What you can do with Ternary MCP

Finance jobs Ternary MCP supports today, including spend analysis, budgets, cost alerts, dashboards, reports, savings recommendations, and AI ROI.

Ternary MCP gives your AI assistant access to the same technology-spend data and permissions you already have in Ternary.

Think in terms of the financial job you want to finish — not in terms of MCP tools. Ask the question in business language. The assistant discovers the relevant Ternary data and chooses the right capabilities.

Analyze actual spend

Ask for a total over a defined period, then break it down using dimensions available in your Ternary data.

What did we spend last month?

Break down last month's spend by provider.

Show AWS spend by service for August.

Show the top five services by spend this quarter.

Ternary can discover the dimensions and values available in your tenant before running the query. You do not need to know the underlying schema names.

📘
Current data scope
Ad-hoc MCP cost queries currently use the FOCUS-normalized billing dataset, FOCUS_BILLING.
Other Ternary data sources — including Kubernetes, carbon-footprint data, and custom datasets — are not currently available through this query path.

Review budgets

Ask the assistant to review the budgets you can access and surface the ones that need attention.

For each budget, Ternary can provide the figures shown on the Budgets page:

Budget amount
Spend to date
Projected month-end spend
Current period
Health status

Example questions you can ask:

Which budgets are at risk or over budget?

Show me our budgets ordered by severity.

What is the current status of the Infrastructure budget?

🚧
Budget figures are a snapshot
These numbers come from the budget's last evaluation. They are not a historical series, and the projected month-end spend is not Ternary Forecasting.
Read Capabilities and limitations before using budget figures for historical or period-by-period analysis.

Review cost alerts

Ternary MCP can retrieve the cost alerts fired by your existing Ternary alert rules.

This includes:

Anomaly detections
Threshold breaches
The observed value
The baseline the value was evaluated against
Alert status
Information about the rule being monitored

Example questions you can ask:

Which cost alerts fired in the last 30 days?

Show recent anomaly alerts.

Which threshold alerts fired this month?

Ternary MCP reads alerts Ternary has already detected. It does not independently decide whether raw spend is anomalous.

Review dashboard changes

If a dashboard already represents a financial view you monitor, your assistant can compare that dashboard across two time windows.

The comparison can include:

Current value
Prior value
Absolute change
Percentage change
Dimension groups with the largest movements

Example questions you can ask:

What changed on our Executive dashboard?

Compare the last seven full days with the seven days before that on the Cloud Spend dashboard.

Which widgets moved the most?

📌
Strongest current comparison path
Dashboard comparison is the most reliable way Ternary MCP compares two periods today. Ternary computes the current and prior values on the server.
Some widget types cannot be compared honestly. Those come back with an explanation instead of a misleading figure.

Work with saved reports

Your assistant can find reports by name and retrieve their data.

Find our Monthly Executive Summary report.

Show me the data from the Cloud Spend by Team report.

List the reports available to me.

If an analysis is worth keeping, you can also ask the assistant to save it as a private report:

Save this analysis as a private report called Q3 Provider Spend.

Creating a private report is currently the only MCP action that writes data to Ternary.

Review savings opportunities

Ternary MCP can retrieve current savings recommendations and rank them by estimated monthly savings.

Recommendations can include idle, rightsizing, and other supported optimization types.

What are our largest current savings opportunities?

Show AWS recommendations with the highest estimated monthly savings.

List currently applicable savings recommendations.

The recommendations returned by MCP match the current Recommendations experience. They are not a complete commitment-management view.

Weigh AI spend against revenue

If your revenue is in Ternary, your assistant can compare AI spend with revenue for the whole tenant, or per product, team, or other Custom Label.

How does our AI spend compare with revenue over the last 30 days?

Show AI cost against revenue by product for last quarter.

Which products have the highest AI cost relative to revenue?

📘
Revenue has to be in Ternary first
AI cost comes from Ternary's own AI classification of billing data. Revenue comes from a bring-your-own-data import. A per-product ratio also needs a billing rule that assigns the Custom Label to your revenue rows.
If anything is missing, the assistant reports AI cost and tells you what is missing instead of estimating a ratio.

Explore with follow-up questions

You can keep going after the first answer.

For example:

Show last month's spend by provider.

Then:

Now show AWS by service.

Then:

Limit that to the top five services.

The assistant can run additional Ternary queries to narrow the analysis.

Be careful with causal questions

Questions such as:

Why did spend increase?

are broader than a simple cost query.

Outside dashboard comparison, the assistant may need to run several Ternary queries and do its own comparison or ranking. That can help you investigate a change. A group that moved with the total is not automatically the root cause.

For finance-critical analysis, prefer questions with an explicit:

Scope
Period
Metric
Comparison baseline

See Questions to ask Ternary for prompts that match current MCP capabilities.