Skip to content
By Role
By Vertical
pp wd
Product blog

Saige Assist Q&A: How We Use Beta Testing to Better Our CLM Platform

, , , , , | April 12, 2024 | By | 6 min read
Saige Assist Q&A: How We Use Beta Testing to Better our CLM Platform

At IntelAgree, we see innovation as a shared journey, and you — our users — are at the heart of it. This is why beta testing is vital to us; it lets us weave your insights and feedback directly into our development process, so we can deliver a CLM platform that exceeds your expectations.

Since our platform was meticulously crafted with AI and ML technology from the beginning, for us, Saige Assist is a natural evolution into the world of generative AI. However, user feedback is even more critical during this transition. We want to ensure Saige Assist doesn't just look impressive but genuinely streamlines your tasks and amplifies your productivity.

In light of this, we caught up with Michael Schacter, our Director of Product Management, to talk about how important beta testing is — especially for Saige Assist. He sheds light on how we choose our beta testers, the invaluable feedback they provide, and how this insight influences the direction of our product roadmap:


Q: How integral is beta testing to IntelAgree's overall product development process?

Michael: Beta testing isn't a new frontier for us; it's the cornerstone of how we've refined our products over time, though the Saige Assist beta testing phase might be our most visible and extensive yet. Part of our strategy is to constantly keep our fingers on the pulse of the market, listening closely to customer feedback and keeping an eye on what our competitors are up to. This insight helps us decide which features to prioritize. For high-impact features, we engage with diverse user groups, varying in size and composition, to ensure we capture a wide range of perspectives. This targeted approach allows us to refine these features effectively, based on real user feedback about what's hitting the mark and what needs fine-tuning. It's all about ensuring usability, simplifying training for teams, and fostering a partnership between us and our clients. By working closely with our customers during the beta phase, we can refine our offerings based on their direct feedback, ultimately benefiting all future users of our platform.


Q: How do you select participants for the Saige Assist beta testing phase?

Michael: Right from the start, our mission was to find users who would not just participate but do so with zeal and enthusiasm. Initially, we reached out to people we knew were interested from earlier conversations. However, our strategy quickly evolved; we really wanted to emphasize that this beta testing phase is a partnership. We're interested in understanding participants' journeys — how they managed before, during, and after integrating Saige Assist into their workflows. Selection involved a cross-functional effort from sales, customer success, and product teams, focusing on clients primed for significant transformation. Participants needed to be open to experimenting with new technology, embodying a "trust then verify" mentality. This approach ensured we partnered with early adopters ready to embrace generative AI with an open mind while understanding it's still evolving. 


Q: What features of Saige Assist are being tested in this beta phase?

Michael: We're putting several innovative components of Saige Assist through rigorous testing. Saige Assist: Search leverages natural language processing for more nuanced and efficient searches. Saige Assist: Contract Advice acts like ChatGPT for contracts, assisting users in navigating their agreements and answering pertinent questions. With Saige Assist: Redlining, we're automating redline suggestions during negotiations based on predefined playbooks. Lastly, Saige Assist: Editing aims to streamline the negotiation process by recommending precise revisions and edits. 


Q: How is user feedback being collected and evaluated during the beta testing phase?

Michael: We start with an initial kickoff call to demo Saige Assist, which usually prompts questions from the team. We then conduct internal training sessions involving key members from various departments — like legal and procurement — or anyone who will be using the tool extensively. We also provide our beta testers with a document to track their experiences. Every 7-10 days, we reconvene with these teams to discuss their experiences based on what they record in the document: what's working, what's not, and what additional use cases they'd like us to support. This structured approach helps us understand not only the specific needs of each customer, but also how these insights can benefit other clients or inspire new features. 


Q: What kind of support is available to beta testers if they encounter issues with Saige Assist?

Michael: Beta testers can book a meeting with me anytime. My schedule is always up-to-date because we're committed to their success. It's crucial for us to learn about any issues as quickly as possible so we can address them and adjust our roadmap if necessary. 


Q: How long is the beta testing phase planned to last, and what comes next in the development timeline?

Michael: The beta testing phase is set for three to four months. This period allows us to gather comprehensive feedback before a wider rollout. We're continuously bringing on more adopters to ensure that Saige Assist meets the diverse needs of our user base.


Q: Have you noticed any common trends or patterns in the feedback received from beta testers so far?

Michael:  Yes, we've observed some interesting trends, particularly with Saige Assist: Redlining. This feature is proving to be a game-changer in terms of efficiency, especially with the standardization and rules to ensure client-favorable outcomes. For instance, if there's a discrepancy in a background check clause, Saige Assist: Redlining can immediately highlight this difference or deviation from standard language, providing a clear starting point for negotiations. The feedback has been overwhelmingly positive, with users amazed by the efficiencies gained and the convenience of not having to sift through entire documents to find discrepancies.


Q: How many users are involved in the beta testing phase?

Michael: Currently, we have 10+ customers participating in the beta test, and that number is growing. We're adding more users in phases to gather a broad spectrum of feedback.


Q: Can you share some improvements beta testers have reported in their workflows thanks to Saige Assist?

Michael: With Saige Assist: Contract Advice, users can get straight to the information they need in contracts without having to wade through hundreds of pages. Whether it's identifying specific clauses, translating terms, or summarizing content, it simplifies the entire process. Similarly, Saige Assist: Search allows users to find contracts expiring soon or locate specific clauses like Force Majeure using plain language. It understands context; users can simply ask, "Show me all my contracts expiring soon," and Saige Assist understands "soon" as within the next 30 days. Then there's Saige Assist: Editor, which lets users rewrite clauses effectively using prompts. Lastly, Saige Assist: Editor allows users to easily modify clauses by providing prompts, akin to conversational AI tools. What sets it apart is not just the provision of answers but also the detailed explanation of the rationale behind these suggestions. This transparency helps users understand the reasoning, ensuring that adjustments can be made more effectively if needed.


Q: Is there a specific focus on how Saige Assist caters to users from different departments such as legal, procurement, and sales?

Michael: Absolutely. Saige Assist is designed to be versatile, accommodating the unique needs of various departments. For procurement teams, it ensures vendor agreements align with their standards and rules. Paralegals, who often handle initial reviews, find that Saige Assist makes verifying compliance with company standards much simpler. Sales teams benefit too, as Saige Assist highlights non-negotiable rules set by legal teams in vendor agreements, ensuring compliance before proceeding. This creates a standardized review process across the company, relieving the review burden and ensuring consistency in contract management.


Q: How are you ensuring data security and privacy during the beta testing phase?

Michael: Data security is our top priority. At IntelAgree, we use Microsoft Azure OpenAI (MS OAI) to power the generative AI features in Saige Assist. All the data processed by MS OAI is securely stored in Microsoft Azure datacenters in the U.S. geography, which comply with a broad range of standards including ISO, HIPAA, HITRUST, GDPR, FINRA, NIST, SOC, and CCPA. We've also prepared a comprehensive generative AI Security FAQ for those looking for more in-depth answers.


Q: Has the beta testing phase influenced any future development plans for Saige Assist?

Michael: The feedback we're receiving is instrumental in shaping our roadmap. By validating this feedback with other beta customers and monitoring market trends, we’re not only addressing current use cases but also expanding Saige Assist's capabilities. This includes refining rules based on organizational needs, enhancing prompt glossaries, developing fallback options, and improving attribute recognition. We're also working on incorporating markup values in responses and streamlining navigation within documents. Additionally, we're developing Saige Assist: Help for internal documentation and guidance throughout the application.


Q: What is the process for beta testers to provide feedback on Saige Assist?

Michael: We've provided our beta testers with a document to track their experiences. During our 7-10 day review meetings, we go through this document together. This structured approach helps us understand not only the specific needs of each customer, but also how these insights can benefit other clients or inspire new features.


Q: How is the performance of Saige Assist being measured during the beta testing phase?

Michael: We're closely monitoring various metrics, such as rule usage, search frequency, and user engagement. Through questionnaires and surveys, we're comparing the efficiency of tasks before and after implementing Saige Assist. This helps us understand the tangible benefits it brings to drafting, redlining, and searching processes.


Q: How is the beta testing phase helping to prepare Saige Assist for a wider release?

Michael: The beta phase is crucial for identifying bugs, understanding user satisfaction, and confirming the tool's effectiveness and efficiency. It's about ensuring that Saige Assist not only meets but exceeds our users' needs before we consider a wider release. Our goal is continuous improvement, informed by real user feedback, to prioritize future developments effectively.


Q: How do you plan to continue to incorporate user feedback even after the launch?

Michael: Post-launch, we'll continue our practice of encouraging users to submit feature requests. From there, we'll assess, prioritize, and look for patterns among our customer base. We also keep an eye on market trends to ensure our roadmap aligns with broader industry movements. Essentially, all user feedback will influence our development roadmap. This ongoing dialogue with our users ensures that Saige Assist remains dynamic, responsive, and aligned with evolving needs.


Subscribe and Stay Informed

This partnership in innovation is a shared path toward making technology that genuinely benefits its users. As we continue to refine and expand our solutions based on your invaluable input, stay connected with us. Subscribe to our blog for all the latest updates on Saige Assist and to keep your finger on the pulse of where your contributions are leading the future of tech.