Your Pre-Live Checklist for AI Live Selling

Des Damakov
Your Pre-Live Checklist for AI Live Selling

The most expensive live stream is the one that drops ten minutes in because something was not checked. This pre-live checklist for AI live selling groups the checks into tech, content, compliance and go-live so an operator can clear a room for launch in minutes.

Ready to put an AI host live on your shop?Get StartedSee how it works
Syntopia BOX live-streaming software interface
Inside Syntopia BOX — the AI live-selling software you run on your own PC.
⚡ Key takeaways

  • Dropped streams are almost never mysterious — they are a skipped step, so run the same runbook every time.
  • Tech first: PC spec, a stable wired network, the RTMP/capture feed, the correct account and the right BOX session per machine.
  • Content: avatar and trained voice loaded, scene set, catalog and product cards linked before launch.
  • Load the slot’s script with claim guardrails applied, and confirm AI disclosure and a supported platform.
  • Never go live without a named operator, an analytics wall and a decided drop-recovery plan — especially for unattended runs.

The most expensive live stream is the one that drops ten minutes in because something was not checked. A pre-live runbook — the same checks, run the same way, every time — is what keeps rooms up and selling instead of failing silently. This checklist for AI live selling groups the checks into tech, content, compliance and go-live, so an operator can clear a room for launch in minutes and catch problems before viewers do.

Why a pre-live runbook prevents dropped streams

Dropped streams are almost never mysterious — they are a skipped step. A loose network cable, an account that was not logged in, a scene that was not loaded, a claim that should have been pulled. A runbook converts hard-won lessons into a standing list so nobody has to remember them under pressure. At one room it is a habit; across many rooms it is essential, because without it the same avoidable failure repeats on every machine. The discipline is the same one pilots use: the point of a checklist is not that the steps are hard, but that skipping one under time pressure is easy and costly. Run it the same way, in the same order, whether it is your first stream of the day or your fortieth, and the room that would have dropped simply does not.

Tech check

Start with the plumbing, because everything else rides on it:

  • PC and spec — the machine is up and meets the hardware requirements for the room it will run.
  • Network — a stable, ideally wired connection; weak Wi-Fi is the most common cause of a drop.
  • RTMP and capture — if a real host is driving rooms, the camera, capture card, OBS and local RTMP feed are confirmed and each room PC is pulling the feed cleanly.
  • Accounts — the right shop account is logged in on the right room, with no mix-ups.

One BOX session runs one room — confirm the right session is on the right machine before anything goes live.

Content check

With the plumbing confirmed, check what the room will actually show. The avatar and the trained voice are loaded and correct for this room and language; the scene is set and on-brand; and the product catalog and product cards are linked so viewers can buy the moment they are convinced. A room that goes live with the wrong scene or unlinked cards is live but not selling — worse than not being live at all, because it burns airtime looking broken.

Script and guardrails

Load the script for this room and this slot, and confirm the claim guardrails and banned-phrase list are applied. Because the host will repeat the script continuously, this is the moment to catch an unsafe claim — before it runs on a loop, not after a flag. Check that the room’s script is distinct from your other rooms’ scripts too, so concurrent rooms are not running near-identical streams.

Compliance and disclosure

Run the policy checks that keep the room online. Confirm the platform disclosure that the host is AI is in place as that platform requires, that you are on a supported platform, and that nothing in the stream trips the rules. This is the pre-flight version of moderation and compliance and overlaps with how to hand off to a human if a live issue arises. Clearing compliance before launch is far cheaper than clearing a strike after.

Go-live and monitoring plan

Finally, confirm how the room will be watched once it is up. Who is the operator, what does the analytics wall show, and what is the plan if this room drops — auto-restart or wait for the next check? For scheduled and unattended runs this matters most, because nobody is sitting at the screen. A room should never go live without a named answer to “who notices if this stops, and how fast”.

Printable pre-live checklist

TechPC up to spec, stable wired network, RTMP/capture feed confirmed, correct account logged in, right BOX session on the right machine.
ContentAvatar and trained voice loaded, scene set and on-brand, catalog and product cards linked.
ScriptCorrect script and slot loaded, claim guardrails and banned-phrase list applied, script distinct from other rooms.
ComplianceAI disclosure in place per platform, supported platform confirmed, nothing that trips the rules.
Go-liveOperator assigned, analytics wall ready, drop-recovery plan decided — especially for unattended runs.

For the full room build behind each line, keep the complete Syntopia BOX guide alongside this checklist.

Frequently asked questions

What should I check before going live with an AI host?

Work through four groups: tech (PC, network, RTMP/capture feed, correct account and the right BOX session on the right machine), content (avatar, trained voice, scene, product cards), script (correct script with claim guardrails applied), and compliance (AI disclosure and supported platform). Then confirm your go-live monitoring and drop-recovery plan.

Why do AI live streams drop, and how do I prevent it?

Drops are almost always a skipped step — a weak network, an account not logged in, a scene not loaded. A pre-live runbook run the same way every time converts those lessons into a standing checklist, so the same avoidable failure does not repeat on every machine. A stable wired connection prevents the most common cause.

Do I need a disclosure that the host is AI?

Confirm and apply whatever disclosure the platform you are using requires, as part of your pre-live compliance check. Clearing disclosure and the other policy checks before launch is far cheaper than clearing a strike afterwards, and it is a standing line on the checklist for every room.

How do I run a pre-live checklist across many rooms?

Use the same grouped runbook — tech, content, script, compliance, go-live — on every room, and confirm each room has the right BOX session, account and a script distinct from the others. Across many machines a shared checklist is what stops one avoidable failure repeating everywhere.

What goes in the go-live step?

A named operator, a ready analytics wall, and a decided drop-recovery plan — auto-restart or wait for the next check. This matters most for scheduled, unattended runs: no room should go live without a clear answer to who notices if it stops and how fast.

Ready to put an AI host live on your shop?

About the author

Des Damakov
Article written by

Desislav Damakov

I’m the Co-Founder and CEO of LiveBuzz Studio

Get Monthly Newsletter in Your Inbox

Share Post

FAQ

Frequently asked questions

SYNTOPIA creates hyperrealistic, interactive AI avatars that revolutionize how businesses and individuals connect with their audiences. Our avatars can:
How long can an avatar live broadcast?

Once the avatar has been trained, it can live broadcast depending on the plan you have purchased.

No, unused hours do not accumulate. Every month, the hours reset.

Yes, you can order a custom avatar. There are two options:

  1. Record a model yourself, submit the model to us, and we will upload it to your Syntopia platform.
  2. We can shoot a model for you.

If you choose to record yourself, please follow the shooting guidelines here: Custom Avatar Shooting Guide.

Yes, the custom avatar will only be visible to you on your Syntopia platform.
When submitting a custom avatar, we require a model release form from the model, as well as an ID, to ensure that the model has granted appropriate permission for the use of their likeness.
Training an avatar involves uploading the model data or shooting the model, after which our system processes and fine-tunes the avatar to ensure it behaves and interacts as expected during live broadcasts. The training time can vary depending on the complexity and type of avatar.

Ready to try our AI Video Platform?

The first ever hyperrealistic interactive digital live streams. Yes, they can reply to comments, greet viewers, and sell your products or services 24 hours a day!

🇬🇧English