DOCApproved
99.9%
"The commitment is 99.9% monthly uptime for the FlowSight cloud platform."
FlowSight service terms: availabilityDaniel Reyes, 26 Aug
Cognition by corteXolutions
Cognition reads your documents, chat and meeting recordings, and keeps one map of what the company says. When two sources give different figures, it flags them before a customer sees either. Every answer carries its source.
Question: What uptime do we promise?
DOCApproved
99.9%
"The commitment is 99.9% monthly uptime for the FlowSight cloud platform."
FlowSight service terms: availabilityDaniel Reyes, 26 Aug
DOCDiffers
99.95%
"FlowSight runs with 99.95% uptime, backed by service credits."
Enterprise sales deck: trust slideLucas Moreau, 5 Sep
CHATDiffers
99.5%
"Last time I told a customer 99.5%."
Customer asks what uptime we promiseMateo Novak, 20 Sep
Three sources, three figures for the same promise.Compared with the approved figure. Two sources differ.
99.9%
Monthly uptime for the FlowSight cloud platform.
Nobody lied. Three people quoted three different documents.
Cognition reads what you already have. Nobody rewrites a wiki.
Documents, chat channels and meeting recordings. You choose the sources.
People, topics and figures, each linked to the sentence it came from.
Someone with authority signs off uptime, prices, headcount. That is the reference.
New sources are checked against it. Differences are flagged; answers show their source.
Each one ends at a source. Nothing is shown without a link back to it.
A question in plain English. A short answer, with every claim linked to its source, author and date. No source, no answer.
with the sourceWho actually knows a topic, ranked by what they wrote and answered, not by title.
by what they wroteA weekly feed of new sources, changed figures and newly opened conflicts. Back from leave, read one page.
this weekSame subject, same metric, different values. Each source listed with its date, and the approved value when there is one.
open and resolvedPaste a proposal or a deck. Each figure is compared with the ones leadership approved.
before you sendExperts: uptime and SLA
Fictional example.
What changed: week 2
Fictional example.
Paste the proposal. Cognition lists each figure as matched, different or not on file, so the fix happens before the send.
| Figure | In the draft | Approved | Result |
|---|---|---|---|
| Gateway lead time | 6 weeks | 11 weeks | Differs |
| Platform uptime | 99.9% | 99.9% | Matches |
| Active customers | 57 | 42 | Differs |
| Pro plan, monthly | $1,450 | $1,450 | Matches |
| Support response | 2 hours | none on file | Not on file |
2 differ. Lead time and customer count are older than the approved figures. Fix both, then send.
We run Cognition on a slice of your documents, chat and recordings under NDA, and bring you the conflicts it finds. You decide from that list.
One team, one set of channels and folders. You choose, we read only those.
People, topics and figures, linked to their sources.
The few figures that must never differ: uptime, price, headcount.
Walked through with your ops team, source by source.
It does not. A person with authority approves a figure, and Cognition compares everything else against it. With no approved figure, it shows the conflict and the sources, and waits.
Documents, chat and meeting recordings. We agree the list with you during scoping, and we read only what you point us at.
A pilot is scoped to sources you choose. Per-person access rules are part of what we settle with you before any wider rollout.
No. Search returns documents. Cognition compares the figures inside them and tells you where teams disagree.
Agreed after the pilot, once you have seen the conflict list on your own sources.