Getting started

Welcome

This guide describes the first steps with Kontrakte.AI: logging in, configuring your environment, and running your first checks. Each section includes the required inputs, the resulting output, and a reference to the screen on which the action takes place.

Logging in

Open https://app.kontrakte.ai. Bookmark the address for future use.

The entry page offers two actions:

• Log In — for users who already have an account.
• Register — for users who want to create an account through self-registration. An account can also be provisioned for you after contract conclusion; in that case skip this step and use Log In directly.

The interface language can be changed at any time via the flag icon in the top right corner of the entry page.

After clicking Log In, the Sign in to your account form opens.

• Enter your Email and Password.
• Use the eye icon next to the password field to toggle password visibility.
• Activate Remember me to remain authenticated on the current device.
• Use Forgot Password? to receive a reset link by e-mail.
• The interface language can also be selected directly inside the login card.

Confirm with Sign In.

Self-registration

Clicking Register on the entry page opens the registration form. Self-registration creates a test account intended for evaluation and demonstration of the Kontrakte.AI platform; it is not designed for productive use.

Provide the following information:

• Company name — the name of your legal entity.
• First and last name of the account holder.
• E-mail address — the address used as login identifier and for system notifications.

Read the Usage Policy on the right-hand side of the form, then activate the I accept the terms of use and would like to test the Kontrakte.AI platform checkbox and confirm with Register. Use Back to return to the entry page.

A successful registration starts a 14-day free trial that includes 50 ContractCoins. ContractCoins are the usage-based units consumed by AI analysis features and certain functions, depending on your tariff.

After submission a confirmation screen is displayed and a message is sent to the e-mail address provided in the form. Open this message and follow the instructions to complete the activation of your account; only afterwards can you log in.

The activation e-mail contains a link. Opening this link in a browser leads to the Perform the following action(s) page, which lists the steps required before the account can be used. For a self-registered account these are typically:

• Verify Email — confirms that the e-mail address belongs to the account holder.
• Update Password — sets the password used for future logins.

Use Click here to proceed to start the listed actions.

The first action prompts you to define a password for the account. Enter the new password in New Password, repeat it in Confirm password and confirm with Submit. The eye icon next to each field toggles password visibility.

The password must be at least 12 characters long.

Activate Sign out from other devices if you want existing sessions on other devices to be terminated when the new password becomes effective.

After successful submission a confirmation screen is displayed. Use Back to Application to return to the regular Log In flow; the account is now active.

First-time setup

On first login, a multi-step setup wizard is displayed as a series of modal dialogs over the home page. Each step ends with a Continue button that advances to the next step. The values entered during the wizard can be changed later in the Settings area.

Step 1 — Welcome

The first step introduces the platform with a short welcome message. Confirm with Continue to proceed.

Step 2 — Company name

Enter the name of your legal entity. This value is used so that documents are read and evaluated from your organisation's perspective; it must match the official name of the legal entity exactly. Confirm with Continue to proceed.

Step 3 — Language and law

The final step collects two required values:

• Language — the language used for analysis output.
• Law — the legal framework against which documents will be evaluated, for example English law.

Both values serve as defaults and can be overridden for each individual analysis when starting a new process. Use Back to revisit a previous step or confirm with Get started now to close the wizard.

After confirmation the wizard closes and the home page becomes fully usable.

The home page

After login (and once the first-time setup wizard has been completed) the home page is displayed as the central entry point.

Layout

The home page is structured as follows:

• Top bar — navigation menu (icon, top left), the current page title, a Contact us button and the Kontrakte.AI logo.
• Module cards — entry points to the main feature areas, arranged around the centre of the page. Curved connection lines visually link each module to the Tutorial Hub.
• Tutorial Hub — central element of the page; see Tutorial Hub below.
• Language switcher — flag icon in the bottom right corner.

Trial-specific elements

Accounts created through self-registration are trial accounts. On a trial account the home page additionally shows:

• a banner at the top of the page (“You are currently in a free trial version — Contact our experts now to get the full version.”),
• a Get in touch call-to-action panel in the bottom right corner. The panel can be dismissed via its close icon.

On accounts that were provisioned after contract conclusion, neither element is displayed.

Modules

The home page exposes the following entry points:

• Smart Check — quick risk assessment of a contract with clear guidance on how to proceed.
• Playbook Check — review of a contract against your own rules, guidelines and standards.

Depending on the features included in the booked tariff, additional modules may appear on the home page.

Tutorial Hub

The Tutorial Hub sits at the centre of the home page and provides guided tours for the available modules. A small badge above the hub indicates the number of tutorials still to discover (for example “3 tutorials to discover”); the badge can be closed via its dismiss icon.

Activating the hub opens a full-screen overlay with the available tutorials. The header shows the title Tutorial, an overall progress indicator (0% – 100%) that reflects the share of tutorials already completed and a close icon (top right) to dismiss the overlay.

Each available tutorial is started via its own Start now button. The set of tutorials shown depends on the features included in the booked tariff; more or fewer tutorials may be visible than in the screenshot below. The screenshot shows the tutorials available on a trial account:

• Smart Check — guided introduction to the Smart Check module.
• Checklists — guided introduction to checklists and how to use them in a Playbook Check.
• Playbook — guided introduction to playbooks and the Playbook Check module.

Following a tutorial

Activating Start now on a tutorial closes the Tutorial Hub overlay and starts a tooltip-based walkthrough on top of the regular interface. Each step contains a short explanation, a progress bar at the top, a close icon to leave the tour at any time and a Continue button to advance to the next step. The detailed contents of each tour are not documented here; follow the on-screen instructions of the tour itself.

Use the Tutorial Hub for a structured introduction to the platform; the recommended sequence is described in Recommended next steps.

Running a Smart Check

The Smart Check is the AI-supported initial review of a single document against general law. It can surface possible legal infringements, ambiguities and potentially critical wording for further review.

Creating a new check

Open the Smart Check page from the home page or the main navigation. The page lists all Smart Checks already created and provides a Check New Document entry point. Activate Check New Document to start the two-step wizard.

Step 1 — Files

Upload the main document by drag-and-drop or via Browse Computer. Only PDF files are supported. Optionally add a reference document; if provided, the two documents are compared during the check to detect deviations and inconsistencies. Confirm with Continue.

Step 2 — Check details

The system pre-fills key metadata extracted from the uploaded document. Review and adjust as needed:

• Document and Optional Reference document — the uploaded files are shown with their filename and can still be replaced or edited.
• Parties — Own Party (auto-detected) and Other Parties (comma-separated). Use the swap icon to invert the assignment between own and other parties.
• Properties:
  • Name (required) — name of the document check.
  • Contract Type (comma-separated).
  • Language — pre-filled from your environment settings; can be overridden for this check.
  • Law — pre-filled from your environment settings; can be overridden for this check.
• Further options — collapsible area with additional configuration options.

Above the Start Check button, an indicator shows an estimate of the ContractCoins the analysis will consume with the current settings (for example ≈2 Coins required). The actual amount may vary depending on the document and the chosen options. Use Start Check to run the analysis or Back to return to step 1.

Processing and result tile

After Start Check, a new tile appears in the Smart Check overview. The tile shows the document's designation, description, creation metadata (Created on, Created by, Assigned to) and a status block:

• Initial processing — the AI is analysing the document; a progress bar indicates the current progress.
• Ready — analysis is complete. The tile then displays a summary of all findings, grouped by severity (errors, warnings, notes) plus a total count. Open evaluation opens the detailed result view.

Evaluation view

The evaluation view is structured in three columns plus a top bar.

Top bar

• Document title with an info icon for additional metadata.
• Review progress — a status bar (for example Not checked) that indicates how many of the AI's findings have already been reviewed by the user, and how many are still outstanding. The bar fills as findings are accepted or rejected.
• Action icons in the top right for sharing the results — see Sharing the results below.

Document viewer (left)

The uploaded document is rendered with navigation and zoom controls. Passages relevant to the analysis are visually highlighted; the highlights are kept in sync with the centre column.

Annotations list (centre)

Findings are listed as cards. Each card shows the section number, the quoted passage from the document and one or more severity badges indicating the type and number of issues. Selecting a card loads its details in the right side panel and scrolls the document viewer to the corresponding passage.

The toolbar above the list provides:

• Filter toggles by severity type — Error, Warning and Note.
• A clear-filter action and an indicator of the total number of annotations.
• A free-text search across the annotations.

Side panel (right)

The side panel uses tabs whose availability depends on the current selection:

• General information — overview of the document: type, Own and Other Parties (with role tags), Important dates (with labels), Summary, and Risks identified from your own perspective. The section starts with the disclaimer “The AI can make mistakes, so please check the information for accuracy.”
• Problems — aggregated view of all problems found in the document.
• Selected block — appears when a card is selected in the centre column; see Reviewing a finding below.

Reviewing a finding

The Selected block tab is the working area for reviewing an individual finding. From top to bottom, it contains:

• Question posed — the question the AI asked itself when analysing the passage.
• Evaluation — the AI's detailed reasoning that led to the finding.
• Severity of the matter — a slider from low (teal) over medium (orange) to high (red). Reference marks on the slider show ratings already validated by you or other users in your account for comparable cases, providing context for the AI's classification.
• Proposed adjustment — a concrete revision text generated by the AI to resolve the issue.
• The disclaimer “The AI can make mistakes, so please check the information for accuracy.”
• Comment — an optional free-text input to explain your decision.
• Adopt or reject the AI's evaluation — three actions:
  • Thumbs up — accept the AI's assessment.
  • Thumbs down — reject the AI's assessment.
  • Skip (eye-with-slash icon) — leave the finding unrated for now.
  Each rated or skipped finding contributes to the review-progress bar in the top bar.
• Evaluation history — chronological list of previous assessments, comments and decisions made for this finding by you and your colleagues.

Sharing the results

The action icons in the top right of the evaluation view allow the result of a Smart Check to be shared:

• Send the results as an automatically generated e-mail.
• Export the results as a PDF report.
• Transfer the Smart Check to other users or teams within your account for further review.

Working with playbooks

A playbook (also referred to as checklist in some places of the user interface) is a structured set of questions used to evaluate a document against specific requirements — for example, internal contract standards or sector-specific compliance rules.

Structure

A playbook is organised into folders that group related questions by subject area (for example Term and Termination, Scope and Handling of Confidential Information, Liability, Disclaimers, and Indemnification). Each folder contains the individual checklist entries (for example 1.1 Reasonable duration?, 1.2 No contradictions in the specified contract periods?).

Per checklist entry, the AI can be configured to additionally produce a concrete improvement suggestion (the Proposed adaptation shown later in the evaluation view). The setting is toggled per entry when editing the checklist; the suggestion is only generated when the entry is evaluated negatively (unfulfilled).

Where playbooks come from

Depending on the tenant configuration, playbooks can be obtained in three ways:

• Sample playbooks provided with the platform for evaluation purposes.
• AI-generated playbooks created automatically from existing reference documents.
• Manually created playbooks, written by users or administrators.

A new playbook can be created either from the dedicated checklist management area (see Creating a checklist) or directly from step 2 of the Playbook Check wizard via Create Checklist.

Creating a checklist

Checklists are managed via the main menu under Settings → Checklists. The Checklists page lists every checklist available in your account with its name, group, creator, description and status. Each row offers actions to edit, delete or directly Create a playbook check from this checklist.

Two entry points at the top of the page create a new checklist:

• Generate with AI — opens the Checklist Generator. Provide a name, the language of the checklist and free-text Instructions describing the scope (for example Legal certainty of the employment contract). Optionally set a Minimum and Maximum number of questions to bound the result. The AI then generates a starting point that can be refined afterwards.
• Create manually — opens a dialog asking for the name, status, an optional Description (shown in the overview) and an optional Group (used to bundle related checklists). Save creates an empty checklist and opens the editor for adding entries.

The checklist editor

The editor opens after creating a checklist or by clicking a checklist's name in the overview. It shows the checklist's entries in a hierarchical list with drag-and-drop reordering. The top right of the page provides:

• Generate new questions — let the AI add further entries to the existing checklist. Same fields as the initial generator, plus an Exclude existing questions checkbox to avoid duplicates.
• Create entry — add a single entry manually.
• Edit checklist — change the checklist's metadata (name, status, description, group).
• Create Playbook Check — start a new Playbook Check directly with this checklist.

Entry types

Each entry has a type that determines how it is processed:

• Organizational — a folder used to group entries underneath, with its own progress indicator. The Group A / Group B / Group C structure visible in the evaluation view consists of organizational entries.
• LLM-question — the question is answered by the AI; this is the default for AI-generated entries.
• Simple question — the question is answered manually by the user during the check.
• Rule-question — the question is answered based on a previously defined rule.

Configuring an entry

The Create/edit entry dialog is organised in three tabs:

• General settings — entry type, the Question for AI (the actual prompt for LLM-questions), and the Answer behaviour. The expected answer format can be set to yes-or-no, whole number, decimal number, text or date. For yes-or-no answers, each possible value is mapped to either Accept (fulfilled) or Reject (unfulfilled), so the AI's answer can be classified correctly regardless of how the question is phrased.
• Advanced settings — short Name of the entry in keyword form, additional Instructions for the AI (limit values, test criteria), the Suggest improvement in case of negative evaluation toggle that controls whether a Proposed adaptation is generated when the AI's evaluation is negative, and external knowledge sources used for that adaptation (Clause Libraries, document vaults, MCP functions — see More features).
• other — additional Description and Further information for the auditing party (extra context shown to reviewers during the check), an Access to entry setting that controls who sees the entry (for example both parties or only one side), and free-text Tags for filtering.

Running a Playbook Check

The Playbook Check evaluates a document against a playbook (checklist) defined by you. It checks the document against the points defined in the playbook and helps surface potential gaps.

Creating a new check

Open the Playbook Check page from the home page or the main navigation. The page lists all existing Playbook Checks and provides a Create Playbook Check entry point. Activate it to start the three-step wizard.

Step 1 — File

Upload the document via drag-and-drop or Browse Computer. Only PDF files are supported. Confirm with Continue.

Step 2 — Playbook

Select the playbook to apply. The dialog lists all available playbooks together with their description. Use the search field to find a specific playbook quickly. The Create Checklist entry at the bottom of the list creates a new playbook directly from the wizard. Confirm with Continue.

Step 3 — Check details

The system pre-fills key metadata extracted from the uploaded document. Review and adjust as needed:

• Document — the uploaded file with an edit icon to replace it.
• Properties:
  • Name (required) — name of the playbook check.
  • Selected playbook (required) — the playbook chosen in step 2; can be changed via the dropdown or replaced with a newly created one.
  • Choose language (required) — pre-filled from your environment settings; can be overridden for this check.
  • Select law (required) — pre-filled from your environment settings; can be overridden for this check.
• Further options — collapsible area with additional configuration options.

Above the Start Check button, an indicator shows an estimate of the ContractCoins the analysis will consume with the current settings (for example ≈2 Coins required); the actual amount may vary depending on the document and the chosen options. Use Start Check to run the analysis or Back to return to step 2.

Processing and result tile

After Start Check, a new tile appears in the Playbook Check overview. The tile shows the document's designation, description, creation metadata and a status block:

• Initial processing — the AI is evaluating the document against the playbook; the duration depends on the document length and the size of the playbook.
• Ready — evaluation is complete. The tile then displays icons summarising the result:
  • Total number of checklist points evaluated.
  • Fulfilled points (green check) — requirements met.
  • Unfulfilled points (red cross) — requirements not met.
  • Uncertain points (purple) — could not be classified clearly.
  Open check opens the detailed evaluation view.

Evaluation view

The evaluation view is structured in three columns plus a top bar.

Top bar

• Document title with an info icon.
• Review progress bar (for example Not checked) — fills as you accept or reject the individual checklist points.
• Action icons in the top right for sharing the results — see Sharing the results below.

Document viewer (left)

The uploaded document is rendered with navigation and zoom controls. Passages used as evidence by the AI are highlighted in three colours:

• Red — referenced as a source for a negative (unfulfilled) evaluation.
• Green — referenced as a source for a positive (fulfilled) evaluation.
• Purple — referenced more than once or could not be clearly assigned to a positive or negative evaluation.

Checklist (centre)

The full playbook is shown grouped by folder. Each folder is collapsible and carries its own progress bar reflecting the share of entries you have already assessed. Each entry in a folder shows:

• A radio-style status circle on the left.
• The numbered question (for example 1.1 Reasonable duration?) with a help icon for additional context.
• An assessment indicator on the right whose colour reflects the AI's classification (fulfilled, unfulfilled or uncertain).

Selecting an entry loads its details in the right side panel and scrolls the document viewer to the relevant passages.

Entry details (right)

The right panel shows the playbook name and description at the top, followed by the detail view of the currently selected checklist entry:

• AI Response — the AI's answer to the question. The bot icon next to it changes with the assessment: a happy bot for a fulfilled point, a worried bot for an unfulfilled point, and a neutral bot when no clear classification was possible.
• Source(s) — the relevant text passages used for the evaluation. Each source contains the quoted excerpt, a clickable position reference (clause / page / text passage) that jumps to the corresponding passage in the document viewer, and an optional Hint with additional context from the AI.
• Proposed adaptation for this text — a concrete revision suggestion. This block is only generated for entries that the AI evaluates negatively (unfulfilled), and only if improvement suggestions are enabled for the entry in the underlying checklist. The setting can be toggled per entry when editing the checklist.
• Below the AI's evaluation, two actions allow you to record your own assessment:
  • Adopt result — accept the AI's classification for this entry.
  • Answer yourself — provide your own answer instead of the AI's.
  An optional Comment field can be used to justify the decision.
• Evaluation history — chronological list of previous assessments and comments for the entry; shows No reviews / comments found until the first decision is recorded.

Sharing the results

The action icons in the top right of the evaluation view allow the result of a Playbook Check to be shared in the same way as a Smart Check:

• Download the results as a PDF report.
• Transfer the check to other users or teams within your account for further processing.

More features

This Getting started guide covers the parts of Kontrakte.AI that most users encounter first: logging in, the home page, the Smart Check, working with playbooks and the Playbook Check. Several adjacent areas of the platform are referenced from these sections without being described in detail here. They are available depending on your tariff and configuration; we are happy to demonstrate them on request and will add dedicated documentation pages over time.

Areas not yet described in this guide include:

• Mass Check — automated review of many documents in a single run. Apply a Smart Check or Playbook Check to a whole batch of files at once and aggregate the results in one overview.
• Matter Management — central workspace for complex, long-running matters. Group related documents, checks and decisions, keep track of the current state and re-evaluate the matter as new documents are added.
• Document search — full-text and metadata search across documents stored in the platform. Helps locate clauses, parties or contract types without opening individual files.
• Document comparison — side-by-side comparison of two document versions with highlighted differences, so changes between drafts can be reviewed efficiently.
• Connections to external sources — integrations with external systems such as Document vaults, Clause Libraries and MCP functions (Model Context Protocol servers) that provide additional knowledge to the AI during analysis.
• Contracts — central management of contracts, including adding new ones and tracking their status. Available as a module on the home page.
• Document vaults — collections of documents that can be attached to a checklist entry as an additional knowledge source for the AI.
• Clause Libraries — collections of clauses referenced by the AI when generating an improvement suggestion for a negatively evaluated checklist entry.
• Other Settings areas — Campaigns, Clauses, Teams and User management, all reachable via the main menu under Settings.

If you would like a walk-through of any of these areas, please contact us.

Recommended next steps

For a structured introduction to the platform, the following sequence is recommended:

1. Complete the first-time setup described above.
2. Take the Smart Check tour in the Tutorial Hub.
3. Run a Smart Check on one of your own documents.
4. Take the Playbook tour.
5. Take the Checklist tour and create a checklist of your own.
6. Run a Playbook Check on a real document.
7. From this point on, continue with the feature that best matches your day-to-day work.

Further questions are collected on the Frequently asked questions page.