Can the AI assistants see data from other companies?

No. Every AI assistant tool call is scoped to the logged-in user's own company, the same as every other part of the platform. The AI cannot see or answer questions using another company's data, and it only ever reads data — it never creates, edits or deletes anything.

What kind of questions can the AI assistants answer?

Questions grounded in a company's real, current data — such as how many guards are currently deployed, which sites have attendance gaps this week, what the current payroll status looks like, or a summary of pending leave requests. The Advisor persona in particular is built for general operational questions across the whole company.

Which roles can use the AI assistants?

AI Assistants are available to Company Admin and HR roles, since answering meaningful operational questions typically requires access to company-wide employee, attendance and payroll data.

AI Assistants for Security Agency Operations

A security agency owner or HR manager running dozens of sites and hundreds of guards ends up asking the same kinds of questions constantly — "how many guards are unassigned right now," "which sites had attendance issues this week," "what does this month's payroll look like so far." Answering these normally means opening several different report screens and mentally combining them. Raksha Kavach's AI Assistants exist to shortcut that — a chat interface, grounded strictly in a company's own real data.

Ten personas, one underlying principle

The platform includes ten distinct AI assistant personas — including an Advisor for general operational questions, a Portal Manager, a Security Agency Manager, a Field Officer, a Data Researcher, a Data Collector, regional heads for Gujarat and Rajasthan operations, and personas oriented toward social media and digital promotion. Each has a slightly different framing and system prompt suited to the kind of question it's meant to help with, but every one of them works the same underlying way: it can read a company's own operational data through a defined set of tools, and it can never write, edit, or delete anything.

Read-only, company-scoped, by design

This is the part worth being explicit about: the AI assistants have a fixed set of read-only tools — covering company overview, employees, clients, sites, deployments, attendance, leave, and payroll — and every single one of those tools is automatically scoped to the logged-in user's own company. There is no tool that lets the AI see across companies, and there is no tool that lets it change data. It answers questions using real numbers pulled live from your database at the moment you ask, not a static or hypothetical summary, and it cannot accidentally leak another company's data because it structurally has no access to it.

A conversation, not a one-shot query

Each assistant maintains its own conversation history per user, so a follow-up question ("and how does that compare to last month?") can build on the previous answer rather than requiring a fully re-specified question every time. Conversations are private to the user who started them.

Practical uses for a busy operations team

In practice, this tends to get used for the kind of quick operational check that would otherwise mean opening three different report pages — a fast summary of today's coverage gaps, a plain-language explanation of why a particular payroll period's numbers look the way they do, or a quick pull of which guards have pending leave requests. It's a faster way to ask a question you'd otherwise dig for manually, not a replacement for the underlying reports and dashboards themselves.

An AI assistant is only useful to a business if it can't quietly make things up — every answer here is grounded in a live, read-only query against your own company's real data.

Conclusion

AI Assistants in Raksha Kavach aren't a generic chatbot bolted onto the product — they're a genuinely scoped, read-only interface onto a company's own operational data, built so an admin or HR manager can ask a plain question and get a fast, accurate, company-specific answer instead of navigating multiple report screens by hand.