Platform

AI workflow platform: five building blocks that carry real work together.

An assistant that builds. Skills and workflows that run. Documents and your own database for agents to work on. And workflows that reach outside the company. All within one frame of approvals, logs, and undo.

Platform

Matching integrations

  • SiSiigo

    Read and create invoices, customers, and postings in Colombian accounting.

  • DIDIAN

    Fetch electronic invoices from DIAN and store them as records with line items.

  • ShShopify

    Read orders, customers, and products from the shop.

  • AtAirtable

    Read tables and create or update rows.

  • NoNotion

    Read pages and databases, create entries.

  • SuSupabase

    Query and write tables in your own Postgres database.

  • SPSharePoint / OneDrive

    Import folders from SharePoint and OneDrive, sync changes automatically.

  • @E-Mail (IMAP / SMTP)

    Read mailboxes via IMAP, send emails via SMTP. Attachments become documents.

Common questions

No. Most teams start with documents and one skill: connect a folder or a mailbox, describe the task, run it, look at the log. Workflows come in when the task needs a trigger or an approval, records when results should live somewhere structured, and partner workflows when a process crosses the company boundary. Each block is useful on its own and gets more useful with the others.

OpenAI, Anthropic, and Google Gemini are connected as integrations. You choose per skill which model works, can restrict the choice per space, and see the cost of every run per step. Switching a skill to another model is a one-line change in the skill, not a migration.

In the EU, separated per company and encrypted. Integration credentials are stored encrypted and redacted in logs. Partner workflows exchange only what both sides agreed in the handoff; the partner never sees your spaces, documents, or records.

Automation tools move data between systems. adlass keeps the shared state the team and the agents work on: documents, records, and the history of every change. The workflow builder is part of it, but the point is that a run leaves a traceable, reversible result rather than a webhook payload.

Describe the task. The rest takes shape on the platform.

Start for free, create your first skill with the assistant, and see what a run looks like before you connect anything.