FinOps Analyzer

Turns live Lansweeper asset data into grounded answers on hardware refresh, license exposure and warranty risk, so IT and finance can make cost decisions backed by real numbers instead of estimates.

The FinOps Agent takes a question about IT asset spend — license waste, hardware refresh, end-of-life risk, contract and warranty exposure, total cost of ownership, or chargeback — and answers it end-to-end using live Lansweeper data. It pulls the relevant asset, software and lifecycle data, tells the user plainly what Lansweeper does and doesn’t cover, and asks for any missing financial or contract data rather than guessing. It then runs the analysis and produces a report or set of remediation tasks, always noting which numbers came from Lansweeper and which were user-supplied.

Where general AI can only answer “what does end-of-life hardware cost us?” in the abstract, this agent answers “which of your assets are end of life, what do they run, and what’s the remediation budget” — because it is working from the user’s actual estate, not a general estimate.

Requirements

Lansweeper MCP

https://mcp.lansweeper.com/mcp

LLM

Atlassian Rovo, Claude 3.5+, GPT-4o, Gemini 1.5 Pro or equivalent

Designed to work with any ITSM tool connected via MCP, specifically Rovo, but also Jira Service Management, ServiceNow, Freshservice.

Lansweeper plan

Starter, Pro, Enterprise

Additional info

Lansweeper does not hold purchase prices, licence entitlements, contract terms or replacement pricing. The agent asks the user to supply these — a spreadsheet, Confluence page or PDF — and states clearly what analysis it can’t do without them, rather than fabricating figures.

Skills It Uses

  • itsm-finops

    Claude skill — core FinOps analysis logic across licence waste, hardware refresh, EOL remediation, contract exposure, TCO and chargeback.

    Lansweeper MCP

    standard instruction (site confirmation, query catalogue, active-asset filtering, dataset reference)

Step by step setup guide

This agent is available on four platforms: Atlassian Rovo, Claude, Microsoft Copilot Studio and Google Gemini. Steps 1 to 3 are the same shape everywhere; the platform-specific detail is called out where the steps diverge. Follow the column for your platform.

Step 1 — Download

Download the agent package.

Step 2 — Open your agent platform

  • Rovo: open Atlassian Rovo from your Atlassian site and go to Agents.
  • Claude: open Claude Desktop or Claude Code.
  • Copilot Studio: open Microsoft Copilot Studio and create a new agent.
  • Gemini (Antigravity): open Google Antigravity (Google’s agentic development platform for Gemini).

Step 3 — Load the instructions

  • Rovo: create a new agent, paste the contents of the Rovo instruction set into the agent’s instructions field, and connect the Lansweeper MCP.
  • Claude: load the itsm-finops skill file into your skills folder (Claude Desktop) or plugin directory (Claude Code). No manual paste needed.
  • Copilot Studio: paste the contents of the Copilot Studio instruction set into the agent’s instructions field, and connect the Lansweeper MCP through Copilot Studio’s connector settings.
  • Gemini (Antigravity): load the Gemini SKILL.md file into your Antigravity workspace and connect the Lansweeper MCP through Antigravity’s connector settings. No manual paste needed.

Step 4 — Connect Lansweeper

Authenticate the Lansweeper MCP connection (EU or US endpoint, as required) and confirm the agent can retrieve your site list. All four platforms need this step; only the connection method in each platform’s settings differs.

Step 5 — Connect Confluence and Jira (optional)

If you want the agent to publish FinOps dashboards or raise remediation tasks directly, connect the Confluence and Jira MCPs (Rovo) or confirm your platform’s equivalent connection. This step isn’t required if you only want the agent to answer questions in-session.

Step 6 — First-run configuration (Claude skill only)

The Claude skill asks which Lansweeper site to use if more than one is available, and confirms whether you want output as a report, a Confluence page, or Jira issues. The Rovo, Copilot Studio and Gemini versions ask the same questions conversationally on first use.

Step 7 — Run a test question

Ask the agent a FinOps question, for example “Which of our assets are running an end-of-life operating system, and what would it cost to remediate them?” Confirm it leads with the number from Lansweeper, asks for any missing pricing or contract data rather than guessing, and states clearly what came from Lansweeper versus what you supplied