Answers to the most common questions about what ArcPrime's AI can do, how it handles your data, and what to expect from AI-generated outputs.
Which features in ArcPrime use AI?
AI runs throughout ArcPrime. Prior art search, office action response recommendations, and final response drafts use it. So do the Filing Recommendation for disclosures, detectability predictions, and the Keep Score for pruning. Continuation recommendations, draft proofreading, claim charts, deep research, crucial elements, and the AI assistant round out the list. Each feature names what it produced so you know what to review.
What is the Filing Score?
Every submitted invention disclosure receives a Filing Recommendation built from five factors: Patentability, Detectability, Taxonomy Fit, Competitive Value, and Prosecution Outlook. Each factor shows its reasoning inline. Administrators set the weight of each factor and add custom factors under Settings > AI Recommendations > Filing. See understanding the Filing Score.
What are crucial elements?
Crucial elements are the points of novelty in a patent's claims, anchored to the exact claim language. ArcPrime suggests them, and you can confirm, edit, or add your own with Add crucial element. They carry across claim charts, family comparison, continuation recommendations, and office action responses, where The Core Novelty block uses them. See crucial elements.
What can't the AI do?
ArcPrime's AI does not make final decisions on your behalf. It does not file documents, submit responses, or take any irreversible action without your explicit approval. It recommends, suggests, and highlights. The final judgment always stays with you.
How accurate are AI outputs?
AI outputs are designed to be thorough and well-sourced, but they are not a substitute for professional judgment. Prior art results include citations so you can verify references directly. Office action recommendations include the reasoning behind each strategy. The AI assistant lists its sources under each answer. Always review AI outputs before finalizing any work product.
Can I use my own AI tools with ArcPrime data?
Yes. ArcPrime supports MCP (Model Context Protocol), so AI clients such as Claude, ChatGPT, Microsoft Copilot, and Gemini can connect to your workspace. Connected clients can read your portfolio and, with the right permissions, make changes such as updating patents, adding notes, or creating projects. Set it up under Settings > My Account > MCP. See MCP integration settings.
Does the AI use my data to train its models?
No. Your data is never used to train AI models. AI features process your data only to generate results within your workspace. See the Security and Privacy FAQ for more details on data handling.
How do I get the best results from the AI assistant?
Be specific in your questions. Include patent numbers, claim numbers, or technology areas when asking the AI assistant for analysis. The more context you provide, the more targeted and useful the output. See Best Practices for AI Features for detailed recommendations.
How long do AI analyses take?
The AI assistant and draft proofreading answer within seconds to a few minutes. Office action response generation takes longer and shows its progress on the page. A prior art search typically takes one to two hours, and deep research can take four to eight hours. ArcPrime emails you when a long-running analysis completes, so you do not need to keep the page open.
Still Have Questions?
Contact our support team at [email protected] or use the chat widget in the bottom-right corner of any ArcPrime page.
Related Articles
Best Practices for AI Features — Get more from ArcPrime's AI with better questions, review habits, and calibrated expectations
AI Assistant Overview — What the AI assistant can do on each page
Security and Privacy FAQ — Learn how ArcPrime protects your data and handles compliance
Troubleshooting: AI Analysis Not Completing — What to do if an AI analysis is stuck in a loading state

