Solve Intelligence vs. DeepIP (2026): Pricing, features, and ease of use compared

If you're evaluating AI patent drafting tools in 2026, Solve Intelligence and DeepIP are probably both on your list. They're two of the most established options in the market, but they work very differently.

Solve Intelligence is a comprehensive, browser-based platform that covers drafting, prosecution, figure generation, and (as of December 2025) claim charting. DeepIP takes a leaner approach: it's a Microsoft Word add-in that enhances your existing workflow without asking you to learn a new system.

This guide compares Solve Intelligence vs. DeepIP across the factors that actually matter when choosing an AI patent drafting tool: pricing, ease of use, onboarding time, core features, and fit for different practice types. By the end, you'll have a clearer sense of which tool matches how you work.

Disclaimer: This guide is based on publicly available information, including company websites, published reviews, press coverage, and product documentation. Features and pricing may have changed since this comparison was written.

Solve Intelligence vs. DeepIP: The core difference

Before comparing features, it's worth understanding how each tool actually works from a design and philosophy standpoint.  

Solve Intelligence asks you to work inside their system. It's a browser-based copilot — you upload disclosures, draft in their interface, generate figures with their tools, and export to Word when you're finished. The pitch is everything in one place, fully integrated, with no context-switching.

DeepIP, on the other hand, involves downloading a Word add-in from the Microsoft Marketplace that’ll add a sidebar next to your document. Your drafting process stays the same; you just have an AI assistant riding shotgun. The pitch is enhancement without disruption.

This shapes everything downstream:

Solve Intelligence: An overview

Solve has raised $55M to build a platform that handles the entire patent lifecycle. Their December 2025 Series B — $40M led by investors including Microsoft's M12 and Thomson Reuters Ventures — funded expansion into claim charting and analytics, moving the product beyond pure drafting.

The founding team (Chris Parsonson, Sanj Ahilan, and Angus Parsonson) comes from machine learning research at UCL and institutions like The Alan Turing Institute. They built Solve as AI researchers entering legal tech, not legal tech veterans adding AI. 

What Solve offers:

• In-browser drafting with chat-based AI assistance
• Figure generation from text descriptions, sketches, or photos
• Office action response tools with MPEP citations
• Invention disclosure questionnaires and tracking
• Claims Charts (launched December 2025) for charting and analysis
• Jurisdiction templates for USPTO, EPO, and others
• Style training to match your drafting preferences

Explore our list of top Solve Intelligence alternatives. 

DeepIP: An overview

DeepIP raised $15M in March 2025 — smaller than Solve's raise, but backed by notable AI-focused investors including Resonance, Headline (Mistral AI's backer), and Balderton. According to their marketing, the company has helped draft over 10,000 patent applications since its launch.

The founders (François-Xavier Leduc and Edouard d'Archimbaud) spent years building AI for research institutions like Airbus and IBM before turning to patents. Their insight: attorneys don't want new platforms; they want their existing tools to work better.

What DeepIP offers:

• Word sidebar for drafting assistance (claims, specifications, abstracts)
• AI Reviewer for catching inconsistencies and unsupported claims
• Office action response drafting
• Basic figure generation from sketches or descriptions
• Style-matching that learns your firm's templates and tone
• Support for 10+ patent offices (USPTO, EPO, PCT, JPO, KIPO, etc.)
• Audio input — upload inventor interview recordings

Solve Intelligence vs. DeepIP features: What does each tool offer?

Drafting

Both tools use prompt-based generation. You describe what you need ("expand this embodiment," "rewrite claim 1 with broader scope"), and the AI produces text.

Solve wraps this in a custom editor with patent-specific UI elements. You can navigate between sections, manage claim dependencies, and see how changes propagate. The editor is designed for patents, but it's still an editor you have to learn.

DeepIP delivers the same capability without the custom environment. The AI responses appear in a panel; you drag them into your document or reject them. It’s less structured, but also less to learn.

Figure generation

Solve treats figures as a first-class feature. You can describe a diagram in text, upload a napkin sketch, or import a photo, and Solve converts it to patent-compliant line art. Editing tools let you add labels, adjust elements, and keep figures synchronized with your specification.

DeepIP offers figure generation, but it's more basic. Upload a sketch or flowchart description, get line art back, and export to Visio or PNG for refinement. It works, but reviewers note it lacks fine control. You probably still need a dedicated illustration service for complex figures.

If figures are a major bottleneck in your practice, Solve has the edge.

Office actions and prosecution

Solve has built extensive prosecution tools. The platform analyzes office actions, suggests amendments, and generates response frameworks with case law citations. The December 2025 Charts launch added claim charting and invalidity analysis, moving Solve into a territory that previously required separate tools.

DeepIP handles office action responses, but doesn't go as deep. You get AI-assisted arguments and amendment suggestions, but no integrated claim charting or invalidity workflows. 

Essentially, DeepIP is a drafting tool that helps with prosecution; Solve is trying to be a prosecution platform.

Quality control

Solve includes proofreading capabilities, but they're less emphasized. The platform focuses more on generation than review; you're expected to QC the output yourself or use external tools.

DeepIP has a dedicated AI Reviewer that scans for vague phrasing, unsupported claims, inconsistent terminology, and figure numbering errors. It's a proofreading layer built into the workflow.

Solve Intelligence vs. DeepIP pricing: How do costs compare?

DeepIP publishes straightforward pricing: $350 to 420/user/month, depending on billing frequency. A five-person team runs ~$21,000-25,000/year. No sales call required to know if it fits your budget.

Solve Intelligence doesn't publish pricing, but we can estimate. A NAPP member discount offers 15% off annual subscriptions, valued at ~$1,400, which suggests a full-price annual subscription around $9,300/user/year (~$775/month).

‍

The takeaway: Solve costs roughly 2x what DeepIP costs. For a five-person team, that's ~$25,000/year more. Whether it's worth it depends on whether you'll use Solve's broader feature set or whether you just need faster drafts.

Solve's pricing is estimated from a promotional offer and may vary based on team size or negotiated terms. DeepIP pricing is current as of early 2026. Verify with each vendor directly.

Solve Intelligence vs. DeepIP: Risks and known issues 

Solve Intelligence

• You have to actually use it: A platform only works if your team adopts it. If half your attorneys keep drafting in Word anyway, you're paying for a system that isn't being used. Solve requires organizational commitment, not just individual interest.
• Chat-based prompting interrupts flow: User reviews note that constantly asking "rewrite this" or "expand that" breaks drafting rhythm. You're switching between writing and prompting rather than staying in a single mode.
• Overkill for focused needs: If you just want faster first drafts and don't need figure generation, claim charting, or invention disclosure tools, Solve's breadth becomes clutter. You're navigating features you don't use to reach features you do.
• Software patents are the sweet spot: Solve performs best on software and computer-implemented inventions. Mechanical devices, chemistry, and biotech require more iteration and manual cleanup.

DeepIP

• It's just drafting: DeepIP won't help with claim charts, invalidity analysis, or portfolio management. If your work extends beyond preparing applications, you'll need additional tools.
• Figures are an afterthought: The figure generation exists, but isn't robust. For complex mechanical drawings or detailed system diagrams, you're still outsourcing to illustrators.
• Word dependency is absolute: If your team uses Google Docs, or if you're in an environment where Word isn't standard, the tool doesn't work.
• Premium pricing for limited scope: $350-420/month per user isn't cheap for a tool that only covers part of the patent workflow. You might pay similar rates for Solve and get more features (if you're willing to use them). 

Who should use Solve Intelligence vs. DeepIP?

Choose Solve if

• You're ready to standardize your practice on a new platform. 
• You want drafting, prosecution, and analysis in one system. 
• You have the time and organizational will to train your team properly. 
• You're drafting primarily software patents. 
• You have enterprise budget and don't mind the sales process.

Choose DeepIP if

• You want AI assistance without changing how you work. 
• Your team lives in Word and isn't moving. 
• Fast adoption matters more than feature depth. 
• You want to see pricing before scheduling demos. 
• You value zero data retention as an explicit policy. 
• You work across patent types and need broad jurisdiction support.

Choose neither if

You're frustrated by prompt-based AI in general, as both tools require learning how to ask for what you want, and output quality depends on your prompting skills. If you'd rather have AI that structures the invention before generating text, that's a different approach entirely.

Consider Patentext as an alternative to Solve Intelligence and DeepIP

Both Solve and DeepIP share an assumption: you prompt, the AI generates, you review. The quality of output depends on how well you phrase your request. Get the prompt wrong, get bad text back.

We built Patentext around the opposite idea.

When you upload a disclosure to Patentext, the AI doesn't immediately generate text. Instead, it creates a visual map of your invention, including components, relationships, steps, and logic flows. You can see exactly what the AI understood and edit the structure, add missing elements, and correct misinterpretations.

Only then does drafting begin. And because every section generates from the same underlying structure, terminology stays consistent, claim language aligns with the specification, and you can trace any paragraph back to its source.

Pricing is also straightforward: roughly $200 per patent application, with no seats, no monthly fees, no annual contracts. You can file one application this month, ten next month, and pay for what you use.

That said, Patentext isn't the right fit for everyone. If you want a comprehensive platform, Solve offers more features. If you want minimal disruption, DeepIP stays in Word. But if your frustration is specifically with prompt-based generation — with the unpredictability of asking AI to write and hoping it gets it right — Patentext solves a different problem.

We’re currently inviting firms and solo practitioners to start free, 4-week pilots. Schedule a demo to learn more.

‍

Frequently asked questions

Does Solve Intelligence or DeepIP offer a free trial?

DeepIP offers trials through its website. Solve requires scheduling a demo — trial availability depends on the conversation.

How does the pricing of Solve Intelligence and DeepIP compare?

DeepIP publishes pricing at $350-420/user/month. Solve doesn't publish directly, but based on a NAPP member discount offer, annual pricing appears to be around $9,300/user/year (~$775/month), which is roughly 2x DeepIP's cost. 

These estimates may not reflect current pricing; contact each vendor directly for accurate quotes.

Which is considered easiest to use: Solve Intelligence or DeepIP?

DeepIP. It's Word with a sidebar; no new platform to learn. Solve requires learning a new browser-based environment, which takes more time but offers more centralized control.

What integrations with other tools does Solve Intelligence offer?

Solve is primarily a self-contained platform. You can export drafts to Word, but the tool doesn't offer native Word integration like DeepIP. Solve focuses on keeping everything within its system rather than integrating with external tools.

Which handles more patent types?

DeepIP emphasizes technology-agnostic performance across software, mechanical, life sciences, and chemistry. Solve reportedly works best on software patents, with mechanical and biotech requiring more iteration.

How long does it take to fully onboard Solve Intelligence compared to DeepIP?

DeepIP can be productive on day one since it works inside Word. Solve typically requires more onboarding time — expect a few weeks for teams to get comfortable with the new platform and workflows.