Analyze data with an AI assistant
Profile a table, calculate totals and make the source of each result clear.
Define the question and inspect the data
Start with a specific question, such as “What is the total amount in Orders?” Identify the table, amount field, filters and reporting period. Install the Cynexio Skill, which includes the CLI, and configure its connection; keep credentials in the supported credential store or a trusted secret environment, outside the conversation.
cynexio field list --table 2234567890123456789 --profile work
cynexio data profile --table 2234567890123456789 \
--fields 3234567890123456789 --max-records 1000 --profile workReview the source row count, collection time, completeness and quality indicators. A profile capped at 1,000 records describes the sampled snapshot; a truncated result is not evidence about every record.
Calculate a total
Save this as aggregate.json and replace the field ID with your numeric Amount field. With no filter or grouping, it requests a single sum over the accessible table data.
{
"metrics": [{
"op": "sum",
"fieldId": "3234567890123456789",
"alias": "sales"
}],
"groupBy": [],
"filter": null
}cynexio data aggregate --table 2234567890123456789 \
--data @aggregate.json --profile workCheck the returned sales metric against the intended field and filters. The aggregate operation refuses a dataset beyond its full-read limit instead of presenting a partial total as complete.
Make the answer verifiable
- State the table, selected fields, filters and time period used.
- Include the collection time, source row count and completeness from the result.
- Keep exact decimals intact. Explain missing values or filters that affect the interpretation.
- Separate calculated facts from the assistant’s interpretation. If data is incomplete, say so before presenting conclusions.