AI NativeEvery scan in AI Native runs in one of two modes: demo or live. The distinction is fundamental and the platform makes it visible everywhere. This article explains what each mode does, why they are kept completely separate, and when to use each one.
What demo mode is
Demo mode produces a complete, realistic scan result with no external calls and no credit cost. The answers are synthesised from deterministic fixtures, not retrieved from real AI engines. The metrics are computed from those fixtures using the exact same scoring, sentiment, and accuracy logic that runs on live answers.
The purpose of demo mode is to let you learn the platform, walk through the screens with a client, and understand what a scan result looks like before spending credits on a live measurement. Demo results show a believable set of mention and recommendation rates, ladder positions, opportunity scores, and accuracy findings. They read like real scan output because the same engine that produces live results processes them.
What demo mode does not show is what AI engines actually say about your brand today. The answers are deterministic constructs, not real responses. The numbers illustrate what the metrics mean; they are not measurements you should act on.
What live mode is
Live mode sends your prompts to the real AI engines configured for your product. The answers come back from those engines at the time of the scan, and the resulting rates reflect the actual current state of your AI visibility. Live is the mode you use when you need to know where you stand, track a change you made, or report to stakeholders.
Live scans take time because they involve real external calls, and they consume credits because those calls have real costs. Before a live scan launches, the platform shows you a cost estimate calibrated from your own recent scan history so you know what to expect.
Why the two never mix
Demo and live results are kept in separate series at the data level. The trend lines in the dashboard are built from one mode only. A demo run does not appear in the live trend line, and a live run does not appear in the demo trend line. There is no mechanism for demo results to influence live metrics, and no view that averages across both.
This separation exists because mixing the two would corrupt every trend comparison. A demo run's fixture-based numbers are deterministic and stable by design; a live run's numbers reflect the genuine stochastic behaviour of AI engines. Putting them on the same trend line would make it impossible to read whether a rate change was real or an artefact of switching modes.
Every result in the platform is labelled with its mode. You never have to guess whether a number came from a real scan. The demo label on a result is permanent and explicit.
How demo fixtures are built
Demo answers are synthesised per prompt and per engine using a seed derived from the prompt and engine identifiers. This makes them deterministic: the same demo scan for the same product always produces the same numbers. The synthesis constructs realistic branded and unbranded answer text that includes real facts from your Brand Truth Studio where you have verified them. On a deterministic fraction of branded answers, the fixture deliberately introduces a wrong number so the accuracy checker has a real contradiction to catch.
This means demo mode exercises every part of the platform end to end, including the accuracy audit. If your verified facts are set up correctly, the demo will surface inaccurate verdicts in the same way a live scan would.
When to use each mode
Use demo mode when you are learning the platform, demonstrating it to a new colleague, or exploring the screens without wanting to spend credits. Run a demo scan before your first live scan to understand what the output looks like and to confirm your product profile and prompts are set up correctly.
Switch to live mode when you need real measurements. That means your brand is set up correctly, your verified facts are confirmed, your prompt set reflects the questions your buyers actually ask, and you are ready to see where you actually stand with the AI engines your buyers use. Live is the mode that produces numbers you can report, compare to a prior scan, or act on with confidence.
There is no requirement to run demo before going live. Demo is available whenever you want it, at no cost, for as many runs as you need. But the numbers it produces are illustrative, and the platform makes sure that is always visible.
Questions
Can demo results appear in trend charts alongside live results?
No. Demo and live results are tracked in separate series and never appear on the same trend line. A trend chart built from live scans shows only live scans. The two modes are permanently labelled and kept separate at the data level.
Does demo mode cost credits?
No. Demo scans are free and do not touch your credit balance. Only live scans draw down credits, because only live scans make real external calls to AI engines.
If demo results look bad, does that mean my real AI visibility is also bad?
Not necessarily. Demo results reflect a deterministic fixture set designed to show all the platform's capabilities across the full range of ladder positions. They are illustrative, not a real measurement of your current AI visibility. Only a live scan tells you where you actually stand.
Why is demo not just called "test mode" or "preview"?
Because demo serves a real purpose beyond testing: it is a complete, scored, realistic scan result that you can use to walk through the platform end to end, train a new team member, or present to a client who wants to see what scan output looks like before committing to a subscription. "Demo" describes what it is: a demonstration-grade result, not a measured one.
How do I know which mode a scan was run in?
Every scan result in the scans list and the results view is labelled with its mode. The demo label appears consistently alongside the date and product name. There is no unlabelled scan result.
When I switch from demo to live, what carries over?
Your product profile, verified facts, personas, and prompt set all carry over. What changes is the answer acquisition: live mode sends those prompts to real engines instead of synthesising fixture answers. The scoring, accuracy, and sentiment logic is identical in both modes.
What is the best way to prepare for my first live scan?
Set up your product profile completely, add and verify your key product facts in the Brand Truth Studio, and run a demo scan to confirm the output looks sensible. Check that the prompt set covers the questions your buyers actually ask. Then run a live scan with a modest prompt set and a single region to get your baseline, and use that as the comparison point for every subsequent scan.