What's new

Director, Manager, and Associate coordinate across every command and skill. They hand off between each other and only involve you for judgment calls. Kick off any conversation with "hi team, ..." — no commands or skills to remember.

Edit roles or add members like a Reviewer, Formatting Associate, or domain specialist. Personal or firm-wide scope.

First live artifact. Workspace-wide tracker showing every work product across all groups. Status, due date, group, item type. Filterable. Updates as work progresses. Configure which fields and views matter to you.

Summon the expert you need from anywhere in your workflow. Switzerland-specific transfer pricing rules, DEMPE analysis guidance, financial transactions nuance, pharmaceutical industry benchmarks — 136 experts covering countries, topics, transaction types, and industries. Ask for a review of your draft or a conversational advisory session.

Review mode produces a structured deliverable check — strengths, gaps, missing elements, specific fixes tied to section references. Conversation mode opens a freeform advisory dialogue where the expert answers follow-up questions and helps you think through positions.

Choose upfront how much the expert should dig. Quick uses embedded knowledge for a fast take. Standard reads the full knowledge base article for the jurisdiction or topic. Deep also searches official tax authority sources and recent publications for the latest developments.

Ask for a cross-jurisdictional review and the system coordinates multiple country experts simultaneously — Germany, France, Netherlands, UK, US all reviewing the same draft in parallel. Findings come back consolidated into one deliverable report.

Every local file and benchmark report now generates a Word document in the same folder as the PDF. Same content, same structure, editable in Word. Tables, charts, org structures, cover page, and table of contents all render natively in Word.

Configure how your Word deliverables look. Fonts, body and heading colors, margins, line spacing, logo placement on the cover, header and footer text with dynamic placeholders, cover layout, table style. Save one format or many. The agent uses the active format across every deliverable.

Create firm-wide formatting styles stored in your library, or import visual hints from an existing Word document you upload. Multiple styles coexist — advisory vs audit, internal vs client-facing — and the agent asks which one to apply when generating a deliverable.

Select your benchmarking standards, upload data in any format, review the proposed screening approach, build a seven-tab Excel workbook, and optionally generate a standalone benchmark report. Each step requires your confirmation before proceeding.

Three qualitative screening stages run strictly in order: database assessment from your data, web screening on official company websites, then detailed financials screening from downloaded annual reports. Each stage reduces the comparable set for the next. The search matrix traces every company from initial search to accept or reject.

The agent navigates company investor relations pages, downloads annual reports as PDF, and takes homepage screenshots of all accepted comparables. Every document is saved to the working papers folder for reviewer traceability.

Default standards ship with the plugin. Your firm can override quantitative thresholds, rejection criteria, PLI selection, adjustment formulas, range computation, and the benchmark report playbook. Place your standards file in the firm library and the agent detects and offers it automatically.

Seven tabs covering search strategy, quantitative screening, search matrix, business descriptions, PLI report, and range analysis. Every computed cell is a formula. Accounting and capital adjustments are fully documented. A reviewer can click any cell and trace the calculation back to the source data.

Ownership structures render as tree diagrams with percentage labels, country badges, and highlighted entities. Large structures automatically split into a simplified body chart and a full appendix version on a landscape page.

Intercompany flows render as directed graphs showing how goods, services, IP licences, and payments move between group entities. Solid arrows for goods and services, dashed arrows for payment flows. Works for both high-level value chain overviews and per-transaction diagrams.

Cross-entity DEMPE contributions displayed as structured tables grouped by Development, Enhancement, Maintenance, Protection, and Exploitation. Each entity rated on a seven-point scale from Sole/primary to None.

Market size trends, revenue splits, and competitive positioning render as line charts, bar charts, or bubble charts. Data sourced from your files and public information, with clear attribution.

Functional profiles are now written per profile type first (e.g. Toll Manufacturers, Limited-risk Distributors), then confirmed or adjusted per entity. Reduces repetition when multiple entities share the same characterisation.

Six core principles from the International Code of Ethics for Professional Accountants (2025 edition) guide every interaction. Includes the 2025 tax planning revisions covering credible basis, the stand-back test, and anti-avoidance awareness.

A pre-filled conduct file is created in your workspace at setup, covering output review, data handling, client-facing language, and risk flagging. Sensible defaults work out of the box. Customise to match your firm's policies.

The agent applies conduct principles at key decision points: method selection, range positioning, conclusions, and risk flagging. It will push back on unsupported positions, decline compliance sign-offs, and flag risks before proceeding.

Each playbook is a folder with a playbook.md definition and a verbatim/ directory for boilerplate content. Structure and prose live side by side.

Playbooks can extend other playbooks. Start from a firm-wide base, customize per jurisdiction or entity type. One change at the top cascades down.

Every playbook carries a version number. Track changes across reporting periods and ensure consistency across your team.

Personal, firm, group, entity, and deliverable. Each scope captures knowledge at the right level of abstraction for your TP workflow.

Memory files survive across sessions. Your firm conventions, entity-specific transfer pricing policies, and methodological choices are always at hand.

Running /setup and /prep-local-file creates the right memory folders for your scope. No manual file management required.

AI builds your transfer pricing documentation from local data. Files are saved on your machine or your cloud.

Start a workflow to prepare transfer pricing local files. The agent guides the transfer pricing professionals, learns, and follows your instructions to generate a complete first draft.

See exactly where every piece of content comes from. Color-coded origin tags and generation indicators give you full transparency over AI output.