
In this use case, the customer owns an expensive MRI machine that requires regular servicing or repairs. LitenAI assists technicians through a conversational interface, providing expert information sourced from ingested manuals and log data. By consolidating all data into LitenAI Data Base, LitenAI enables intelligent searches and in-depth analysis. The LitenAI Agent leverages this data to provide precise technical support for these devices.
Preventive maintenance tasks are routinely carried out. Below is a chat session illustrating a preventive maintenance interaction between a technician and the system.
For this demo, ensure you select techassist as your AI database. Select reason as your AI Agent. Scroll down for instructions on how to select these.
Preventive Maintenance
Customers can simply request tasks to be performed on a machine. In this example, we have ingested an MRI machine manual to demonstrate the process. You can ask a question like this.
I need help with quarterly maintenance of Philips MRI machine. Do not provide details. Include any test images.
The answer provided is specific and derived directly from the ingested data, ensuring relevance over generic responses.

Customers can now query the service record table to review previous maintenance work. They can access the latest records and learn from them as needed.
Query the servicerecord to show me the PMs done on this MRI machine.
This query performs a search on the service record table.

Customers can request specific steps for each test and ask clarifying questions as needed.
Provide me exact names of test that are part of Calibration Checks for MRI machines from manuals. Don’t include any details on the steps.

In the example below, steps for medical field tests are provided, derived directly from customer-ingested documents.
Show me the exact setup steps for medical field homogeneity test from manuals. Do not include details.

Once completed, customers can generate reports and add them to the records if needed. Agents can also be customized to ensure the report format aligns with customer requirements.
Generate report showing the test result and setup using manuals.

Break Fix
When machines break down, they need to be repaired as quickly as possible to minimize operating costs. LitenAI assists technicians by providing expert answers, reducing delays and idle time caused by waiting for human experts. Customers can initiate troubleshooting by describing the problem they observe in the system.
Image produced is noisy and does not look correct. Can you suggest me a few fixes from Philips Insignia MRI manual.

Customers can now inquire about additional scenarios and explore various possibilities.
Tried these steps, but did not fix the issue. Could a faulty coil be an issue?

Machine logs are available in the LitenAI Smart Lake, allowing customers to search for failures or anomalies. LitenAI leverages big data technology to securely store all data and elastically scale to accommodate large log volumes.
Query using sql for fault or gradient in message field of mrilogs table.

Customer can now ask for exact steps to fix this possible issue.
From the manual, tell me exact steps to fix coil failures. Include any test images.

Customers can now generate detailed reports.
Can you create a text report with headers and tables summarizing the work done on break fix.

The report includes a conclusion, and the reporting agent can be customized as needed. Reports can also be recorded in a table if required.

Reasoning Agent
In reasoning agent, a technician poses a high-level question, and LitenAI generates a plan, executing steps to provide a final answer.
Can you reason out and provide a single resolution. My Phillips MRI machine has been giving incorrect images. Look for faulty or gradient in mrilogs table. See if that is an issues. If not, look for other causes of failures. I want you to analyze and come back with the most likely cause of failure.
LitenAI first generates a plan.

It then generates the query code and runs it on LitenAI accelerated big data platform.

It then analyzes it and generates the likely cause of failure.

The reasoning agent enhances its accuracy and effectiveness as LitenAI’s Smart Lake ingests more data and users engage with the agents.
Business Analytics
LitenAI securely stores machine data and customer documents within customer-controlled storage. The LitenAI Agent enables the extraction and visualization of various metrics and trends for business purposes.
This example examines the service record to analyze the number of tasks completed daily over the course of a year.
For servicerecord table, use sql to count the number of rows for every day before year 2020. Execute the generated SQL.

Data can also be plotted and shared for better visualization and communication.
Plot the data in line.

Using LitenAI agents, analytics can be completed on this data. Reports and plots can be generated to help with understanding the machine usage as well as its failures among other tasks. These all get done using a conversational interface.
LitenAI Database
In the LitenAI database, the customer has ingested their knowledge documents and connected to the necessary databases. All data remains securely stored in the customer’s own storage. Data can be ingested in multiple ways—programmatically from stored files, via streaming for continuous ingestion, or directly through the GUI.
For this test:
- Select the Reason Agent from the AI Agent card.
- Select the techassist data lake from the AI Database card in the sidebar.
See a screen shot below with these selections.

Contact us to explore how our tools can support your equipment and use case.