Ask who can help.
A sentence in, up to three colleagues out, each with one line of evidence from work they shipped. If nobody here has done it, Keegi says so.
LoadingAsk who can help.
A sentence in, up to three colleagues out, each with one line of evidence from work they shipped. If nobody here has done it, Keegi says so.
LoadingA sentence in, up to three colleagues out, each with one line of evidence from work they shipped. If nobody here has done it, Keegi says so.
A directory tells Keegi who reports to whom, where people are, and what time it is for them. Reachability and load spreading read it, so Keegi can avoid returning somebody who is asleep or already carrying four introductions this week.
It is not evidence and it never becomes evidence. A directory row says somebody exists and where they sit. It is not a claim that they have done anything, so it is never quoted in an answer. Classifying it Citable does not change that.
There is no Personio, BambooHR or HiBob connection. You upload a JSON file, and Keegi never holds a live link to your HR system.
A JSON array of people. id, email and displayName are required. Everything else can be left out. managerId points at another row's id, and that is what builds the org chart.
[
{
"id": "<id>",
"email": "you@company.com",
"displayName": "<name>",
"role": "<role>",
"timezone": "Europe/Berlin",
"location": "<city>",
"managerId": "<manager-id>",
"startedAt": "2020-01-01"
}
]Skills, competency ratings, compensation and performance fields are dropped rather than stored. You do not have to strip them out of the export first, and Keegi will not keep them if you forget.
On /app/[org]/sources, under Load a directory, choose the file and press Load. It is read in your browser, so a file that cannot be parsed is refused before anything is sent.
Keegi reports how many people it read and how many matched somebody it already knew about. That second number is the one to look at. Email is the only thing joining a directory row to a person, so a load that matches almost nobody means your HR system and your other tools disagree about addresses. The unmatched addresses are listed under that, so the pattern is visible. If one address pointed at more than one person, Keegi says so. It still joins that row; the number is there so you can see it.
Loading again merges rather than replaces, which is how a directory is kept current. Unmatched rows stay in the directory and do not create people.
The HRIS row is classified like any other source. Excluded stops Keegi using the directory, which turns reachability off: answers still come back, without the timezone and workload sense behind who is returned.