How soon after payment do I see the first results?

After payment you configure the project and the prompt set is frozen; the baseline batch is created immediately after that, and results become visible in the workspace as each task completes. We do not promise a specific number of hours, because collection speed depends on how the third-party AI platforms respond and on their anti-automation measures. What we do commit to is that the baseline has the same planned volume as every later cycle, and that a new cycle follows every 72 hours.

Four steps from payment to the baseline

1. Payment: self-serve plans accept WeChat Pay; each payment buys one month with no auto-renewal.
2. Project set-up: enter the brand details (including aliases and the website domain), choose the platform mix (1–3 of the six on the cross-platform plan, fixed for the term once chosen at sign-up) and settle the 50 prompts.
3. Prompt-set freeze: once set-up is complete the prompt set is frozen and the anchor of the 72-hour cycle is fixed at the same moment.
4. Baseline batch created immediately: 50 prompts on each selected platform, each once by default; with three platforms on the cross-platform plan the baseline is 150 planned observations, not a reduced sample.

Why we do not state a number of hours

Every observation runs in a real browser through each platform's web interface. Page timeouts, CAPTCHAs, expired sessions and UI changes can slow or interrupt individual tasks; technical retries keep the original task ID, and tasks that still fail are recorded as failed or blocked rather than dressed up as successes to meet a deadline. Evidence and reports are then reviewed separately by people, and only reviewed observations enter the denominator. That is why /legal/service-levels explicitly makes no promise that any cycle completes within a given number of hours, and offers no public status page or blanket uptime percentage.

After the baseline

• A new cycle is created automatically every 72 hours from the baseline. The anchor is fixed and does not drift with platform response times; expect roughly 10–11 cycles a month.
• Every cycle has the same planned volume, 50 prompts per platform. The baseline is the reference that later cycles are compared against on the same basis.
• Nothing is re-asked outside the cycle once it closes; a single failure costs at most one observation, and the next cycle asks the prompt again.

Tracking progress in the workspace

• The monitoring log lists, per cycle, planned tasks, finalised tasks, valid observations, runtime exceptions (failed / blocked), unconfirmed results, cancelled and pending tasks, together with an evidence-completeness percentage and a runtime-exception rate.
• Every task drills down to its stable task ID, sampling time, status, answer text, screenshot and sources.
• Incidents that affect your data or span several cycles are flagged in the workspace and communicated to your contact email.

Can I get a refund before the baseline has finished?

Both monthly and annual plans are refundable in full before the baseline batch is created. After that, monthly plans are not refunded, while annual plans can be terminated for the remaining term. Full rules are at /legal/refund-policy (Chinese).

Is the baseline result a complete report?

The baseline is shown first as a web report in the workspace, with answer text, screenshot, sources and status per task. The formal PDF report is delivered monthly.

What if one platform fails entirely in the baseline?

Failures are recorded as such, excluded from the denominator and never backfilled. Repeated failures are flagged in the workspace and discussed with you by email or WeCom; see /en/help/failed-runs-and-retries.

Last updated 2026-09-12. The capabilities page and the pricing page are authoritative where they differ.