Juro, 2024
From manual grind to AI-powered efficiency
Challenge
Sales and legal teams were spending 15–30 minutes per contract on low-impact manual tagging. Over five months and six releases, we introduced AI Extract, a one-click, human-in-the-loop data extraction experience.
My role
I led end-to-end design for AI Extract, including cross-team collaboration, preliminary and post-launch research, data analysis, design for initial and following iterations and development handover.
Team
Product trio: Staff Product Designer (Myself); Product Manager; Tech Lead;
Engineers: QA; Back-end x2; Front-end x2
Timeline and milestones
6 releases during 5 months (2024)
Impact
50% AI functions adoption by monthly active users
86% satisfaction score earned from customers
8% ARR lift driven directly by AI Extract
15x cut in document tagging time (15 to 1 min)
Juro is an all-in-one platform where companies create, negotiate, and sign documents. It's like G-Docs meets DocuSign, but only built specifically for contracts. When documents come into the storage, customers tag them to optimize filtering and search, automate approvals and renewals. In short, it helps them to stay in control of their vendor relationships without chasing the documents in folders or emails.
We confirmed the scale of the pain through interviews, support data, and analytics.
UX methods
16 user interviews, ～100 customers' feedback analyzed, lost prospects and churn reasons analytics, in-product data
Key insight
Third-party paper volumes are ～50% of all legal contracts. In the Juro repository, the portion of external contracts is only ～20% due to poor support for external documents. Only 15% of uploaded documents are tagged, making the remaining invisible legacy.
User sentiments
Juro is great for our own paper workflow. But we need a clean way to upload vendors’ documents, pass them through the contracting process, track changes, and send them back to the counterparty as a Word file without breaking the formatting. Additionally, having to manually tag these contracts is a waste of time.
Sr Corporate Counsel, mid-size business
Our first prototype rewrote the entire intake flow, but data and stakeholder signals led us to pivot.
Tagging pain spans new and legacy contracts
AI extraction is great! It’s so much clearer and cuts the time our team spends re-uploading documents. Can it handle existing contracts?
Legal Ops, enterprise
Executive mandate for quick wins led us to prioritize AI tagging over a full pipeline rebuild
What will have the highest impact and push us toward an AI-native future? Let's focus on this for 1-2 sprints.
Richard, CEO
Cut manual tagging time by 10×
Legal and ops teams need to extract key contract details — like dates, amounts, or notice periods — without reading the whole document. AI-powered extraction can save up to 15 minutes per document while let legal teams stay in control of extracted data.
New experience insight
I use AI for routine fields and confirm the results manually. It makes repetitive tasks easier and faster, while keeping me in control. It would be helpful if you could click a smartfield and jump to where the AI found it in the document, so you can check instantly
Senior Legal Ops
Reduced verification time by ~1 minute per field
Users want to verify what the AI has pulled out without having to chase the data position in the document. Seeing extracted data in the context it came from helps to build trust with AI and stay in control while keeping verification quick.
Accuracy delta
Improved accuracy from 50% → 85% for complex data
Companies care about different terms — one size doesn’t fit all in the legal world. The playbook interface with the ability to fine-tune what AI needs to find or explain required conditions improves the accuracy for complex fields from 50% to 85%, making it a drastically different product.
Scaling opportunity
300–1500 legacy documents:
An average number of files that new customers need to tag during onboarding. Enabling bulk extraction with AI unblocks efficient and quick customer activation.
Made mass extraction scalable across contract archives
Ops teams often deal with hundreds of legacy contracts stuck in shared drives or inboxes. I enabled bulk AI extraction for our CS team to pull data for clients from the entire base in minutes, not hours and build reporting and workflows on top of it.
It saves me around 30 seconds per field, Across contracts, that adds up fast and significantly reduces manual work. But it's more than that. Going through a contract to fill fields with data isn't particularly interesting work. So this is just a better way to do the job.
Legal Counsel, Network Security
I upload a contract and click the button. It’s integrated into our CLM, so I don’t even think about it—it’s just there. Makes it easy and efficient.
Head of Legal Ops, enterprise
We’re planning a project to bulk upload all our legacy contracts. This feature will be really useful—it can pull the same smartfields for reporting and help us build custom table views in Juro.
Legal Operations Lead, mid-size business
Highlighting parts of the document where the extraction has been taken from is the best thing ever.
Legal Counsel, SaaS
So far it has been a great feature! We even have some contracts in German that we've had to upload and it has been helpful in getting those tagged correctly.
Head of Legal, SaaS
Impact across key metrics
Retention: ~40–45% weekly retention aligns with document upload and indicates co-dependent feature usage.
Adoption: ~40% of customers used AI Extract ≥3 times in Month 1, trending toward ~60% by Month 3.
Accuracy: 84% of AI suggestions accepted by beta customers. Ranges 70–90% by company.
Impact on the North star: 60% lift in uploads vs. the pre-launch period; some accounts experienced 3–4× growth.
What surprised me: The human-in-the-loop highlights were more valued than pure speed. Users wanted control as much as automation.
Why it mattered: By marrying generative AI with clear user controls, we reduced busywork and drove measurable business impact, paving the way for further AI-first automations at Juro.
Next priority: We found the biggest lift in legacy/bulk uploads. Scaling this allows customers to have more control over their data.