PreviousTo Changelog

Mar 2026 Update

πŸ“¦πŸš€ Here’s what we shipped in March at ArcPrime! πŸŽ‰πŸ› οΈ

βœ… MCP for External Agents

β€’ You can now connect external agents to ArcPrime through an OAuth-based MCP (Model Context Protocol) server, making it much easier to securely work with ArcPrime data and tools from your broader AI stack.

β€’ The MCP server exposes tools across portfolios, disclosures, projects, dashboards, and analytics, so your agents can actually do useful work instead of just reading static exports.
β€’ This means you can plug ArcPrime directly into your workflows for things like pruning, categorization, filings, spend analysis, and claim charts β€” all powered by agents that can reason over your portfolio instead of operating in silos.

πŸ’‘ Smarter, More Proactive Disclosure Intake

β€’ Disclosure intake now catches missing required fields automatically, helping you clean up incomplete submissions before they become follow-up work for legal or program teams.

β€’ ArcPrime can now notify inventors automatically when something important is missing, so intake keeps moving without manual chasing.

β€’ We also improved settings and templates around disclosure notifications, giving you more control over how these workflows run.

πŸ“š Deep Research Now Has a Real Document Library

β€’ You can now upload your own documents and organize them in folders, making Deep Research much more usable for real, ongoing work.

β€’ Run claim matching against your library content so you can compare ideas against the material that actually matters to your team.

β€’ You can also archive evidence sources and generate concept charts directly from Deep Research.

β€’ This is especially useful for SEP workflows, where source documents are long, fixed, and authoritative. Deep Research lets you map claims directly to those specific texts instead of pulling from scattered web sources, making analysis far more precise and reliable.

Stronger Patents.
Better Coverage.
Lower Costs.

See how ArcPrime helps IP leaders continuously optimize portfolio performance with domain-specific AI.