Skip to content

How Telha reads your content

For administrators

Once a source is connected, Telha reads each item and turns it into memory: not a copy of the file, but the facts inside it, each one linked back to the exact place it came from. This page explains what Telha can read and what happens to it.

Mostly automatic

You do not run this yourself. Once your sources are connected, Telha reads new and changed items on its own. This page is here so you know what to expect and what formats are supported.

What Telha can read

Content Examples What Telha understands
Documents PDF, Word, PowerPoint Headings, sections, paragraphs, tables, page and position, so answers can cite the exact spot
Spreadsheets Excel, CSV Each row as a record, with typed columns (numbers, dates, text)
Web pages Intranet and public pages The readable content, headings, and links
Email Messages and attachments Sender, recipients, subject, thread, and the body, with attachments read too
Code Source repositories Files, functions, classes, and how they reference each other

Attachments are read as well: a spreadsheet attached to an email, or a PDF in a Slack thread, is picked up and understood in its own right.

From a document to memory

When Telha reads an item, it does three things:

  1. Extracts the facts and how they relate, keeping the structure (which section, which row, which page).
  2. Records where each fact came from, down to the exact passage, so every answer can show its sources and link back to the original behind its own login.
  3. Notes the dates the item carries, so Telha can answer questions about how things were at a point in time.

When something is unclear

If Telha meets something ambiguous while reading (a date that says "next quarter", two figures that disagree, an undated scan), it does not guess silently. It makes a best guess at lower confidence and asks the right person to confirm. That is a clarification, and answering it makes the memory more accurate.

Re-reading and duplicates

  • Unchanged content changes nothing. If a document is re-synced but its content is identical, Telha recognises it and does no extra work, so nothing is duplicated and no one is notified for a non-change.
  • A new version is remembered as a new version. When a document actually changes, Telha keeps both, so you can always see what it said before and after. See Understanding your answers.