Analytics
Everything this workspace keeps a record of, and what each record can answer. Pick a channel and a range once; they carry into every view you open from here.
What this workspace can measure
11 measures · 8 reporting today · 3 recorded but not yet reported
1,842 messages, comments and conversations arrived. Instagram carried 1,104 of them, WhatsApp 612, Facebook 126.
Every message stores the minute it arrived and the minute we answered, so this is countable now. Nothing reads it yet.
1,253 of 1,842 interactions closed without a person. 405 waited for review and 184 went to a human.
405 drafts reached review and 402 are decided. Approvals ran three to one over rejections, median wait 41 minutes.
214 of 252 answers quoted at least one document. The other 38 found nothing and the agent answered from the brief alone.
The agent hid 74 comments, deleted 9 and replied to 188. A person overrode it 6 times, all unhides.
No campaign was broadcast in this range. All 1,036 outbound sends came from automation flows instead.
Every flow session and every step it passed is already stored, so drop-off per step is countable. Nothing reads it yet.
1,269 contacts are on file, each with the channel it came from and its opt-in state. Nothing charts how that number grew.
4 of 1,036 sends failed, all on numbers that no longer exist. WhatsApp is connected, quality green, on the 1,000 a day tier.
18.42 dollars of model spend, averaging 20 cents a day. Cost for each captured lead stays blank until leads are attributed.
Every figure is counted from this workspace's own records: conversations, messages, delivery receipts, agent turns, moderation acts, campaigns and contacts. Nothing is estimated and nothing is pulled from a third party except the WhatsApp quality rating, read from Meta on 29 Jul at 06:10.
Move with j and k, open with Enter, come back with u.
Analytics · Agent outcomes
Agent outcomes
Handled alone means the agent replied or acted and no person touched it. The channel and range you picked on the map still apply.
1,253 of 1,842 interactions closed without a person. 405 were held for review before sending and 184 were passed to a human. The agent chose to stay silent 297 times, mostly because the comment was not a question.
Per situation
Sorted by how much work each situation carried| Situation | Interactions | Handled alone | Held for review | Passed to a human | Share handled |
|---|---|---|---|---|---|
| Comment on a post | 908 | 701 | 158 | 49 | 77% |
| Buyer enquiry | 514 | 331 | 129 | 54 | 64% |
| Job seeker | 263 | 178 | 63 | 22 | 68% |
| Trigger reply | 148 | 43 | 55 | 50 | 29% |
| Everything else | 9 | 0 | 0 | 9 | 0% |
| All situations | 1,842 | 1,253 | 405 | 184 | 68% |
Why it stayed silent, and how often it was right
297 times the agent read something and chose not to answer.
Your verdicts on the agent's decisions, collected since 12 Jun.
Thin on purpose. Labels are collected by hand from Activity, so a low share is expected and does not weaken the figures above.
Loading. Row-shaped skeletons, never a spinner. The group headers and the tally line are the last things to fill, so the map's shape is stable while it loads.
What this workspace can measure
Loading the coverage map
Error. Full width, persistent, carries the way out. It names what failed and says plainly that nothing was lost.
Partial. One channel is missing from every row, so the map says so once at the top rather than eleven times.
Offline. The last good figures stay on screen with their age stated, instead of the page emptying itself.
Filtered to zero. A channel with no records in the range. The measure keeps its place in the map, states why the figure is zero, and offers the filter back, never a create-first prompt.
First run. A workspace connected today. The map is still the full eleven measures, so what is coming is visible before any of it has data.
What this workspace can measure
11 measures · 0 reporting today
WhatsApp connected 40 minutes ago. The first figures appear here as soon as a message, comment or send is recorded.