AI dashboards: asking your data questions in plain language
An AI dashboard lets you get charts of your business numbers by asking in plain language. A dashboard is a screen that shows numbers such as sales or open orders as charts. Normally someone has to set up each chart by hand. With AI you type "show open orders by month", and the chart appears. The numbers still come from your own records. The AI does not make them up.
Who does what: the AI and your records
This is the most important thing to understand. AI is good with words and unreliable with sums. A well-built AI dashboard uses each side for what it does well.
Think of a waiter and a kitchen. The waiter understands your order and passes it on. The kitchen cooks. The waiter does not invent the meal.
- The AI understands your sentence. It works out which records you mean and which kind of chart fits.
- The software counts. The chart is calculated from your records in the usual, exact way.
- The AI describes. If you ask, it reads the finished chart and puts it into a few sentences.
If a product lets the AI produce the figures itself, be careful with every number.
An example in four requests
The owner of a small trading company prepares for Monday's team meeting. She has ten minutes.
- "Show me the open offers by stage, and the total value." The AI suggests a chart and one big number. She confirms, and both appear.
- "Only offers we expect to close in the next three months." The AI narrows both down. She did not have to know how.
- "Put the chart on the left and make it larger." The layout changes.
- "What stands out here?" The AI answers in a few sentences. Most of the value sits in offers that were sent but never answered.
She now has a picture and something to discuss. She still has to phone those customers to learn why they are silent. The AI showed her where to look.
Where this helps most
- People who have never built a chart. A setup screen with many options stops most people. A sentence does not.
- Questions you ask only once. A view for one meeting is not worth ten minutes of setup. It is worth one sentence.
- The "so what?" moment. A chart nobody explains is decoration. A short first reading gets the discussion going.
Where it does not help
- Unclear words. If "active customer" means three things in your company, the AI will pick one. Agree on the meaning first.
- Things nobody records. No sentence produces a chart of information that was never entered.
- The reason behind a number. The AI sees the values. It does not know that a big customer paused orders because of a move.
Three things to check in any AI dashboard
- Can you see what was built? You should be able to see and change what the chart counts.
- Does it ask before it changes something? This matters most for a dashboard the whole team uses.
- Does it follow your access rules? Nobody should get a chart of records they are not allowed to open.
How to ask
Name three things: what to count, how to split it and which records to include. "Tasks by status for my team, created this month" works. "How are we doing?" does not. If the result is not right, correct it with a second sentence.
How Worknano handles this
Worknano splits the two jobs between two AI agents. The Dashboard Assistant creates dashboards, adds and filters charts and rearranges the layout when you ask. Its charts point at your existing records. It reads only what you may read, and it shows every change as a preview that you confirm. The Chart Analyst then writes a short reading of a chart from a summary of the rows it shows. It cannot change anything, and a number that is not in that summary does not appear in its answer.
- Dashboard Assistant: build dashboards and widgets by asking
- Chart Analyst: a short, honest reading of any dashboard chart
- Dashboards in Worknano
Create a free workspace and build your first chart with one sentence.