EstoppelBotSign in

AI processing disclosure

Effective October 6, 2026

EstoppelBot uses AI to save paralegals time. It never makes a legal decision: every AI result is a draft that a person on your team reviews, and nothing is sent to a tenant or issued as a certificate without a human action.

Where AI is used

Extraction reads a lease or rent-roll file you upload and proposes lease fields, or reads a purchase agreement, amendment or access agreement and proposes the deal’s critical dates; you review and confirm before anything is saved or added to a calendar. Response review triages the differences between the lease terms presented and a tenant’s answers by how much they matter to the deal. Every disputed item is always shown, whatever the AI concludes. Drafting proposes certificate wording from an approved response; issuance is blocked unless every tenant exception appears word for word. Each feature can be switched off in Settings → AI.

What the model sees

Only the data needed for the task at hand: the file you uploaded (with, for a deal document, that deal’s name and its dates so far), or one request’s presented lease terms and that tenant’s response. The model has no access to the rest of your workspace, the database, the internet, or other customers’ data, and it cannot take actions.

Provider and training

By default, AI features run on Anthropic’s Claude models through Anthropic’s commercial API, under terms that do not permit using API inputs or outputs to train models. A workspace can instead connect its own account with Anthropic, OpenAI, Google (Gemini), Microsoft (Azure OpenAI), Amazon (Bedrock) or another OpenAI-compatible provider (Settings → AI); calls then go to that provider under the workspace’s own agreement with it, and that provider’s data terms apply. Connected keys are stored encrypted and are never shown again. For providers that don’t read PDFs directly, EstoppelBot sends the document’s extracted text instead of the file.

Connecting your own AI app

Members can connect an AI app they already use (for example Claude, ChatGPT, Cursor or GitHub Copilot) to EstoppelBot through the Model Context Protocol (Settings → Connect your AI). The app can read the workspace data that member can see — deals, tenant responses, exceptions, questions and deadlines — and cannot change anything. What the app does with that data is governed by the member’s agreement with its provider. Access is per member, can be revoked at any time, stops when the member leaves the workspace, can be switched off for the whole workspace by an owner or admin, and every lookup is logged.

Safeguards

Tenant-written text is treated strictly as data, never as instructions to the model. Every AI call is recorded in your workspace’s audit log with its input and output. Usage on EstoppelBot’s key is capped per workspace. See also the Privacy Policy and subprocessors.

AI processing disclosure · EstoppelBot