Security & Trust
It speaks to your customers. It is built accordingly.
This sits between your business and the people who pay you, holding what they tell you. The questions your security team will ask are the right questions, and these are the answers.
Roles, permissions and the trail
| See conversations | Join and reply | Change the agent | Export records | |
|---|---|---|---|---|
| Owner | ||||
| Administrator | ||||
| Member | ||||
| Viewer |
Audit trail
- 09:41 · Maya joined a conversation with Priya Nair
- 09:12 · Alex published a change to what the agent knows
- 08:55 · Records exported by the owner
It stays yours, and you can prove where it is.
No copy of your customer record in somebody else's product that you cannot get back.
- One database
- Conversations and records in a database you control
- Exportable
- Your data out, in full, whenever you want it
- Your environment, if needed
- The whole platform inside your own infrastructure
- Deletable
- A clear route to erase a customer on request
Only the right people, and it is recorded.
Who can see a conversation, change what the agent knows, or export a record is decided by role, and every consequential action leaves a trail.
Roles and permissions
Members, administrators and owners with distinct powers. People see the conversations and records their role allows, not everything by default.
Audit trail
Who changed the agent, who exported what, who joined which conversation, and when. Written as it happens, not reconstructed afterwards.
Separated by workspace
Each business, brand or unit is bounded. One team cannot read another team's conversations because they share a platform.
Encrypted in transit
Everything between the customer, your site and the platform travels over TLS. At rest, it is protected by the database and infrastructure you run it on.
An agent that invents an answer has not saved you work. It has created a liability.
The failure mode that matters is not silence. It is a confident wrong answer given in your name, to a customer who believes it.
What stops it saying something you would not.
The risk with this technology is not that it fails to answer. It is that it answers confidently and wrongly, in your name.
It answers from your material
Your pages, documents and policies, with the source recorded against the answer, so anything it says can be checked against what you actually published.
It says when it does not know
Rather than filling the gap. Unanswered questions are collected for your team instead of improvised in front of a customer.
You set what it must never attempt
Subjects that always go to a person, and things it is never permitted to promise on your behalf.
Everything is reviewable
Every conversation is kept in full, agent and human together, so you can see exactly what a customer was told.
Your own model providers
Run it against your own provider accounts under your own commercial and data terms, rather than ours.
We will tell you plainly where we are.
We do not claim certifications we do not hold, and we will not put a badge on a website to get through a procurement gate. Tell us what your organisation requires and we will give you a straight answer about what we have today, what is in progress, and what we would need to do for you.
Experience it
Send us your security questionnaire.
We would rather answer the hard questions early than discover them at the end of a procurement process.
Monday
10:04
Every answer traceable to the sentence it came from.
- Traceable answers
- Back to the sentence they came from
- Your retention policy
- Not a vendor default
- Removal on request
- And provably