How Triad uses your data

Triad can observe something most systems cannot: how partner relationships behave over time. That raises a reasonable question about what we do with that information once more than one company is using the platform. This page sets out the answer in full.

We do not train on your content

The notes you write, the transcripts you forward, and the correspondence you log are not used to train any model, for your organization or for anyone else's, and they are not used in any analysis that spans customers.

What we analyze is the shape of activity: how often, how recently, across how many people, moving in which direction, ending in what outcome. Numbers and categories. Content is excluded before analysis begins, which means de-identification is a consequence of how the system is built rather than a step we perform on data that could have identified you.

We analyze patterns, not entities

An example of something we may learn: relationships in which a partner's delivery lead stops attending within 60 days of a late-stage opportunity close at a lower rate than those in which they continue. That finding is useful to every customer and identifies none of them.

An example of something we will not produce: a statement that a named partner is disengaging across our customer base. Entity-level analysis is excluded permanently, not deferred or subjected to a threshold.

The reason is structural rather than editorial. Large partners work with many companies simultaneously, including companies that compete with one another. Any system that reasons about a named partner across its customer base discloses one customer's position to another, however carefully the output is phrased. We enforce this boundary in code rather than in policy, because a boundary that depends on judgment is one that will eventually be crossed.

Results require a minimum number of contributors

Cross-customer measurements are presented only when enough distinct organizations and underlying records stand behind them that no single organization can be identified from the result. Below that threshold, no result is shown.

Contribution can be switched off

Contribution to cross-customer analysis is enabled by default and can be disabled by a workspace administrator at any time, along with a request to remove aggregates already contributed. Features within your own workspace are unaffected either way.

Why this is published

These commitments are more useful to you as a published document than as an assurance given on request. A customer who can read where the boundary sits does not have to rely on our characterization of it.

Questions about anything on this page: privacy@gettriad.io. See also our Privacy Policy and Terms of Service.