As enterprises grapple with the security risks of cloud-based AI, a new approach is gaining traction: processing sensitive data directly on-device. Leading this charge in the legal tech space is SpotDraft, a company whose on-device contract AI has captured significant investor attention. We sit down with Vijay Raina, a leading expert in enterprise SaaS and software architecture, to unpack SpotDraft’s recent strategic funding from Qualcomm Ventures, which nearly doubled its valuation toward $400 million. We’ll explore how they’ve managed to bring powerful AI capabilities offline, the technical hurdles they overcame, and what this privacy-first model means for the future of AI in highly regulated industries.
Your company’s valuation recently doubled toward $400 million in a year. What key customer growth metrics, such as the 173% increase in contract volume, drove this new valuation, and how will this strategic investment from Qualcomm Ventures help you sustain that momentum? Please share some details.
The jump in valuation to around $380 million is a direct reflection of our explosive growth and the market’s validation of our on-device AI strategy. It wasn’t just one metric, but a confluence of powerful indicators. We saw our contract volumes surge by 173% year-over-year, and our platform is now processing over a million contracts annually for more than 700 customers. This isn’t just passive use; we have nearly 50,000 active users every month. This deep engagement proves the demand. The investment from Qualcomm Ventures is far more than just capital; it’s a strategic partnership. It allows us to deepen our AI capabilities and accelerate our expansion across the Americas and EMEA through joint go-to-market efforts, ensuring we’re perfectly positioned as the next wave of AI-ready hardware hits the market.
Many enterprises in regulated sectors like pharma and defense are slow to adopt cloud AI due to security concerns. How does your on-device VerifAI workflow directly address these data privacy and residency requirements? Could you describe the biggest technical challenge you overcame to make this offline functionality a reality?
This is precisely the problem we were built to solve. For sectors like pharma and defense, sending a contract filled with intellectual property, privileged information, and critical deal terms to an external cloud server is a non-starter. Our VerifAI workflow completely sidesteps this issue by performing the core intelligence tasks—reviewing the contract, scoring risks, and suggesting redlines—entirely on the user’s local machine. The document never leaves the device for these critical functions. The biggest technical hurdle was a battle against compromise. We had to shrink our powerful, proprietary AI models to run efficiently on a laptop’s processor without sacrificing the quality of the legal analysis. It took immense fine-tuning to achieve this balance, ensuring the offline experience was not just secure but also fast and effective enough to rival cloud-based alternatives.
With on-device AI now showing as little as a 5% performance difference from frontier cloud models, what specific fine-tuning was required to achieve this? Please walk us through how you balance the trade-offs between on-device speed and the capabilities of larger, cloud-based systems for legal teams.
Closing that performance gap to just a 5% difference from leading cloud models was a monumental effort in optimization. It wasn’t about simply shrinking a large model; it was about rebuilding a highly specialized, fine-tuned model specifically for the architecture of platforms like the Snapdragon X Elite. We focused intently on the core tasks a lawyer needs—comparing clauses against playbooks and identifying risks—and stripped away extraneous computational weight. This allows us to achieve speeds that are now about one-third of what you’d get from a cloud API call. The trade-off is managed through a hybrid approach. The most sensitive, latency-critical work happens on-device. For less sensitive, collaborative features or initial user authentication, we still leverage the cloud. This gives legal teams the robust, secure, and immediate intelligence they need for deep work, without sacrificing the collaborative tools they also rely on.
VerifAI integrates into Microsoft Word, applying playbooks directly where legal teams already work. Could you share a specific example of how this in-workflow, on-device support has transformed a customer’s contract review process? What feedback have you received from your nearly 50,000 monthly users?
The in-workflow support is transformative because it eliminates friction. Imagine a lawyer at a pharmaceutical company receiving a complex third-party agreement. Previously, they would have to manually cross-reference it against a 100-page internal policy document. Now, they open it in Microsoft Word, and VerifAI instantly gets to work, offline. It flags a non-compliant liability clause, suggests company-approved language, and explains why the change is necessary, all within seconds and without the document ever being uploaded to a server. The feedback from our users has been overwhelmingly positive. They love that the AI meets them where they already are, in the tool they use all day. They constantly tell us how this seamless, secure assistance not only saves them hours of painstaking work but also gives them the confidence that they are not missing critical risks buried deep in the legal text.
Your partnership with Qualcomm includes joint go-to-market efforts. What are the first practical steps you will take together to expand your enterprise presence in the Americas and EMEA? Please explain how you plan to navigate the market as compatible AI PC hardware becomes more widely available.
Our partnership with Qualcomm is a two-pronged attack. The first practical step is joint development, ensuring our software is deeply optimized for their Snapdragon platforms to deliver an unparalleled on-device experience. The second, and equally important, step is a coordinated go-to-market strategy. We’ll be working directly with their enterprise sales teams to target regulated industries in the Americas and EMEA that are actively seeking secure AI solutions. As AI PCs become the new standard, our strategy is to be the killer app for legal and compliance. We won’t be trying to catch up to the hardware; we’ll be riding the wave with it. By demonstrating a powerful, privacy-critical workflow that runs flawlessly on this new generation of devices, we become an essential part of the enterprise AI PC value proposition.
What is your forecast for on-device enterprise AI over the next three to five years, particularly within legal and other highly regulated industries?
The future of enterprise AI is not in one central, all-powerful cloud; it’s hybrid and distributed. Over the next three to five years, I predict that on-device AI will become the default standard for any workflow involving sensitive or regulated data. The current debate around data privacy and security isn’t going away; it’s intensifying. For industries like legal, finance, and defense, the ability to say with 100% certainty that your critical data never left your control will shift from a competitive advantage to a baseline requirement. We will see a major market movement toward architectures where the AI is brought directly to the data, not the other way around. This “close to the document” approach will unlock AI adoption for the most critical, high-value enterprise use cases that have been held back by security concerns.
