Early previewFor small online stores
Know why your store's numbers changed.
When conversions, traffic or site health change, SitePulse AI shows you what moved, the evidence behind it, and what to check next. The numbers come from plain, deterministic code, never from an AI guess.
No sign up. The demo runs on a seeded sample store, not real customer data.
Conversion rate fell 29.9% (3.85% to 2.70%) in the last 7 days compared with the 7 days before.
- Conversion rate2.70%-29.9%
- Checkout errors131+773.3%
- Checkout LCP, p753.89s+69.9%
Checkout errors rose 773.3% (15 to 131) in the same period, starting Sep 22, within a day of when conversion rate started to fall.
Review checkout deploys and payment errors from Sep 22
How it works
From what changed to where to look
One clear order, every time: what the data shows, what might explain it, and what to check. Facts and possible explanations are never mixed.
- Step 1
Evidence first
SitePulse compares your last 7 complete days with the 7 before, metric by metric: conversion rate, sessions, checkout errors, checkout page speed and uptime. Each number shows its method, its dates and how fresh the data is.
- Step 2
Possible explanations, labeled as such
When other signals moved around the same time as your conversion rate, they are listed as candidates, with the day each one started. They are never presented as the proven cause.
- Step 3
What to check next
Each candidate comes with a concrete next step, like reviewing checkout deploys from the day errors started, so you know where to look first.
What makes it different
Built to be checked, not just believed
Dashboards show numbers and chat tools tell stories. SitePulse keeps the evidence and the explanation apart, so you can trust one and question the other.
Numbers from code, not a model
Every metric and comparison is computed by deterministic code you could check by hand. When the AI assistant arrives, it will explain these numbers, not make them up.
Correlation is not causation
Signals that changed together are shown as candidates to check. SitePulse says how confident it is in the data, and why, instead of guessing a cause.
Visible provenance
Source, date range, comparison range, freshness and known gaps sit next to every conclusion. Missing or stale data is called out, not quietly treated as normal.
Where it stands
An honest status
SitePulse AI is early. There are no accounts yet, and nothing connects to a real store today. Here is exactly what exists and what is next.
Works today
Live- An interactive demo on a seeded sample store
- Period comparison with thresholds, data confidence and known limitations
- Candidate explanations and suggested next checks
- Sending feedback straight from the demo
Planned
Not built yet- Connecting Google Analytics 4, the first real integration
- Accounts and a private workspace for your store
- An AI assistant that answers questions from the same evidence, read only
See it on a sample store
Open the demo, read the evidence for a week where conversions fell, and tell us what would make it useful for your store.