Anywhere the work is behind a login and there is no API.

The pattern we automate

Almost every use case on this page is one of four movements. Recognise yours in the list below and the rest of the site will describe it precisely.

• Pull: retrieve a statement, report, remittance advice or status page and land it somewhere you can query.
• Push: submit a filing, an invoice, a claim, a renewal or an attestation through a form nobody will ever expose programmatically.
• Reconcile: sign in to two systems that do not know about each other and compare what each one says.
• Monitor: watch for a change: a status flip, a new document, a deadline approaching.

What makes these hard is not the clicking. It is that the front door is a login, often with MFA, increasingly with SSO, and always with a session the site can revoke at will.

What we do not do

We are deliberately narrow, and saying so saves everyone time.

• We do not bypass authentication, solve CAPTCHAs to gain initial access, or circumvent a site’s bot protection. The human signs in during capture. We replay what the site then granted.
• We do not scrape public pages at scale. If the data is public, use a normal crawler. It is cheaper and you do not need us.
• We do not help violate terms of service. If a portal prohibits automated access in a way that binds you, that is your contract to honour, and we will not design around it.

In practice the sites our customers automate are ones they are entitled to access, their own tax filings, their own bank statements, their own claims, and the only obstacle is that the site offers no programmatic door.