Frequently asked questions

The hard questions, answered straight. If your real concern isn't here, it's because the honest answer is "it depends" — and we'd rather say that to your face than bury it.

Can the AI change what my form is for?

No — and not just by policy. The objective (its purpose, the information it must collect, and how you weight success) is locked in the data model itself. There's no button, no endpoint, and no agent path that can change it — the engine only ever reads it.

What if it makes my form worse?

It always keeps a control running and only promotes a change once the live numbers say it's actually better — and only after a minimum amount of traffic, so it can't jump to conclusions. Anything that would regress your goal is rejected before it ever reaches a respondent. And you can pause autonomy whenever you want.

How do I know what it changed?

The Living View is a running timeline: every change, the reason for it, the before-and-after, and what it did to your results — plus the changes it rejected and why. Nothing happens off the record.

Can it trick the metric by asking less?

No. A response only counts as complete when every required answer is actually there and valid. "Ask less, finish faster" lowers the score, not raises it — gaming the metric is designed out, not patched over.

Which channels does it support, and is this spammy?

Web today, plus your own WhatsApp and SMS lines; iMessage is coming. It only ever reaches people who opted in, on channels you own. No scraping, no number-cycling, no spam — that's a hard line, not a setting.

Do I have to write any questions myself?

No. You define the goal, the information to collect, and the rules; the engine writes and continuously rewrites the questions. You can read everything it writes — you just don't have to author it.

Can I pause it?

Yes, anytime. There's a pause-autonomy switch: the current form keeps running, and the objective is always yours. Nothing about your form is held hostage to the optimization loop.

What makes this different from an AI form generator?

A generator writes a form once. This one writes a form, measures it on real traffic, improves it daily, proves the improvement, and can never drift from your goal. The generation is the easy part; the locked, measured optimization loop is the product.

Who owns the responses and the data?

You do. The responses are yours to read and export, and the proprietary record of what works — which phrasings and orders move your reward — accrues to your objective, not to a generic model. Pausing or leaving doesn't strand your data.

How fast will it actually improve my form?

Honestly, it depends on your traffic — the loop learns from real responses, so more sessions mean faster, more confident changes. We won't promise a specific lift, because anyone who does is guessing. What we promise is the discipline: every change is tested against a control and proven before it ships.