Converses in natural language with your operation's data. Learns your business vocabulary. Never exposes what the user has no permission to see.
The operator asks in natural language, within the screen they are working on.
Nous queries the operation's live data — never a copy exported to another system.
The context of who asked is applied before the response is assembled.
The response comes back with only the fields the user can see — not one more.
If the response involves a screen, Nous takes the user there with a click.
§ 01
What Nous is
Nous lives inside the platform. Not in another tab, not in another SaaS. The conversation happens where work is already happening, and the same governance layer that protects the operation protects the AI.
The most important difference is not the quality of the language model — that changes over time. It is the level of governed access to your operation's data . When the operator asks "which orders are due today?", Nous queries the platform's live data with the user's permissions applied before the response is assembled.
Embedded
Shares the context of the screen the user is on, without requiring an application switch.
Governed
The same access rules that apply to the operator apply to the AI. Always.
§ 02
Each question becomes a query, report, chart, explanation, or navigation — without becoming a data analyst to get an answer.
“Which orders are due today and what do they amount to?”
▸ Query against live data, with the user's scope applied.
“Which customer has accumulated the most overdue balance in the last 90 days?”
▸ Aggregation within your operation's context.
“Generate a sales report by salesperson for last month.”
▸ Report assembled on demand — no parallel spreadsheet.
“Show delinquency by region in a chart.”
▸ Visualization generated right there, without leaving the conversation.
“Why was this entry blocked by governance?”
▸ The rule is cited, not hidden.
“Take me to the order approval screen.”
▸ One click to get where the action happens.
The common operator asks in business language. Nous translates for the platform.
§ 03
With each conversation, Nous learns from the user. The following day, friction is lower. The following week, nearly zero.
Memory boundary
Memory is per user within the institution. A team member does not see your memory. Another company does not either.
If you say "João" and it is always the salesperson with the same record, Nous learns. The following week, ask "how much did João sell?" — the answer comes without asking for clarification.
▸ Less rephrasing. More direct questions.
If you typically request reports with a PDF preview, Nous begins proposing that format by default. The same applies to data slices, header metrics and granularity.
▸ The interface molds itself to how you use it.
"Picking up from yesterday" works. Nous resumes the last conversation without the user having to recap what they were analyzing.
▸ You return to the work, not to square one.
§ 04
Beyond responding, Nous observes the operation and alerts on anomalies. Each alert carries a confidence score: less relevant warnings stay below the radar; relevant ones arrive before the problem becomes urgent.
Three tax invoices issued today carry a risk of rejection by the tax authority.
Item X's stock has fallen below the defined replenishment point.
Customer Y has not purchased in 90 days — behavior outside the observed pattern.
⎯ Confidence is shown alongside the alert. The operator decides what to do with it.
§ 05
This is the most important point for the buyer's decision-maker. Nous does not "know everything" — it knows what the user asking has the right to know.
If the user has no permission to see the "cost" field, Nous does not expose it — even if the data exists in the operation.
If the user only has access to data from one branch, Nous does not query data from another. Scope limits are applied before the query.
If a governance policy blocks an action, Nous explains the reason — it does not try to circumvent it, does not negotiate, does not suggest a workaround.
Brazilian documents — CPF and CNPJ — go through automatic masking on output, even when the user has permission to view them. Additional PII types are covered as business rules are configured.
Financial controller
Scope: all units · Sensitive fields visible
Profile with broad access. CPF and CNPJ still go through masking by minimization principle.
Store operator
Scope: current branch · Sensitive fields hidden
Same question, response appropriate to the profile. No exception for the AI.
Closed by default
Without the user's permission context, Nous refuses the operation. This is not an optional setting — it is the base design. Fail-closed, not fail-open.
§ 06
No extra window, no tab switching, no browser plugin. The AI is where the operation is.
Fluid conversation in natural language, on the same screen the user is working on — no separate window, no context switch.
PT-BR is Nous's base language. Multiple additional languages are supported, with coverage varying by region served.
If the question calls for a chart, table, or map, Nous assembles it right there. The response is not just text — it is the artifact you need.
If the response involves going to a screen, Nous takes the user there with a click. Conversation and navigation are the same movement.
§ 07
Honesty about what Nous does — and, more importantly, about what it does not yet do. A promise larger than delivery erodes trust.
It queries, generates, explains, navigates — and carries out actions in the operation only after your confirmation. Sensitive actions require dual approval, and everything is recorded.
For decisions requiring human judgment (approving a contract, deciding on credit policy), Nous helps see the picture — it does not make the decision.
If the operation's catalog is disorganized, Nous has less context to work with. The platform and Nous grow together with use.
§ 08
A new operator does not need to memorize menus. Ask in natural language — Nous explains in the context of your current screen. The learning curve is drastically lower when the system has an AI that knows the system itself.
First day of use
“How do I generate a sales report by salesperson?”
Nous responds with step-by-step guidance in the context of your current screen, explains where to find the relevant filters, and, if you prefer, assembles the report directly through the conversation.
Want to see Nous operating with live data — respecting permissions by design?