Earlier this year, we introduced Navi as a new interface to the Parsewise Data Engine. Instead of learning the platform first, users could begin with the questions they wanted answered and the data sitting in their documents.
Navi can now inspect the decisions behind an extraction, transcribe dictated prompts, and use screenshots as context.
Decision transparency
Parsewise results have always been traceable to their source. Citations were included to help you find the source of a value, but they do not explain why a value is missing or unexpected.
To inspect a decision, Navi retraces the extraction path for a specific Agent, document, and page. Navi checks whether the page was judged relevant, whether a value was returned, and where the path stopped.
A user can ask Navi:
“Why didn’t page 5 give me a Net Debt value?”
Navi inspects the original decision flow, then replays extraction against the selected page. Here, page 5 discusses Gross Debt rather than Net Debt, which is why Navi skipped it.
Voice dictation
Longer, more detailed instructions are often quicker to say than type. Tap the microphone, say what you need, and Navi adds the transcript to your draft so you can review or edit it before sending.
Screenshot context
Not everything is easy to copy or describe. Sometimes a screenshot is quicker. Paste one into Navi, ask your question, and Navi uses the image as context for its answer.
Here are the financials shown in the chart:
| Company | Revenue | EBITDA |
|---|---|---|
| Aster Group | £42.8m | £6.1m |
| Beacon Ltd | £31.4m | £4.7m |
| Cobalt plc | £27.9m | £3.8m |
But wait, there’s more!
Part 2
Navi, beyond chat.
Find out tomorrow · 9 September 2026