Three modes · One quality process

Processing modes for dictation, interviews and assessments

Not every recording is a dictation. Medico-legal experts conduct assessment interviews, doctors hold patient consultations and experts in Swiss IV cases record their interviews. Tippiti applies the same multi-stage quality process to each recording in the mode that matches it.

You choose the processing mode separately for each recording when you upload it. Instruction groups, the editor, version comparison and Word export are available in every mode.

Quick guide

What do you have, and what do you need?

Each path leads to a different result. The editor, Word export and quality control stay the same.

01 Recording
One person dictates

You dictate and Tippiti turns it into the finished document.

DSS DS2 MP3 M4A WAV etc.
[TOOLTIP] Recordings from dictation devices such as Philips or Olympus can be transcribed directly in DSS and DS2 format, without prior conversion. [/TOOLTIP]
02 Processing
Tippiti Dictation

Recognises which document is being written

  • Clinical letters and discharge summaries
  • X-ray, MRI and CT report
  • Medico-legal report, from records review to expert opinion
  • Operation note, progress note and more
03 Result
Finished document Ready

Dear Dr Patel, thank you for referring Mrs Jones, whom I saw in clinic today …

Corrected and ready for your final review

Editor Word API

The use cases show which mode fits which document, from the clinical letter to the Swiss IV assessment report.

Mode · Dictation

A single dictation in, a finished document out

In the standard mode you dictate medico-legal reports, clinical letters or diagnostic reports, and Tippiti produces the document. The recording is treated as one person’s dictation, with one voice and one document. Dictation commands such as “full stop” or “delete that” are carried out, remarks to the typist are removed and the formal register of medical reporting stays intact.

There is no dialogue to preserve and no speaker labels to add, so the result is simply the document you dictated. Why this goes further than speech recognition is explained on the Why Tippiti page.

  • Understands the dictation as a whole
  • Single speaker, no speaker labels in the result
  • Proven for medico-legal reports, clinical letters, diagnostic reports and more
Mode · Conversation & interview

Several voices, one clean transcript

For patient consultations, assessment interviews and other multi-speaker recordings, Tippiti identifies the speakers and labels their roles rather than simply numbering them. You can choose full verbatim, clean verbatim or a summary for each recording. Timestamps and audible behavioural observations can be enabled if needed.

Tippiti recognises dictation commands even in the middle of a conversation. When the doctor or assessor switches to dictating, Tippiti renders the passage as clean text, and what everyone else says remains part of the conversation, never an instruction.

  • Speaker labelling by role
  • Full verbatim, clean verbatim or summary
  • Timestamps measured from the start of the recording
  • Audible behavioural observations in context
Mode · Assessment interview & case history

From the assessment interview to the case history and interview section

Tippiti turns the conversation between a medico-legal expert or doctor and the person being assessed into the case history and interview section of the report. When you upload the recording, you select the report type. Tippiti then follows the relevant outline for a psychiatric report, another speciality or a Swiss IV assessment report.

Completeness takes top priority. The validation checks every section of the outline against the verbatim transcript. Tippiti clearly identifies any section not covered in the interview, keeping the basis of the report transparent.

  • Report-ready case history and interview section
  • Outline per speciality (psychiatric, Swiss IV and more)
  • Completeness validated against the transcript
The common core

Every mode includes the same quality checks.

The modes differ in their source material and their result, not in the care taken. In every mode, several AI models with distinct roles work together until the result meets the requirements.

01 · Transcription

Capture the recording verbatim

A known weakness of AI transcription is that models condense what they hear. Repetitions are dropped, statements are merged and entire passages may disappear.

Tippiti’s transcription step is designed specifically to counter this. It captures the entire recording word for word and, in conversations, also attributes each statement to its speaker.

02 · Correction

Correct and structure

This is where the transcript becomes a document. Depending on the mode, the raw text is corrected, the speaker-labelled transcript is refined, or the conversation is transferred into the report outline.

This step decides what is content and what was an instruction. Dictation commands are carried out and remarks to the typist removed. This stage also runs a fact-check[TOOLTIP] The fact-check verifies details that can be checked against a source, including proper names, medication names, ICD codes and place and postcode combinations. Clinical findings and diagnoses come from the clinician or medico-legal expert. [/TOOLTIP] and a plausibility check[TOOLTIP] The plausibility check looks for clear logical errors, further reducing transcription and dictation errors. [/TOOLTIP], and applies the rules in your instruction groups.

03 · Validation

Check against the original

An independent validation stage compares the result with the original recording. Has everything been carried over? Has anything been invented? Were the rules followed?

04 · Re-correction

Re-correct and check again

Where discrepancies are found, Tippiti re-corrects the affected passages and validates them again until the validation criteria are met.

How this process works in detail is shown in the interactive demo and on the Why Tippiti page.

Pricing

What the modes cost.

No base fee, no minimum term, no subscription. The mode determines the price per character, and nothing else.

Good to know

Frequently asked questions about the processing modes

What is a processing mode?

The processing mode determines how Tippiti interprets a recording and what result is produced:

Every mode includes correction, fact-checking and validation.

Which settings are available for each mode?

That depends on the mode:

  • Conversation & interview: speaker labelling, transcript style (full verbatim, clean verbatim or summarised), timestamps and behavioural observations.
  • Assessment interview & case history: report type (for example a psychiatric report, another medical speciality or a Swiss IV assessment report) and an optional concluding summary.
  • Dictation: no additional settings needed.

You can further customise the output in every mode by using instruction groups.

How do I choose the mode?

You choose a mode for each recording when uploading it. The upload dialog shows the available modes and the settings for your selected mode. For a batch upload, you can apply the same selection to all files.

You can also specify the mode for each recording through the API.

Does every mode go through the same quality check?

Yes. Correction, fact-checking and plausibility checking work together in every mode. An independent validation stage checks the result against the original recording, whether the output is a dictation, a conversation transcript or a case history section written up from the interview.

The modes differ in what the recording becomes, not in how thoroughly it is checked.

Do instruction groups work in all modes?

Yes. You can assign instruction groups in every mode. Rules for abbreviations, formatting and text substitutions are applied during the relevant correction or structuring stage.

Can Tippiti record audible behavioural observations?

Yes. When behavioural observations are enabled, Tippiti notes audible non-verbal events such as [laughs] or [cries] directly in the transcript.

What does each mode cost?

Dictation and Conversation & interview cost €0.50 per 1,000 characters of the finished text. Assessment interview & case history costs €1.50 because it also structures the conversation according to the report outline and validates completeness against the full transcript.

There is no base fee or minimum term. The pricing overview shows the details at a glance.

See more frequently asked questions and answers

See for yourself.

Upload a test dictation and see what Tippiti makes of your recording. Try it free with no registration and no commitment.

Create account

Loading…