# Nominal Analyst | AI Agent for Hardware Engineering [View this page](https://nominal.io/products/analyst) Nominal Analyst is an AI agent for engineers building hardware in Nominal. It helps teams ask questions about their data, build editable analyses, and turn test plans into pass/fail checks, with sources and calculations engineers can review. ## Example investigation: BP-17 battery pack This is an illustrative demonstration with synthetic data. It is not a customer test result or a performance benchmark. ### Start with the question that matters Ask about a test result. Analyst brings together the requirement, measurements, and evidence behind it. **Engineer:** This battery pack failed its latest test. Which requirement failed, and when? **Analyst:** The temperature-spread check failed. Spread exceeded the 8 °C limit for 102 consecutive seconds during sustained load. The requirement allows less than 10 seconds above that limit. ### Check results against the test plan Analyst reads the requirement in your test plan, stored in Nominal. It turns the limit into a pass/fail check, runs it, and links the result to its source. **Engineer:** Does this meet the temperature-spread requirement? **Analyst:** No. BP17-044 fails THERM-04. During the evaluated sustained-load interval, spread exceeded 8 °C for 102 consecutive seconds. The test plan prohibits 10 consecutive seconds or longer above that limit. ### Follow the result to the measurements Move from an overall test result to the signals and physical locations worth examining. **Engineer:** Which measurements account for the larger spread? **Analyst:** At 150 seconds after load start, T04 measured 56.4 °C and T01 measured 44.0 °C. These were the hottest and coolest measured locations, giving a spread of 12.4 °C. ### Bring context into the investigation Compare the recorded conditions and identify what additional evidence would help explain the result. **Engineer:** Could different cooling conditions explain this? **Analyst:** The recorded coolant-flow and inlet-temperature settings match the previous tests. That doesn’t establish whether cooling behaved the same during the run. This dataset has no coolant-flow trace to evaluate that explanation. ### Leave with a clearer next step Keep the findings, supporting analysis, and open questions together so your team can continue the investigation. **Engineer:** Summarize what we know and what we should investigate next. **Analyst:** The spread exceeded the requirement and was larger than in the three comparison tests. T04 and T01 define the peak difference; the cause remains unresolved. ### Keep the evidence and continue the investigation Return to the editable analysis, review the requirement result, and bring the open questions to your team. **Engineer:** Save this investigation for review. **Analyst:** Four tests, six temperature measurements, and the evidence behind the failed check. ### Test comparison | Test | Peak temperature spread | Check | | --- | --- | --- | | BP17-041 | 4.8 °C | PASS | | BP17-042 | 5.7 °C | PASS | | BP17-043 | 6.2 °C | PASS | | BP17-044 | 12.4 °C | FAIL | **Requirement THERM-04:** During the sustained-load phase, temperature spread across T01–T06 must not exceed 8 °C for 10 consecutive seconds or longer. The cause remains unresolved. The example identifies missing coolant-flow measurements and recommends further investigation. ## In the hands of engineers ### Wisk Aero > I've saved hours and hours by using Analyst. It's been a huge improvement to our iteration speed. Cooper Diamond, Staff Systems Engineer, Wisk Aero. ### Mach Industries > As a new employee it made it very easy to navigate the large dataset and understand which variables were most useful for analyzing certain situations and root causing issues. Spencer Sharpe, Flight Test Engineering Intern, Mach Industries. [Read the Analyst launch story](https://nominal.io/blog/analyst). ## Capabilities ### Build and edit analyses Create and update Nominal workbooks with plots, derived signals, tables, and layouts. ### Turn test plans into automations Read specs and test plans in Nominal, define pass/fail checks with thresholds, run them against test data, and review violations. ### Compare tests and calculate signals Overlay tests, apply filters and derivatives, and run frequency analysis such as FFT, PSD, and Bode plots. ### Write custom analysis code Calculate statistics across tests, explore trends across months of testing, and create visualizations for questions that need more context. ### Use your engineering documents Find relevant specs and test plans in Nominal Drives and cite the documents behind the analysis. ### Find the right data Locate tests, hardware, signals, and existing analyses in your Nominal workspace. ## Frequently asked questions ### What is Nominal Analyst? Nominal Analyst is an AI agent for engineers building hardware in Nominal. It helps teams ask questions about their data, build editable analyses, and turn test plans into pass/fail checks, with sources and calculations engineers can review. ### Is Analyst ready for my team? Analyst is available today in preview for engineers building on Nominal. Engineers are already using it today. ### What data can Analyst use? Your tests, hardware, signals, analyses, and checks in Nominal, plus the specs, test plans, and design documents your team keeps in Nominal. ### What can Analyst see and change? Analyst acts with your permissions. It can only see and edit what you allow it to. ### Can I customize how Analyst works? Yes. Write skills that describe your team's analysis methods, review steps, or reporting conventions, and share them across your workspace so everyone's Analyst works the same way. ### Can I check its work? Yes. You can inspect the calculations, code, and sources behind every analysis, edit the results in Nominal, or ask Analyst to refine them. ### Is my data used to train models? Your data is not currently used to train models, but we do keep traces and product analytics to improve the product. ### How do I get access? Request a demo to see Analyst with the Nominal team. Already on Nominal? Ask your account team about preview access. ### How is Analyst different from a general-purpose AI assistant? Analyst works directly with your data and analysis tools inside Nominal. It builds analyses your team can edit, turns test plans into pass/fail checks and runs them, and follows the playbook your team teaches it through skills. Every conclusion traces back to the data or document behind it. ### How is Analyst different from connecting my own AI agent to Nominal? Both work in Nominal with your permissions. Nominal's MCP server gives the agent you already use, like Claude, Codex, or Devin, access to Nominal tools, which might be the right fit when your work spans other systems. Analyst is built for engineering work inside Nominal: it creates workbooks, writes and runs checks, derives signals, and grounds its work in your team's documents and skills. Already building with Claude, Codex, or Devin? Connect your own agent to Nominal through our MCP, the Model Context Protocol connection that lets external agents work in Nominal. [Learn more about Nominal for Developers](https://nominal.io/products/developers) [Request a demo](https://nominal.io/request-demo)