CSAT by Agent, Intercom

Positive-rating percentage broken out per Intercom admin over 30 days, ranked as a horizontal bar so you can see who customers rate highest and who needs coaching. How to read it, why it matters, and how to act on it.

Metrics type: Supporting Metrics • Category: Helpdesk / Customer support

CSAT (% positive, 4-5★ conversation ratings) split per Intercom admin over 30 days, ranked so you can spot the strongest and weakest performers.

At a glance

CSAT by Agent is a customer-satisfaction metric tracked from your Intercom workspace. It takes the same positive-rating percentage as the headline CSAT gauge and attributes each rating to the admin who handled the conversation, then ranks them as a horizontal bar chart. For a small Blitz support team this is the difference between knowing satisfaction is slipping and knowing whose queue it is slipping on. Use it to recognise your best agents, target coaching where it counts, and protect the customer experience as you scale headcount.

What it counts	For each Intercom admin (agent), the percentage of their rated conversations scored positive (4-5★) over the last 30 days. The rating is attributed to the admin recorded on the rated conversation — typically the assignee or the admin who sent the reply being rated. Agents with no rated conversations in the window are omitted.
Sample type	API-derived from the Intercom conversations endpoint joined to admins. Vortex IQ reads each conversation’s conversation_rating and its assigned/handling admin, then groups positive-share by admin.
Why it matters	Team-wide CSAT can sit at a healthy 88% while one agent quietly runs at 64% and drags the average. This card surfaces that variance so you can coach early, share what your top performers do well, and avoid routing high-stakes conversations to a queue customers consistently rate poorly.
Reading the value	Scan the ranked bars top to bottom. Anyone below the 70% alert line needs attention — but always read it next to that agent’s volume; a 60% score on five rated conversations is noise, while 60% on eighty is a pattern. Confirm volume against Conversations Handled by Agent.
Currency	percent
Time window	30D
Alert trigger	<70%
Sentiment key	csat
Roles	owner, operations

Calculation

For each admin, Vortex IQ counts their conversations rated 4-5★ in the last 30 days and divides by all their conversations carrying any rating (1-5★) in that window. The result is a per-agent positive share, sorted descending for the bar chart. The <70% alert fires for any agent whose positive share falls below the floor (subject to a minimum sample so a single bad rating does not trip it). Unrated conversations are excluded — an agent’s bar reflects only the customers who chose to rate them.

Worked example

A representative reading of CSAT by Agent for the Blitz support team. The ranked bars show Priya at 94% (61 rated), Tom at 91% (58 rated), Jordan at 88% (47 rated), and Sam at 66% (52 rated). Sam’s bar sits below the 70% line and trips the alert. Volume is comparable across the team, so this is not a small-sample artefact — it is a real gap. You open Sam’s negative-rated threads from Negative-Rated Conversations and find a pattern: fast first replies, but a habit of closing conversations before the customer confirms the issue is solved, which is generating low scores and reopens. The fix is coaching, not headcount. You pair Priya with Sam for a week and re-check the card. To trace whether the low scores cluster on a topic or a shift, open Vortex Mind; to ask “show me Sam’s 1-2★ conversations this month” in plain English, use Ask Viq.

Sibling cards merchants should reference together

Card	Why merchants reach for it
ic_csat_score	The team-wide headline this card decomposes — read the average, then find who moves it.
ic_csat_trend	When the team trend dips, this card tells you whether the slide is broad or one agent.
ic_handled_by_agent	Always read CSAT-by-agent next to volume so you do not over-react to a small sample.
ic_resolution_by_agent	Slow resolution and low CSAT often travel together — check both per agent.
ic_reopen_rate	A low-CSAT agent who also reopens a lot is closing conversations prematurely.

Reconciling against the vendor’s own dashboard

Where to look in Intercom’s own dashboard:

In Intercom, open Reports → Team performance (or Customer satisfaction with the teammate breakdown enabled) and set the range to the last 30 days. The per-teammate satisfaction column reports the same positive share of ratings Vortex IQ ranks here. You can also filter the Inbox by teammate and rating to read the underlying conversations.

Why the Vortex IQ value may legitimately differ:

Reason	Direction	What to do
Attribution rule. Vortex IQ attributes a rating to the conversation’s handling/assigned admin; Intercom may attribute by the teammate who sent the specific part being rated, or by current assignee.	Variable	Confirm how your workspace attributes ratings to teammates.
Reassigned conversations. When a conversation passes between agents, the rating may land on a different admin in each system.	Variable	For reassigned threads, check the assignment history.
Minimum sample. Vortex IQ may suppress or de-emphasise agents with very few rated conversations; Intercom shows all.	Marginal	Note the rated-conversation count beside each bar.
Time zone. Intercom uses the workspace time zone; Vortex IQ aligns to your merchant reporting time zone.	Marginal	Confirm time-zone match.

Cross-connector reconciliation: read this card alongside First-Response Time by Team and Workload by Team — an agent rated poorly while carrying double the team’s workload is a staffing signal, not a coaching one. For divergence investigations, use Vortex Mind.

Known limitations / merchant FAQs

Q: How often does CSAT by Agent update? The card refreshes on the standard data refresh (typically every 30-60 minutes). New ratings attribute to the relevant agent as they arrive.

Q: Why is an agent’s number different from Intercom’s teammate report? Almost always the attribution rule. Vortex IQ ties a rating to the handling/assigned admin; Intercom may tie it to whoever sent the exact replied-to part or to the current assignee. Reassigned conversations are the most common source of single-agent discrepancies.

Q: An agent shows a low score but only handled a handful of rated chats — should I worry? Not yet. Below roughly 15-20 rated conversations, one or two bad scores swing the percentage hard. Read the bar next to its rated count, and confirm volume via Conversations Handled by Agent before acting.

Q: Can I customise the alert threshold? Yes. The <70% floor is configurable per profile in the Alert Rules tab. Set it to match the standard you hold the Blitz team to rather than the generic default.

---

Tracked live in Vortex IQ Nerve Centre

CSAT by Agent is one of hundreds of KPI pulses Vortex IQ tracks across Intercom and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English.

Start for free or book a demo to see this metric running on your own data.