The Release Report: July 2026
July brings Vincent Skills, Memory, visual document review, and a sharper GPT-5.6 model, plus a new claimant PI framework in Clio Operate.
A matter doesn’t stay with one attorney for its whole life: it gets picked up, set down, and handed off, and something gets lost each time. This month, Vincent starts closing that gap.
A firm’s drafting standards now travel with every attorney, and case history carries forward across sessions instead of resetting. Vincent also reads more of the file, tables, photos, handwriting, and redlines. The research gets sharper with the move to GPT-5.6.
Claimant personal injury teams get a working England and Wales framework in Clio Operate that covers the full case lifecycle. Read more about what’s new.
Vincent Skills: your standards, applied to every matter
Your firm can set its preferred drafting structure, analytical framework, or house voice once, and Vincent applies it the same way across every matter.
In every draft, it follows how your law firm works, including the format and organization, the reasoning approach it applies across matters, or the tone and style. Once set, Vincent applies your firm’s standards consistently across outputs, for every matter and every attorney.
Read the full Vincent Skills announcement
Memory: matter history that doesn’t reset
Vincent now carries persistent awareness across sessions, so key dates, prior advice, and open questions are still there when you return, even after weeks, without having to be repeated or reconstructed.
Nothing gets lost between sessions. Memory removes the need to repeat details, prior advice, and strategic decisions every time you return.
Visual document review: Vincent reads what’s on the page
Legal documents carry more than typed text. Tables, diagrams, photos, and handwriting appear throughout a file, often on scanned pages. Vincent now reads all of it, leaving no layer behind.
Vincent decides when to look closer, opening an image or section when the prompt points to it, and surfaces the relevant exhibit automatically when the answer depends on a visual element. Attorneys can also ask Vincent to check a specific table, diagram, or exhibit directly, especially useful when visual accuracy has to be confirmed.
Vincent reads tracked changes and comments the same way, picking up both the document’s language and what’s still being proposed around it. Ask it to “summarize the proposed changes” or “list the open comments,” and it walks through what changed round after round and flags which comments are still open.
Sharper research, powered by GPT-5.6
We upgraded Vincent’s underlying model to GPT-5.6 and validated it against our legal research evaluation suite.
Compared with GPT-5.5, it delivers 6% higher completed-answer quality based on expert-defined research criteria, and surfaces 6% more of the expected authorities during the research process, for a more complete first pass.
Better answers and broader research mean lawyers start stronger, before they’ve even opened the first document.
A running start on claimant PI
Standing up a practice area like PI on a new platform normally takes months of configuration. Our new Clio Operate solution accelerator gives you a working claimant PI model on day one, then your team can shape it to how you actually work, so you get to value far faster and with less risk.
The accelerator includes a pre-built configuration pack of content, covering the claimant personal injury case lifecycle in England and Wales, from matter setup through settlement and closure, complete with automated workflows,phase plans, key dates, data-capture forms, a risk scorecard, and heads of loss ready to go.
Start from a working framework and shape it to your firm in Clio Operate.
That’s the July release. The goal stays the same each month: less friction, more fit for how firms already work. See you next month for our September update.
