Elbuz builds the support report from the tickets your agents work on: who is overloaded, how fast they reply, where deadlines slip, and how many orders followed a reply.
When you need this:
- The queue grows for no clear reason — “Right now” shows open and overdue tickets per person.
- A number is bad, but whose is unclear — clicking a breakdown row opens those tickets.
- Comparing two months — “Period…” takes any dates up to a year.
- What does support do for sales? — “Orders after tickets”: orders within 3 days of a reply.
Where to start: “Tickets” window, view button → “Reports” → “Right now” block.
What the reports do
- “Right now” — each assignee's workload at this minute; clicking a number opens the tickets.
- Any period — 7, 30, 90 days or your own dates up to a year.
- Tiles — new and solved, first reply, resolution time, compliance, customer rating.
- “Solved without a human” — chats and emails closed by AI; an email counts only after the customer has been silent for 72 hours.
- “Orders after tickets” — customer orders within 3 days of a reply, per assignee.
- Breakdowns — by assignee, channel, reason and department; a row opens the tickets in the list.
- Customer rating per assignee — the share of “Good” for each agent.
Who is overloaded right now: the “Right now” block
The first block of the report is the workload at this minute, not for a period: the period chips don't change it. Each row is an assignee; “Unassigned” comes first.
| Column | What it counts | Clicking the number |
|---|---|---|
| “Open” | tickets in “New” and “Open” | these tickets in the list |
| “New” | tickets in the “New” status — no reply yet | “New” only |
| “Overdue” | tickets past their SLA deadline, including “On hold” | this assignee's overdue tickets |
| “Longest waiting” | how long the customer who has waited longest for a reply has been waiting | the list, with the longest-waiting customer first |
The list opens with the caption “Selection: …” — a one-off selection, not a saved queue. Zeros are not clickable. The “Team” section in the left queue panel shows the same per-person numbers (working with tickets).
1
Unassigned row2
Open count3
Overdue count4
Longest waiting- Tickets nobody has picked up.
- Click — these tickets open in the list.
- Overdue SLA deadlines, including “On hold”.
- How long the customer who has waited longest has been left without a reply.
How to choose a period
Above the tiles are the chips “7 days”, “30 days”, “90 days” and “Period…”. The default is 30 days. A period starts at midnight on its first day and includes today: “7 days” on a Wednesday runs from last Thursday to this minute.
“Period…” opens two date fields and the “Show” button. Both days are fully included. Use it to compare specific months, or the week before and after a settings change. The longest period is one year: if you set more, the report shows the year ending on the “To” date.
What each tile means
Most numbers are counted over tickets created in the period. A ticket that came in a month ago and was solved yesterday only gets into the solved counter for “30 days”. The median means half the tickets are faster and half slower; time is calendar time, nights and weekends included.
| Tile | How it's counted | What it leaves out |
|---|---|---|
| “New tickets · N solved” | Created in the period; “solved” means set to “Solved” in the period, whenever they came in. Solved can exceed new: the team is clearing an old backlog | Tickets set to “Closed” without “Solved”; tickets the customer reopened to “Open”, until they are solved again |
| “First reply” | Median from the moment the ticket appears in Elbuz to the first reply | Tickets with no reply yet |
| “Resolution time” | Median from the moment the ticket appears to “Solved”, including tickets later closed | Tickets set to “Closed” without “Solved” |
| “SLA met · N breached” | Share of tickets with no “SLA breached” event — for first reply, next reply or resolution | — |
| “Customer rating (CSAT)” | Share of “Good” among ratings in the period, “👍 N · 👎 M · asked K” (details) | Tickets with no rating request and no rating |
An email received Friday evening and answered Monday morning adds over 60 hours to the median, though the SLA deadline wasn't missed: SLA counts only working minutes. For an email, the clock starts when the ticket appears: with automatic AI email analysis, after the analysis or 10 minutes later; without it, at the next ticket collection.
A ticket that is still open but not yet past its deadline counts as met, so the SLA percentage for the last few days always looks better than it will in a week. Compare periods that have already ended. A breach gets into the report even when SLA notifications are turned off.
1
Period chips2
First reply3
SLA met4
Solved without a human5
Customer rating (CSAT)- The period includes today; “Period…” takes your own dates, up to a year.
- Median in calendar time, nights included.
- Share of tickets without a single breach.
- Chats and emails closed by AI without a staff member; an email counts if the customer stays silent for 72 hours.
- Share of good ratings among all ratings, and how many were requested.
How many questions AI closed without a human
The “Solved without a human: chats X, emails Y” tile is stricter than its name suggests.
- Chats. A website chat answered by the bot or an AI employee: no human employee wrote to the customer, the bot didn't hand it over to a person, no ticket was created. The customer must have been silent for at least 3 days, so chats from the last three days are not counted yet. Messenger conversations are not included.
- Emails. AI replies sent to the customer automatically under the “Act on its own” trust level in “Pulse”. An email counts when the customer hasn't written in the same thread for 72 hours after sending. If they write again, even just to say thanks, it doesn't count.
Below the tiles are two explanation lines:
- “N more AI replies were sent by staff…” — AI drafts that a person accepted or edited; they are not included in the tile;
- “An AI email counts if the customer did not reply in the same thread for 72 hours. Still checking: N, the customer came back (not counted): M.” — appears when N or M is above zero. A large M is a reason to reread what AI replies to such emails (AI in tickets).
The percentage on the tile is AI resolutions divided by the sum of AI resolutions and new tickets. This tile covers the whole company, ignoring “Data scope”.
How many orders followed tickets
The “Orders after tickets” block shows the amount and number of orders — “Orders within 3 days of a reply: N” — and a table by assignee with the “Orders” and “Amount” columns. How it's counted:
- it takes tickets whose first reply was given in this period — by an employee or by AI;
- orders from the same customer placed within 3 days of the first reply;
- canceled and returned orders are not counted;
- an order is credited to one ticket — the nearest one before it — and to that ticket's assignee;
- tickets without a linked customer are not counted; the amount is in the default currency.
These are orders after a ticket, not a proven result of support: the customer might have bought anyway. Compare assignees and periods with each other, and don't present the amount as what support earned.
How to find the weak spot: charts and breakdowns
- “New and solved by day” — bars show how many tickets were created per day, the line shows how many were solved. If the line stays below the bars for a week, the queue is growing.
- “In progress now” — a donut with the “New”, “Open”, “Pending” and “On hold” statuses. It's a snapshot of this minute, not of a period.
The four breakdowns — “By assignee”, “By channel”, “By reason (AI)”, “By department” — are counted over tickets created in the period. Columns: “Tickets”, “Solved”, “First reply” (median), “SLA met”; “By assignee” also has “Customer rating (CSAT)”.
Rows are sorted by number of tickets. Tickets without an assignee go to the “Unassigned” row; without a department or reason, to “Not set”. The reason is set by AI analysis; chats and calls often don't have one.
Clicking a row opens that breakdown's tickets for the period in the list, captioned “Selection: ‹row› · ‹dates›”. “Not set” rows in channels, reasons and departments are not clickable; “Unassigned” in assignees is.
1
Solved line2
Status ring3
Unassigned row4
CSAT cell- A line below the bars for a week means the queue is growing.
- A snapshot for this minute, not for the period.
- Tickets nobody has picked up.
- Share of “Good”; in brackets good / bad.
What to decide based on a number
| What you see | What it means | What to do |
|---|---|---|
| The solved line stays below the new bars for a week | The queue is growing | Check “Unassigned” in “Right now”: if it's large, review distribution; if it's small, you need another person or more replies without a human |
| In “Right now”, one assignee has more “Overdue” than the others | The person is overloaded or doesn't see new tickets | Click the number and hand over some tickets; set a limit |
| “First reply” is long but SLA is high | Tickets come in outside working hours | If customers expect replies on weekends, add a day to working hours |
| “Resolution time” shows a dash or a random number | Tickets are set to “Closed”, skipping “Solved” | Agree to set “Solved” — Elbuz sets “Closed” by itself after 4 days |
| SLA is dropping, mostly on resolution | Tickets sit in “On hold” or “Open” while waiting for the customer | Set “Pending”; still can't keep up — raise the deadline (statuses and deadlines) |
| One channel is slower than the others | Nobody watches that channel | Assign its tickets with a process (example) |
| One reason takes a large share | A common question | A knowledge base article and a macro (AI in tickets) |
| “Solved without a human” is zero | The chat bot is off or hands over to a person, and emails have no “Act on its own” trust | Check the chat and “Pulse” settings |
| CSAT dropped | Almost always one or two people | The rating column in “By assignee” (how to read it) |
Scenarios
A manager's morning check
What to do. Open “Reports” and look at “Right now”. What you get. You see who has overdue tickets and how long customers have been waiting; clicking a number opens a list where you can hand tickets over to a colleague with a bulk action.
Compare a month before and after a settings change
What to do. “Period…” — September, then October. What you get. You see how first reply and SLA changed after, for example, turning on a limit. Compare periods that have ended.
Which channel is slow
What to do. The “By channel” breakdown; click the row with the longest first reply. What you get. A list of exactly those tickets for the period: you can see who took them and where they got stuck.
What support does for sales
What to do. “30 days”, the “Orders after tickets” block. What you get. The amount of orders within 3 days of a reply for each assignee — for comparing people and months, not as a proven result.
What to do when the numbers look odd
| What you see | Why | What to do |
|---|---|---|
| “Resolution time” shows a dash even though tickets are being closed | Tickets are set to “Closed”, skipping “Solved” | Set “Solved” (tiles) |
| An agent sees different numbers in the report than the manager | The report applies their department's “Data scope” | This is by design; the “All” scope gives the full picture (access) |
| SLA for last week got worse although there were no new tickets | Open tickets that counted as met have since gone past their deadline | Compare periods that have ended (tiles) |
| The “Not set” row in a breakdown doesn't open the list | An empty channel, reason or department can't be expressed as a filter | Open “By assignee” or the “All tickets” queue (breakdowns) |
| “Orders after tickets” is zero even though customers buy | Tickets are not linked to a customer, or the order came more than 3 days after the first reply | Link the customer in the ticket (how it's counted) |
FAQ
Why is “solved” higher than new tickets?
“Solved” counts all tickets solved in the period, older ones included; new counts only tickets created in the period. The team is clearing an old backlog.
The median first reply is 14 hours but SLA is 98%. How is that possible?
The median counts calendar time, SLA counts only working minutes (tiles).
Why are there zero emails “without a human” for yesterday when AI replied on its own?
An AI email counts after 72 hours with no customer reply in the same thread. Recent ones are in the “Still checking: N” line.
Why does an agent see the same “Solved without a human” as the director?
This tile covers the whole company; all other numbers cover only the tickets the employee can see.
Can I export the report to a spreadsheet?
No, the report has no export button. You can open a breakdown's tickets in the list by clicking a row.
Why doesn't the “Right now” block change when I pick a period?
It shows the workload at this minute, not for a period (“Right now”).
Where to find them and who has access
- Menu
- Directories → Contacts → Tickets; or Organizer → Tickets
- Reports
- the view button above the list → “Reports”; the window remembers the selected view
- Access rights
- anyone who can open “Tickets”, within their department's “Data scope”
- Prepare
- working hours and SLA deadlines — how to set them
An agent with the “Own only” scope sees their own tickets and unassigned tickets in the report, not the whole team (who sees tickets). Merged tickets are not included: the ticket they were merged into is counted.
Related topics
- If you need to know which agents reply poorly — How to collect customer ratings (CSAT).
- If the report shows overload — How to assign tickets to employees.
- If deadlines are missed because the targets are unrealistic — Working hours and reply deadlines (SLA).
- If “Solved without a human” is zero — AI in tickets.
