Frequently asked
questions.

600 minutes of meetings when you sign up, with the core product intact: the notetaker bot on all four platforms, Arabic and English transcription with speaker names, AI reports and templates, action items, email recaps, and PDF export. No credit card required.

A minute of meeting audio processed. A 40-minute meeting uses 40 minutes whether the bot joined a Zoom, Google Meet, Microsoft Teams, or Webex call, you recorded in the browser, or you uploaded an audio file. Generating reports, editing them, translating, and searching never consume minutes.

Recording continues when the allowance is reached; processing can wait for additional minutes. Your transcripts and reports stay available, and to process new meetings you top up (500 minutes for $9, 2,000 for $29) or upgrade your plan. For plan changes and available checkout options, contact us.

Per seat. Each seat on Pro gets 1,200 minutes a month, each seat on Business gets 3,000. Annual billing is equivalent to two months free: $180/seat/year on Pro, $300/seat/year on Business. Larger teams and volume terms: talk to us.

Self-serve plans are priced in USD. Enterprise and government agreements can be invoiced in AED or SAR, and applicable tax is itemized separately. AWS Marketplace private offers are available for procurement through your existing AWS billing.

Deployment and control. Enterprise runs MeetriX in a private cloud or fully on-premise on your own GPUs, white-labels the bot with your name and branding, adds institution administration and deployment-specific controls, and manages custom vocabulary at the organization level. See security and on-prem.

On-premise is scoped per deployment, since it runs on your hardware inside your network. Book a demo and we will walk through the architecture, GPU requirements, and a plan sized to your institution.

No. You can add an online meeting link directly, record in your browser, or upload a recording. Calendar connection adds scheduling and joining rules.

Your joining mode, scope and title-keyword rules determine which calendar meetings it joins. Hosts may still need to admit the assistant.

Yes. Structured meeting records have editing controls for the title, topic, attendees, summary, discussion points and recommendations.

Editing and saving are available. If your organisation needs reviewer stages, formal sign-off or a versioned approval process, discuss that requirement during implementation.

The presentation shows an illustrative reporting concept. Confirm the dashboard, metrics and data sources for your deployment during a demonstration.

No. The reported character-error results evaluate Arabic and English separately. The 40 languages and 32 Arabic dialects describe product coverage.

Yes. M3 handles Arabic-English switching within one system. Test the dialects, terminology and recording setup used by your team during a pilot.

Google Calendar and Microsoft Outlook are supported for meeting discovery and scheduling.

Yes. Set title-keyword blocking rules and choose the appropriate joining mode and internal/external scope.

Connections need the appropriate accounts, permissions, destinations and actions. We agree the required workflow during setup.

Yes. Notion pages and Microsoft Teams channel delivery have destination and automatic-sending controls. Email and PDF provide additional distribution options.

Private-cloud and on-premise deployment are available for enterprise planning. We scope the processing boundary, hardware, connectivity and operating responsibilities with your organisation.

Online meeting capture still connects to the meeting platform, and enabled external integrations need their own connections. Local recording workflows and processing boundaries are agreed for your environment.

A Meeting Type defines what information to extract. A Report Template controls how that information is arranged and presented.

A time box forces a scope decision before the meeting instead of during it. When an item routinely blows its box, it is either two items or a working session pretending to be an agenda line.

Around 10 to 15 percent of the meeting. A 45-minute meeting with 40 minutes of items is realistic; one with 45 planned minutes ends late.

Copy it with one click and paste it into the calendar invite, a document, or a chat. Nothing is stored on our side; the tool runs entirely in your browser.

Yes. When a meeting runs through MeetriX, the discussion is checked against the agenda: covered items are confirmed and deferred ones are flagged in the report and in analytics.

Attendees × average hourly cost × meeting length in hours. Weekly cost multiplies by how often the meeting happens; annual cost assumes 48 working weeks.

Use the fully loaded cost: salary plus benefits, overhead, and employer costs. A quick approximation is annual cost divided by 2,000 working hours.

No. The calculator runs entirely in your browser; nothing you type leaves the page.

Shorter meetings with real agendas, fewer standing meetings, and records good enough that nothing needs re-discussing. MeetriX tracks length, punctuality, and agenda coverage automatically in analytics.

As a rule of thumb: the most active voice under 40% and no participant under 10% reads as a healthy discussion. One voice over 50% means the meeting was effectively a briefing.

Estimate them; the shape matters more than precision. If your meetings run through MeetriX, per-speaker talk time is measured automatically from the transcript.

No. A briefing, a training session, or a board presentation is supposed to be one-sided. The verdict matters for meetings that claim to be discussions or decisions.

No. The analyzer is plain JavaScript running in your browser; there are no network calls.

An hour is highlighted when the local time in every selected city falls inside 9:00 to 18:00. Contiguous highlighted hours are suggested as slots.

Approximately. The offsets are fixed summer 2026 values (London +1, Paris +2, New York -4, San Francisco -7). Verify against a calendar before sending invites near DST transitions.

The planner says so. At that point the honest answer is rotation: alternate who takes the out-of-hours call, and make sure the record reaches whoever missed it.

They cover the corridors GCC organizations schedule across most: the Gulf, the Levant and Egypt, Europe, both US coasts, India, and East Asia. Tell us at support@lisan.com if yours is missing.

Nothing: they are two names for the same thing, converting spoken audio into written text. ASR (automatic speech recognition) is the research term; speech-to-text is the product term.

Diarization splits audio into "who spoke when". MeetriX uses available participant channels or speaker-activity information to help attribute contributions. Identity information depends on capture conditions and may need review.

Lower is better: 5% WER describes five word-level edits per hundred reference words. M3 reports 1.62% Arabic CER and 1.83% English CER in Lisan’s separate September 2026 evaluations. CER is not WER; read the scoring method at /research/m3/.

No, not today. Live in-call captions are on the roadmap. Today the bot records during the meeting and the transcript and reports are prepared after processing. See the notetaker bot.

Because two owners means each assumes the other has it, and a team as owner means nobody does. The owner is not necessarily doing all the work; they are the single person accountable for it getting done and for reporting back. Split shared work into separate items, one per owner.

Convert it before the topic moves on: "Who takes that, and by when?" It takes thirty seconds in the room and feels natural. Two days later over email it feels like an accusation, so it rarely happens at all.

Then the deadline is for the next step, not the whole task: "Faisal to scope the migration and propose a plan by Jul 24." A big vague task with no date becomes a small concrete task with one. There is always a next step that fits in a week.

Out loud, at the review: the owner gives a new date and the list is updated. What kills action lists is not slippage, it is silent slippage. One openly re-dated item preserves the norm; one ignored item teaches everyone the dates were fiction.

Both have a role: the minutes are the record of what was agreed, and a living list (task tool, shared doc, or your notetaker's action-item view) is what you review week to week. What matters is that exactly one list is the working list, and that it opens the next meeting.

About three real items for a 45-minute meeting, each with an honest timebox and a short buffer at the end. If your list is longer, some items are status updates that belong in writing, or the timeboxes are wishful thinking.

The meeting organizer owns the agenda, but each item is owned by the person driving it. A useful pattern: the organizer collects proposed questions from attendees, cuts the list to what fits, and sends the final version about 24 hours ahead.

Cancel it, or convert it to a message thread. This feels rude exactly once. After that, people learn that meetings on your team come with agendas, and the quality of every meeting rises.

In the pre-read, or in a written channel. Reading a status aloud to a room is the most expensive way to transfer information a team has ever invented. Reserve live time for the two things writing does badly: deciding and debating.

More, not less. Tools like MeetriX write the summary and pull out the action items for you, but they document the meeting you actually had. A meeting shaped by clear questions produces decisions worth recording; a wandering one produces a well-documented wander.

Usually not. Forcing a single language pushes people to argue in their second language, and the quality of thinking drops. Let the room switch naturally, have the chair recap across the language line after long exchanges, and fix a single language for the written record instead.

Moving between two languages within one conversation, often within one sentence: an Arabic sentence with an English technical term embedded, or an English discussion with an Arabic aside. It is normal, efficient behavior in bilingual teams, not sloppiness to be corrected.

Remote speaks first. On every agenda item, the chair invites the remote attendees before opening the floor to the physical room. It reverses the natural advantage of the table and costs nothing.

Because attribution is the first thing lost at a distance. Remote listeners cannot see who is talking at the table, and any record of the meeting needs to know who committed to what. It feels awkward for about a week, then becomes automatic.

By hand, barely: the note-taker needs both languages and misses whatever crosses the line mid-commitment. The practical answer is to record and get a diarized transcript that keeps each line in its original language with the speaker named, then write minutes from that. This is the setting where AI notetakers built for Arabic and English earn their keep.

It depends on the meeting type. In a discussion or decision meeting of four or more people, no single voice should sit much above 40%, and the chair should be low on Decide items. A briefing is legitimately lopsided; the problem is a briefing pretending to be a discussion.

Briefly, and mostly in a good way: dominant speakers are almost always unaware of it, and seeing the number is usually enough to change it. Keep the data at the meeting level, review it as a team, and never use it in performance reviews. It is a mirror for the meeting, not a metric for the person.

Publish the number. Track the percentage of meetings starting within two minutes and show it monthly; most teams jump within weeks because nobody wants to be the reason it is red. Structural helps: 25 and 50 minute defaults instead of 30 and 60, so back-to-back meetings get a buffer.

No. The healthy sign is length matching the agenda: three timeboxed questions should fit 45 minutes, and a meeting that ends early because the questions were answered is a win. The unhealthy sign is drift, a recurring meeting growing quarter over quarter without gaining scope.

On-time starts move within a month of being published. Length follows agendas: fix the agenda discipline and the average drops within a cycle or two. Talk-time balance is slowest because it is habit; expect visible change over a quarter of monthly reviews.

You can start with a stopwatch and a spreadsheet: note the start time, the length, and roughly who spoke. It works, but it decays because someone has to do the bookkeeping forever. Meeting analytics from a notetaker automate the collection; keep the monthly ten-minute review human either way.

As a default, no. Minutes record outcomes: decisions, action items, attendance. Attributing positions makes minutes longer, slower to approve, and more political. Formal governance settings sometimes require recorded positions or votes; if yours does, record them precisely and keep everything else outcome-only.

A transcript is evidence of what was said, but most organizations treat the approved minutes as the official record, because minutes go through review and sign-off. Requirements vary by jurisdiction and by body (a board has different obligations than a project team), so check what your governance rules actually require.

The same day, ideally within a couple of hours. Fresh minutes get corrected while people still remember the meeting and acted on while the commitments feel current. Every day of delay costs accuracy and momentum.

The Arabic term (محضر اجتماع) for formal meeting minutes. In many Arabic-speaking institutions it is a signed, filed instrument: written in formal Arabic regardless of how the meeting was spoken, with numbered decisions, full names and titles, and the chair's approval before filing.

Whichever language your institution files records in, and say so on the document. A bilingual pair is useful for mixed teams, but designate one version as authoritative. Tools that generate minutes in Arabic or English from the same transcript make the pair cheap to produce.

The draft can come from anywhere; what makes minutes official is the process around them: a named person reviews the draft, the chair approves it, and the approved version is circulated and filed. Treat AI output as a fast first draft that still earns its status through review.

MeetriX turns the conversation into a structured record: a summary, discussion points, decisions, and follow-up responsibilities. Review and edit it, choose a report layout, and share by email, PDF, or configured integrations.

MeetriX is designed for multilingual meetings, Arabic dialects, and Arabic-English code-switching. M3 supports 40 languages and 32 Arabic dialects. Arabic and English evaluation results are reported separately. Try a representative recording to evaluate your own context.

Yes. Select meetings manually, join meetings you host, or cover all calendar meetings. Configure meeting scope and title-keyword rules. These controls determine attendance; they do not automatically detect sensitive content.

On-premises and sovereign or private cloud deployments are available for enterprise projects. Infrastructure, retention, access, and integrations are scoped with your organisation during a pilot.

Start through meetrix.ai, or plan a pilot around your languages, meeting types, and deployment requirements. See current plans.