> ## Documentation Index
> Fetch the complete documentation index at: https://docs.prisme.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Overview
> Enterprise use cases on Prisme.ai in three shapes: agents, agentic workflows, and full-stack portals. Each with a ready-to-paste prompt for your coding agent over MCP.
Use cases, not features. Whatever the business problem, it takes one of **three shapes**: an agent, an agentic workflow, or a full-stack portal. Start from the shape that fits, then combine them. Most enterprises start on **SaaS** for speed and move to **self-hosted** when criticality or sovereignty require it.
Every example below comes with a **prompt**. Connect the Prisme.ai [MCP](/products/ai-securechat/mcp-connections) to VS Code, paste the prompt into **Claude Code, GitHub Copilot, or Codex** (and use the [starter SPA](https://github.com/prismeai/starter-spa) for portals), then review and version in Git.
## Agents
Talk to it, it acts. A conversational agent that reasons over your knowledge, calls tools, and acts on your systems.
A secure, multi-LLM assistant for daily work, grounded on internal content, no data leakage.
```text theme={null}
Using the Prisme.ai MCP, create a general assistant agent for all employees:
secure multi-LLM chat with our internal knowledge base attached and no external
data leakage. Publish it to the Store.
```
Instant, sourced answers on leave, payroll, and mobility.
```text theme={null}
Using the Prisme.ai MCP, create an HR agent that answers questions on leave,
payroll, and mobility, grounded on our HR policies knowledge base with sources
cited. Restrict access to employees.
```
Conversational research over internal case law, with clause comparison.
```text theme={null}
Using the Prisme.ai MCP, create a legal research agent over our internal
case-law knowledge base: answer with citations and compare clauses across
contracts. Keep it confidential to the legal team.
```
Real-time, cited answer suggestions for support agents.
```text theme={null}
Using the Prisme.ai MCP, create a support-assist agent that suggests cited
answers in real time, reading from our help center and the customer record via
the CRM connector.
```
The agent gathers data across CRM, news, and your systems, and prepares the meeting and the pitch.
```text theme={null}
Using the Prisme.ai MCP, create an agent in the "sales" workspace that prepares
a prospecting brief: read the account from our CRM connector, search recent
news, and draft a one-page brief and a suggested pitch. Ground it on our
sales-playbook knowledge base. Require human approval before it writes back.
```
Query your warehouse in plain language and get exact figures.
```text theme={null}
Using the Prisme.ai MCP, create a conversational data-analyst agent that queries
our data warehouse via the connector, returns exact figures, and explains the
query. Log every access for audit.
```
## Agentic workflows
Headless automation. A process that runs on its own (endpoint, event, or schedule), calls agents where judgment is needed, and acts on your systems. No UI required.
Monitor the ledger, detect anomalies, propose corrective entries for approval.
```text theme={null}
Using the Prisme.ai MCP, add an automation that runs the continuous financial
close: monitor the ledger, detect anomalies, and propose corrective entries for
approval before posting.
```
After a meeting, draft the CRM notes, update SAP, refresh Power BI.
```text theme={null}
Using the Prisme.ai MCP, add an automation to the "sales" workspace triggered by
a webhook after a meeting: take the meeting notes, draft the CRM update, push the
opportunity fields to SAP, and refresh the Power BI dataset. Log every step and
ask for approval before the SAP write.
```
Match invoices to purchase orders, flag exceptions, route approvals.
```text theme={null}
Using the Prisme.ai MCP, add a procure-to-pay automation: match incoming invoices
to purchase orders, flag exceptions, and route approvals, updating the ERP.
```
Resolve tickets end to end with policy-based fixes.
```text theme={null}
Using the Prisme.ai MCP, add an automation that resolves IT tickets end to end:
read the ticket from the service desk, diagnose, apply policy-based fixes, update
the ticket, and escalate when needed.
```
Assemble an AI Act or DORA dossier from evidence.
```text theme={null}
Using the Prisme.ai MCP, add an automation that assembles an AI Act or DORA
compliance dossier: gather evidence from Governe and Insights, produce the
report, and file it.
```
Triage SIEM/EDR alerts and propose containment for approval.
```text theme={null}
Using the Prisme.ai MCP, add an automation that triages SIEM/EDR alerts, filters
false positives, and proposes containment for human approval, logging everything.
```
Extract fields from documents, validate, and route.
```text theme={null}
Using the Prisme.ai MCP, add an automation triggered on new documents that
extracts fields from contracts and invoices, validates them, and routes them to
the right system.
```
## Agentic portals (full-stack)
The highest-value shape: a React front-end, agents, and workflows in one custom portal that replaces a patchwork of siloed tools, connected to your systems and governed.
A React front-end over your CRM, with agents that draft, summarize, and update records.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a CRM portal: a React
front-end over our CRM, with agents that draft, summarize, and update records.
Version it in Git.
```
Ticket queue with agent triage and resolution, connected to your service desk.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold an ITSM portal: a ticket
queue with agent triage and resolution, connected to our service desk.
```
Usage map, regulatory watch agent, and audit-ready reporting in one place.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a compliance portal: an AI
usage map from Governe, a regulatory-watch agent, and audit-ready reporting.
```
Onboarding, scoring, and contract-risk agents over your procurement systems.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a supplier portal:
onboarding, scoring, and contract-risk agents over our procurement systems.
```
One front-end over apps that do not talk to each other, with workflows replacing manual copy-paste.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a single front-end over our
[legacy apps], with workflows that replace the manual copy-paste between them.
```
Aggregate alerts, enrich with agents, drive response on approval.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a security portal: alert
aggregation, enrichment agents, and response actions on approval.
```
Connect entities across your data and let agents navigate and answer over the graph.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a knowledge-graph portal:
connect entities across our data and let agents navigate and answer over the graph.
```
The prospecting brief and the post-meeting updates, in one UI.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a sales cockpit: an account
list and a brief view, wired to the prospecting agent and the post-meeting
workflow, plus connectors to CRM, SAP, and Power BI.
```
Stand a portal up fast for an event or a crisis cell, retire it when done.
```text theme={null}
Using the Prisme.ai MCP and the starter SPA, scaffold a short-lived crisis portal:
a dashboard, an intake form, and a notification workflow. Make it easy to retire.
```
The full theme: one pattern, many portals, plus the economics that make it pay off.
## One need, three shapes
The same business need can take any of the three shapes. A **Sales assistant** can be an **agent** (prepare the prospecting brief), an **agentic workflow** (update CRM, SAP, and Power BI after each meeting), or a **full portal** (a sales cockpit that does both, with a UI). You pick the shape per use case, on the same governed foundation.
## By sector
Regulated industries most often deploy self-hosted, for data residency and sovereignty.
| Sector | Flagship use case | Typical deployment |
| ----------------------- | ------------------------------------------------------------------------------ | ------------------------------------- |
| **Defense** | Multi-source intelligence fusion (OSINT/SIGINT/IMINT), briefs for validation | Self-hosted / air-gapped |
| **Aerospace** | Autonomous AOG resolution: parts sourcing, airworthiness check, logistics plan | Self-hosted / private |
| **Banking & Insurance** | Continuous financial close, reconciliation, audit-ready reporting | Self-hosted / sovereign |
| **Energy** | Autonomous grid outage triage and balancing (AMI/OMS/SCADA) | Sovereign / self-hosted |
| **Public Sector** | End-to-end citizen request management: classify, route, track | Sovereign private cloud / self-hosted |
| **Manufacturing** | Agentic supply chain: inventory, suppliers, logistics | Self-hosted / edge |
| **Telecom** | Network supervision, customer resolution, field orchestration | Your control |
## Next steps
Frame and prioritize your first use cases
The tool every one of these use cases is built with