tickets are excluded by default. Why not count them as abandoned? Because snooze is an explicit team decision to defer; it represents triage, not abandonment. The risk is that teams use snooze as a “hide from view” tool rather than a deliberate defer. To catch that pattern, watch Snoozed Conversations: if snoozed count keeps growing without findings closing, snooze is being misused. Why doesn’t Intercom ship an “abandoned” view natively? Intercom’s filter primitives (last_admin_action_older_than, tag, state) let you build the abandoned view yourself, but the curated view is not pre-built. The card’s added value: cross-Inbox aggregation, customer-reply / Fin-auto-reply exclusion, threshold-based alerts, historical trend.
Is this card the right primary watch metric for Intercom-using merchants?
Yes. Findings Open tells you the queue size; this card tells you the queue health. Use Findings Open to size the team’s workload; use this card to detect execution-discipline regressions. Both are hero cards because they answer different questions; if you only watch one on an Intercom-using team, watch this one.
A finding important enough to fix manually but we never used Intercom to track it. Do we close the ticket?
Yes. Set the ticket to closed once the work is done. Otherwise it stays in the abandoned bucket. Vortex IQ does not auto-close findings just because the underlying audit signal cleared.
Can Intercom Workflows auto-close low-impact abandoned findings?
Yes. Build a Workflow: IF tags CONTAINS 'vortex_iq' AND last_admin_action > 21 days AND priority NOT IN [high, critical] THEN close conversation AND apply tag 'auto_closed_low_impact'. Auto-closes the bottom-tier abandoned findings on a 21-day timer, leaving high-impact ones for human triage. Most Intercom merchants who adopt this see abandoned-bucket size drop 40 to 60% within a month.
My team uses multiple Intercom workspaces. Does the card aggregate?
Yes by default; configure vortex_iq.workspace_filter if you want per-workspace cards. Most Intercom merchants run one workspace.