A natural-language query is a question asked in ordinary words rather than by building a chart or writing a formula. “What was revenue by region last quarter?” “Which customers have not ordered in 90 days?” “Show me margin by product for the Dubai store.” The tool works out which tables, filters and calculations the question needs and returns a chart, a table or a sentence.
It matters because most questions are one-offs from people who will never build a dashboard. A managing director on the way to a meeting wants one number. A branch manager wants to know which lines slowed down. Previously each of those went to the person who “does the reports”, who added it to a list. With plain-language questions the list shrinks, and the analyst’s time goes on the questions that deserve it.
The caution is trust. An answer is only as good as the definitions and permissions behind it. If “revenue” means something different to the question tool than to the finance dashboard, the two will disagree and the tool will be blamed. Rule of thumb: plain-language questions should use the same defined metrics as the dashboards, respect the same row and column rules, and show the chart and the filters it applied so the answer can be checked rather than believed.
In Klayara, the AI chat answers questions from the same governed data as the dashboards — same metric definitions, same permissions — and shows its working as a chart you can save to a dashboard.