What we touch, and how we handle it.
We operate inside your HR systems and see your employees’ data. Here is the plain-language version of how we treat it: access, tooling, retention, and where we draw the line on AI.
Part 1 · Posture
Four commitments the rest of this page just makes specific.
1.1
Least access, by default
We take the narrowest access that does the job: scoped to named people, and removed when an engagement or a team member ends.
1.2
Your data stays in your systems
We operate inside the HRIS you own. Your employee records are not copied into a People Street database to run our service.
1.3
Never sold, never trained on
Your data is never sold, never shared beyond the subprocessors that run your own systems, and never used to train a public AI model.
1.4
A hard line on AI
We use AI throughout our own work. Where your data is involved it goes into enterprise tools configured not to train on it, never a consumer one. The specifics are below.
Part 2 · Access
The narrowest key that opens the right door.
2.1
Named, not shared
Access is tied to specific People Street people. No shared logins, no generic service accounts standing in for a person.
2.2
Least privilege
Each person gets the role the task needs and no more. Admin rights only where the work genuinely requires them.
2.3
Multi-factor everywhere
Every system that touches your data sits behind multi-factor authentication and a managed password vault.
2.4
Removed on exit
When an engagement ends or someone rolls off, revoking their access is on the offboarding checklist we run for our own team.
Part 3 · Systems
Your systems, plus the shortest list on top.
3.1
HR & payroll platforms
Rippling and HiBob: the system of record you’ve licensed and asked us to operate. We take other platforms on case by case.
3.2
Recruiting systems
Ashby, Gem and Greenhouse, where we run embedded recruiting into your HRIS.
3.3
Our own tooling
A password manager, a ticketing desk, and productivity tools, kept to the shortest list that does the job.
A current subprocessor list and a data processing agreement come with onboarding. Ask for them any time
Part 4 · AI
Everywhere in our work. Never into a model that trains on your people.
4.1
What we will not do
- Knowingly submit your confidential or personal data to any AI tool that trains third-party models on it
- Use consumer AI tools where your people’s data is involved
- Move your records out of the systems you own in order to process them
4.2
How we use it
- Throughout the work, without limitation: research, analysis, drafting, document review, process automation
- On enterprise or comparably configured tools wherever your data is processed
- With responsibility for the work ours, regardless of the tools used
- To configure the AI inside your platform against this same line
What teams ask before they hand this over.
Do you store our employee data on People Street systems?
No. We operate inside the platforms you license. Your records stay in your system of record; we don’t replicate them into a database of our own to run the service. Where a task needs an export, it’s minimal, time-bound, and deleted when the task is done.
Do you use our data to train AI?
Never. Your data is not used to train any model, ours or a third party’s. We use AI throughout our own work, and where your data is involved we use enterprise tools configured so that what goes in is not used for training.
Who can see our data, and how is that controlled?
Named people on your engagement, each with the least access the work requires, behind multi-factor authentication. No shared logins. When someone rolls off or the engagement ends, revoking their access is part of the offboarding checklist we run for our own team.
Can we get a subprocessor list and a DPA?
Yes. We’ll share the current list of the systems that touch your data and sign a data processing agreement as part of onboarding. Both are named in your agreement rather than described on a call.
What happens to our data if we leave?
You keep everything: it was always in your systems. We hand over documentation and runbooks written as the work happened, transfer to a named internal owner, and remove our access on a schedule you can see.
Related
Keep reading
A security question we didn’t cover?
Ask it directly. We’ll answer in plain language, share the documents, and tell you where the line is, before you’ve signed anything.