A question that asks for a list comes back as a table. A question about a trend comes back as a chart.
Five steps; the wording of the question decides the first one.
Wording carries shape. “Show me”, “list” and “by business unit” ask for rows, so ChatLima builds a table. “Through the year”, “month by month” and “trend” ask for a series, so it draws a chart. “How much” and “one line for the board pack” ask for a single figure, so two sentences carry it.
The size of the result counts as well. Four rows read well as a table; forty rows read better as a chart with the leading rows named underneath. ChatLima uses both signals and states the shape it chose above the answer.
Units come from the mapping your team confirmed. A quantity keeps the unit of measure its column holds, an amount keeps the currency of the company code, and a date follows the format your estate writes. So 48.2 arrives as $48.2M, because the mapped column is a value in dollars.
Every shape carries the same six things: the terms used, the systems read, the time it took, the access that applied, the query and the shape itself. Switching from a table to a chart keeps all six and re-draws the same rows.
The same eight rows served a table, a chart and two sentences in under eight seconds.
Inventory value stands at $48.2M at the end of August, which is $3.2M above the budget line of $45.0M. Four business units carry the gap.
| Business unit | Inventory value | Against budget | Days of inventory |
|---|---|---|---|
| Controllers | $18.6M | +$2.1M | 64 days |
| Power systems | $14.9M | +$0.6M | 55 days |
| Sensors | $8.4M | +$0.3M | 51 days |
| Connectors | $6.3M | +$0.2M | 49 days |
Here it is for January to August. Inventory value rose above the budget line in May and stayed there. The chart is in the next section, drawn from the same eight rows.
Inventory value closed August at $48.2M, $3.2M above budget and 13 days of inventory above the 45-day target. Controllers hold $2.1M of the gap, most of it in material 7741-X.
| How the question is worded | What comes back | The shape ChatLima chooses |
|---|---|---|
| “by business unit”, “list the lines”, “show me” | a handful of rows | A table, with the columns the question named |
| “through the year”, “month by month”, “trend” | a series over time | A line chart, with the target or budget line drawn on it |
| “how much”, “where does it sit”, “one line for the pack” | one figure | Two sentences, with the figure in bold |
| “compare”, “against last year”, “by plant and month” | two series or two groups | A bar chart, or a table with both columns side by side |
Three questions, one result, three shapes. The wording in the left column is what people typed in the showcase estate.
One glance gives the trend; one click gives the figures behind it.
Charts follow a few plain rules. The axis starts where the values live, so a rise reads at its true size. The target or the budget is drawn beside the measure when the mapping holds one. A mark on the axis names an event the estate recorded, such as the customer push-out in May.
Colour carries meaning: one tone for the measure, a grey line for the target, an amber tone for a value past a limit. A reader takes the point from the shape of the line and confirms it in the rows underneath.
The rows behind the chart stay one click away, so a reader can read the figures that made each point.
The query is the query the table ran. Switching shape re-draws the result and leaves the query as it was.
A chart carries the same access line as any other answer. A person cleared for two plants sees a chart of those two plants.
Five steps turn a month-end review into something one person runs in a minute.
A walkthrough is a saved line of questions. It holds the wording, the shape each answer took and any clarifying choice made along the way, so a replay a month later is the same review.
Replays read today’s data, so the figures move while the questions stay fixed. Juno keeps each run, which lets a reader put August beside September and see what moved.
A walkthrough travels between people. The questions are shared; the rows stay each reader’s own, because every replay runs with the access of the person who started it.
Five routes out of one answer, each one carrying the asking person’s access.
| Route | What travels | Who uses it | What holds |
|---|---|---|---|
| The ChatLima screen | The question, the answer, the shape and the history of the thread | Anyone with a Juno login | The asking person’s row and column access |
| Power BI connector | The result set behind the answer, refreshed on the report’s own schedule | Report authors and analysts | Row and column rules travel with the connection |
| A downloaded file | The rows of a table answer, with the query and the terms beside them | Anyone who shares a figure by e-mail | The working comes with the file |
| The Ask API | The question in; the answer, the shape and the query out | Applications, SmartApps and scripts | The calling person’s access applies |
| A scheduled digest | A saved walkthrough, run to a clock and sent as a summary | Managers and weekly reviews | Each reader sees their own rows |
A finance team that builds its pack in a reporting tool keeps that habit. The connector hands the result set to the report and the report refreshes on its own schedule. Both the pack and the conversation read the same registered tables, so the figures agree.
Every route carries the access of the person who asked. A downloaded file holds the rows that person was cleared for, and the query beside them states how the figure was built.
Showcase estate.
On Wednesday morning L. Tan had her answer: 31 open purchase order lines worth $348,000, inside their cancellation window until Friday 18 September. By the afternoon the question had moved to the finance floor, where J. Harrison asked what those lines would do to quarter-end inventory value.
The first answer came back as a table of four business units. The second, asked as a trend, came back as a line chart across eight months with the budget line on it. The third, asked for the board pack, came back as two sentences. All three read the same rows, each took under four seconds, and each carried the same query.
How context carries from one question to the next, what a new question replaces, and the single clarifying question.
Juno proposes the shape from the wording and the size of the result. The reader switches with one click, and that choice is kept for the next question of the same kind.
Currency, number and date formats come from the mapping and from estate settings, so a table, a chart and a digest all read the same way across teams.
ChatLima shows the leading rows with the measure that ranks them and draws the spread as a chart. The whole result travels through a download or the reporting connector.
Yes. The rows sit one click under the chart, and the query that produced them stays beside the answer, the same query the table ran.
Yes. The questions are shared and the rows stay personal. Two people running one walkthrough each see the plants and customers they are cleared for.
Each one builds on the one before it.
Reads your rows in place.
Connect and move ↗One record per real thing.
Master and govern ↗Apps for every decision.
Operate and decide ↗Asks and acts, at your level.
Ask and act ↗Bring two systems and one decision your planners make every week. In the first week we register the data, confirm the mapping with your team, answer your questions in ChatLima and run the first agent at the level you choose.