Reports
In the app: Reports
Needs the 'View Reports' permission.
Reports, titled "Reports & Analytics" on the page, shows how support is going: ticket volume, speed, the team's and each teammate's work, service levels, customer satisfaction and what the AI did. This page holds nine tabs; the other tabs in the strip are reports with a page of their own.
Period, timezone and filters
These controls sit above the tabs and apply to every tab on this page except "Live Agents", which shows the present moment.
Period. Two date fields set the first and the last day; both days are included. "7d", "30d" and "90d" set the start to 7, 30 or 90 days before today and include today too: 8, 31 or 91 calendar dates. The page opens on "7d". A period can be at most 366 days and cannot end after today. The volume and AI trend charts show one point per day when the start and end dates are at most 60 days apart, and one per week otherwise. The SLA and CSAT history charts always show days.
"Times in". Chooses the timezone that decides where a day starts and ends: the organization's "Report timezone" from Settings → Organization (the default) or the timezone of your own device. It applies to the period, to every chart by day or hour, and to exports. With your device's timezone selected, days can differ from the ones your SLA policies and business hours use.
Filters. "Stores", "Agents", "Teams", "Tags", "Channels" and "AI" narrow the tickets every number is counted over. Each filter can hold several values and can be set to "Is one of" or "Is not one of". "Clear" removes them all.
- "Agents" has two meanings. For numbers about tickets (how many came in, how many are closed) it means tickets now assigned to that person. The Overview first-response average and the Agent Performance reply figures use the reply author, whoever the ticket is assigned to. Team Performance, Busiest Times and the AI tab use the ticket's assignee instead.
- "AI" → "AI handled" keeps tickets on which the AI wrote at least one reply or draft; "Non-AI" keeps the rest. On the "CSAT" tab this filter becomes who answered the customer: "Entirely AI", "Human Agent Involved" or "No AI-written reply".
Ticket exclusions. Most figures exclude deleted tickets, tickets merged into another one, the example ticket a new workspace starts with, and outbound campaign tickets the customer never answered. Macro insertions do not apply these exclusions. Live Agents excludes deleted tickets from its open-ticket count, but does not apply the other exclusions. AI run cost and tokens have their own exception, explained under "AI". The inbox does count the example ticket, so its totals can be one higher.
A teammate whose access is limited to certain stores only sees those stores' numbers, except on "Live Agents". The period, the view, the timezone and the filters are kept in the page address, so a link or the Back button returns to the same report.
"Refresh" reloads the current tab.
Advanced and Compact view
"Advanced" is the full report with tabs and filters. "Compact" is one table that compares your stores, called "Volume by store". Switching to "Compact" clears the filters.
The compact table lists stores with tickets or customer messages in this period or the previous one, with an "All stores" total, for the period and compared with the same number of days before it:
- "Tickets": tickets created in the period, with the change against the previous period.
- "Messages": customer messages received in the period, on any ticket, with the change against the previous period.
- "Closed": of the tickets created in the period, the ones that are closed now.
- "Open now": of the tickets created in the period, the ones that are open now.
- "Waiting 4h+": tickets created in the period that are open, whose newest message is from the customer, and that have been waiting for a reply for 4 hours or longer. This is a fixed 4-hour rule; it does not use your SLA policies.
Click a store's row to open its details, all over the tickets created in the period:
- "First response": the average time from a ticket's creation to its first reply, over the tickets that got one.
- "Resolution": the average time from creation to closing, over the tickets that are closed now.
- "Msgs / ticket": the average number of messages on those tickets, counting every message on the ticket.
- "Handled by AI": the share of those tickets on which the AI sent the customer at least one reply, by itself or as a draft a teammate approved.
- "Open … unresolved tickets" opens the list of the tickets behind that store's "Open now" number.
Click a column heading to sort; the table starts sorted by "Tickets". The "Created vs closed" chart under the table shows, per listed store, tickets created per day or week and tickets closed per day or week. In this chart a close is counted on the day it happened, also for a ticket that was created before the period, provided it is still closed. Click the store controls to hide or show charts; hiding all stores shows them all again.
Export CSV and ticket lists
"Export CSV" downloads selected figures from the current tab, for the current period and filters, or the store table in Compact view. It does not include every chart. The Overview file includes both resolution-time averages, whichever one the tile shows. On CSAT it exports the rating distribution and the current page of comments; the positive, neutral and negative comment filter does not narrow the file. The file states the timezone its dates are in. The button is disabled until the tab has loaded.
Ticket lists. Some numbers open the tickets behind them: "Created Tickets", "Closed Tickets", "Open Tickets" and "Snoozed Tickets" on "Overview", "Replied" and "Messages" in the "Agent Performance" table, and "Open now" in the compact view, through "Open … unresolved tickets" in a store's details. The list shows the ticket, its status, assignee and creation time, 20 per page with page buttons, and each ticket opens in the inbox. "Download all matching tickets" downloads up to 5,000 matching tickets as a CSV file. It is not limited to the page on screen, but a larger list is cut off at 5,000.
Overview
The tiles at the top. Unless stated otherwise, they are counted over the tickets created in the period.
- "CSAT Score": the average rating out of 5 of the surveys customers answered in the period, with the number of responses. A rating counts in the period it was given, also when its ticket is older.
- "Resolution Time": the average time from the creation of a ticket to its close. The switch on the tile chooses between every ticket that is closed now ("Average time to close") and only tickets that are resolved now: the customer confirmed, the resolve follow-up ran out, or a teammate pressed Resolve.
- "Messages / Ticket": the average number of messages on a ticket, in both directions. Internal notes are counted as messages here. Messages sent after the period on those tickets count too.
- "First Response": the average time from the creation of a ticket to the first reply that was sent, by the AI or by a person. Tickets without a reply are left out.
- "Created Tickets": tickets created in the period.
- "Closed Tickets", "Open Tickets", "Snoozed Tickets": of the tickets created in the period, how many are closed, open or snoozed right now. "Closed Tickets" is not the number of tickets that were closed during the period.
- "Messages Received": messages customers sent in the period, on any ticket, also on tickets created before the period. The tooltip on the tile says otherwise; this is what is counted.
All averages are plain averages, not medians, so a few very slow tickets move them.
"Resolution" is a table of the tickets created in the period that were marked resolved at least once, split by who resolved them: "AI" (the customer confirmed, or the resolve follow-up ran out without a reply), "Agent" (a teammate pressed Resolve) and "Total". "Resolved (% of tickets received)" is the count and its share of all tickets created in the period. "Reopened (% of resolved)", marked "Test metric", is how many of the resolved tickets the customer came back on: the ticket is no longer resolved, or it needed another reply before it was resolved again.
"Conversation Volume" shows two bars per day or week. "Created" is the tickets created in that day or week. "Closed" is, of those same tickets, the ones that are closed now and were closed inside the period. A close is therefore shown on the day its ticket was created, not on the day it was closed. For closes by the day they happened, use the "Created vs closed" chart in the compact view.
"SLA Compliance" shows the first response and resolution rates explained under the "SLA" tab. Targets that are still running are left out.
"Channel Breakdown" is the tickets created in the period per channel.
"WhatsApp messages" only appears when a store in view has a WhatsApp number, or had one or WhatsApp traffic in the period. "Sent by AI", "Sent by agents" and "Templates" count messages Meta delivered in the period; "Failed" counts messages that could not be delivered. The figures are added up from a count kept per store per day. Only the store filter and the period apply to it, and its days are UTC days. At most 365 dates are shown; longer periods are trimmed to their final 365 dates, and the end cannot go past today.
"Top Tags" is the ten tags that are on the most tickets created in the period, with the number of tickets each is on.
Team Performance
Figures for the people on the team together. A reply by a person is a message a teammate wrote, or an AI draft a teammate approved and sent. Campaign sends are not counted.
- "Tickets Replied": tickets on which a teammate sent at least one message in the period, whenever the ticket was created.
- "Messages Sent": messages teammates sent in the period. Internal notes written by a teammate are counted in these two tiles as well.
- "Avg Response Time": for customer messages sent in the period on tickets created in the period, the average time until the next message a person wrote on that ticket. An internal note counts as that next message, so a note written before the real reply shortens the figure. A reply sent after the period ended still counts. Customer messages that never got a message from a person are left out.
- "Avg Handle Time": the average active time per ticket, explained below. "Unavailable" means it could not be calculated; try again later.
- "One-Touch Tickets": tickets created in the period that are closed now and got exactly one reply from a person. Internal notes are not replies here.
- "Closed with no human reply": tickets created in the period that are closed now and never got a reply from a person. Many of these are tickets the AI handled alone.
Of the filters, "Agents" narrows these figures by the ticket's assignee.
What counts as active time in "Avg Handle Time". Active time is measured in the inbox while a teammate works in a ticket. It counts when three things are true at once: the ticket is open in the inbox, the browser tab is visible, and the teammate typed, clicked or scrolled inside the ticket in the last 2 minutes. It stops when they open another ticket, hide the browser tab, or leave the ticket untouched for 2 minutes. Scrolling the ticket list or working elsewhere in Squid does not count as working the ticket.
Opening a ticket starts the activity window, so reading it counts for up to 2 minutes without another interaction. Bringing another window in front does not by itself stop tracking if the browser still considers the tab visible. The time of every teammate who worked in the ticket is added together, over the whole life of the ticket and not only inside the period. The tile is the average of that total over the tickets created in the period on which any active time was measured. Tickets no teammate opened are not in the average, and the AI's work is never handle time. The same measured time feeds "People needed" under Reports → Forecast.
Agent Performance
Active teammates are included even with nothing to show for the period, subject to the Agents and Teams selections. Other people can also appear when they have matching assignments or recorded work, including people deactivated since. The AI is not a row. Rows are ordered by the number of tickets assigned.
- "Avg CSAT": the average rating given in the period on tickets assigned to this person, with the number of ratings as "(n=…)". A warning sign marks fewer than 10 ratings.
- "First Response": on tickets created in the period where this person wrote the first message by a person, the average time from the creation of the ticket to that message. An internal note can be that first message; it need not have been sent in the selected period.
- "Resolution": the average time from creation to close of the tickets created in the period that are assigned to this person.
- "Avg Response": for the replies this person sent in the period on tickets created in the period, the average time since the most recent preceding customer message. Internal notes count as messages here, and several messages by the person after one customer message each contribute to the average.
- "Closed": tickets this person closed themselves in the period and that are still closed, whoever they were assigned to. A ticket the customer reopened since is not counted. Closes by the AI, by a rule or by an automatic sweep belong to nobody.
- "Resolution Rate": "Closed" divided by the tickets created in the period that are assigned to this person. Because the two are counted differently, it can be above 100%.
- "% of Closed": this person's share of all the closes in the table.
- "Replied" and "Messages": tickets this person sent a message on in the period, and the number of messages. Approved AI drafts count for the person who approved them. Internal notes also count. Click a number for the tickets.
Labels next to a name: "No activity in period" (no replies and no closes anywhere in the period), "Not counted here" (they worked in the period, but on tickets the current filters exclude), "Activity unknown" (it could not be checked), and "Pending invite" or "Invite expired" for someone who never accepted their invitation. The line above the table sums these up, and "Manage invites" opens Settings → Team.
Busiest Times
A grid of the days of the week by the hours of the day, in the timezone chosen under "Times in". For the three counts, darker means more. For the two times, the colour runs from green for the fastest hours to red for the slowest. The menu chooses what is counted:
- "Conversations": tickets created in that hour.
- "Messages Received": customer messages sent in that hour, on any ticket.
- "Messages Sent": messages written by teammates in that hour, on any ticket. Internal notes are included. AI replies are not, and neither are AI drafts a teammate approved and sent, so this view can show fewer messages than "Messages Sent" on the "Team Performance" tab.
- "First Response Time": the average minutes to the first reply, by the AI or a person, for tickets created in that hour.
- "Avg Response Time": for tickets created in the period, the average minutes between a customer message sent in that hour and the message a teammate wrote directly after it. Both messages must fall inside the period. Internal notes can count as the next message. Customer messages followed by an AI reply or by another customer message are left out.
Live Agents
Who is working right now. This tab ignores the period and the filters, and it is the one tab that is not narrowed to a teammate's stores: it covers every teammate in the organization.
- "Online", "Away", "Offline" and "Pending Invite" count teammates per state. Someone who never signed in, or who has not been seen for twice the organization's idle time (20 minutes with the standard idle time of 10 minutes), is shown as "Offline" whatever their last status was. "Pending Invite" is people who were invited and never accepted.
- Each card shows the teammate's state, when they were last seen, "Open Tickets" (open tickets assigned to them) and "Workload" (their current load against their maximum, from the capacity limits under Settings → Organization and Settings → Team). The bar turns yellow from 70% and red from 90%.
The AI is not shown.
SLA
Results of your SLA policies for tickets created in the period that had a policy applied. Tickets without a policy are not counted. "View SLA Policies" opens Settings → SLA Policies.
- "First Response SLA": "Met" is tickets that got their first reply before the due time. "Breached" is tickets whose first response due time passed without one. Tickets that are still waiting inside their target are shown as still awaiting an outcome and count as neither.
- "Resolution SLA": "Met" is tickets that were closed without breaching their resolution target. "Breached" is tickets that passed it. Open tickets inside their target count as neither.
- The percentage on each card is met divided by met plus breached.
- "Achievement Rate": the plain average of the two percentages, rounded to a whole percent. When only one of the two has an outcome, it is that one's percentage. It shows a dash until at least one target has an outcome.
- "Breached SLA targets": first response breaches plus resolution breaches. A ticket that breached both counts twice.
- "SLA Over Time": met and breached targets per day, both kinds added together, so one ticket can add up to two. The chart is always per day. Each ticket is placed on the day it was created, not the day the target was met or breached.
Macro insertions
How often macros were used.
- "Macro insertions": the number of times a teammate inserted a macro into a reply in the period, with the count for the previous period of the same length.
- "Change": the difference against that previous period, as a percentage. It shows "N/A" when the previous period has no insertions.
- "Macro insertions by macro": the count per macro. An insertion counts when the macro is inserted, also when the reply is edited afterwards or never sent.
Only the "Stores" and "Teams" filters apply on this tab. Deleting a macro removes its recorded insertions from these figures. The percentage in the CSV file is each macro's share of the period's total insertions.
CSAT
Customer satisfaction ratings, from 1 to 5 stars. Whether and when surveys are sent is set under Settings → CSAT. Handling labels and the "AI" filter use replies up to when the survey was requested, or up to the rating itself for chat ratings without an email survey. Later replies do not change that classification.
- "Rating Distribution": the ratings received in the period per number of stars, with the total. A rating counts in the period it was given, whichever period its survey was sent in.
- Under it, the survey line. It counts tickets whose survey email was created in the period: how many were emailed successfully, and how many of those have a rating, whenever the rating was given. "Queued" and "Failed or bounced" are shown when there are any. "Most recent send among the surveys in this period" can be later than the period's end, because a survey created in the period may be sent afterwards. The last line counts the tickets on which a rating prompt was shown in chat in the period; when the period starts before the first recorded chat prompt, it says so instead of giving a count. These survey counts are hidden while the "AI" filter is set, because sends cannot be narrowed by it.
- "CSAT Over Time": the average rating per day. This chart only includes ratings on tickets that were also created in the period, so it can show fewer ratings than "Rating Distribution".
- "Customer Feedback": ratings given in the period that came with a comment, newest first, 20 per page. Each shows the stars, the customer, a label for how the ticket was handled ("Entirely AI", "Human Agent Involved" or "No AI-written reply"), the comment and a link to the ticket. "Positive (4-5)", "Neutral (3)" and "Negative (1-2)" filter the comments on the page you are on; use "Previous" and "Next" to go through the other pages.
AI
What the AI did in the period.
- "Closed, AI-Handled": of the tickets closed in the period that are still closed, the share where the AI answered the customer, no teammate is recorded as the closer, no person is currently assigned, and the ticket is not marked as handed to a team. A ticket is counted by the day it was closed, whenever it was created. Older tickets whose closer was never recorded can qualify when the other conditions hold.
- "Confirmed AI Resolution Rate": closed, AI-handled tickets that are resolved now, because the customer confirmed or did not reply to the AI's follow-up, as a share of the same closed tickets. A ticket reopened since is not counted.
- "Messages by AI": replies written by the AI as a share of all replies sent to customers in the period, by the AI and by people. Internal notes, campaign sends, templated back-office emails and answers from the chat widget's saved FAQ are not AI replies. A reply sent as several chat bubbles counts once.
- "Autonomous AI replies" and "Human-approved AI replies" split "Messages by AI" into replies the AI sent by itself and AI drafts a teammate approved and sent.
- "Manual AI Assignments": tickets a teammate assigned to the AI by hand in the period, as a share of the tickets created in the period. Assignments on older tickets also count in the numerator, so this share can exceed 100%.
- "Who handled the ticket": every ticket created in the period, in three groups: "Handled by AI alone" (the AI replied and no person did), "AI and a person" (both replied, including a ticket whose only reply was an approved AI draft) and "No AI-written reply" (a person replied, only a saved FAQ answer went out, or nothing was sent). The last group does not mean a person resolved the ticket.
- "AI Usage & Cost": "Model run cost" and "Tokens", explained below. It is not your invoice: Squid bills per AI message, see Settings → Billing. "Cost per closed ticket" divides the cost of every run, drafts and handoffs included, by the number of closed, AI-handled tickets.
- "AI Performance Trend": "Closed, AI-Handled" and "Messages by AI" per day or week.
- "AI Activity Summary": the same four shares as bars with their counts.
How "Model run cost" and "Tokens" are added up. Each time the AI runs, Squid records the tokens it read ("In") and wrote ("Out") and the price of that run in US dollars. "Tokens" is the sum of both over every recorded run in the period, and "Model run cost" is the sum of the recorded prices. Every kind of run counts, whatever came of it: replies that were sent, drafts that were never sent, sorting a ticket into a category, and handoffs. A run counts in the period in which it ran, also on a ticket created earlier, and it belongs to the store of its ticket.
Two things differ from the rest of the tab. Runs on deleted, merged and example tickets are included, unless "Agents", "Teams", "Tags", "Channels" or "AI" is set; then only the runs on the tickets those filters keep are added up. And a run that used tokens without recording a price adds its tokens but no cost: a line under the tiles says how many such runs were left out of the cost. Squid never fills in a price for them.
"About these numbers" under the tiles gives the number of saved FAQ replies and what is excluded. A fuller AI dashboard is under AI → Performance.
Other report tabs
The tab strip continues with reports that have their own page: "Refunds", "Ticket Fields", "Return Requests", "Returns Analytics", "AI Handoffs" and "NPS". Organizations with Squid Agent switched on also see "Contact reasons", "AI Tools", "CS economics", "Investigations" and "Forecast". Each has its own period and filters; the ones set on this page do not carry over.