AiThority Interview with Jonathan Rhyne, Cofounder & CEO of Nutrient

Jonathan Rhyne, Cofounder & CEO of Nutrient catches up with AiThority.com to chat about the latest in AI, which AI SaaS enhancements are taking the industry by storm, insights on the current document processing landscape and more:

———

Take us through your journey in the AI realm and tell us about your role at Nutrient.

I originally became interested in Machine Learning approaches to data extraction to solve the problem of data being locked up in image-based PDFs that were originally made as scans or pictures of paper documents. When GenAI emerged, and with the advent of NLP, it really started to pique my interest. I personally have played around a ton with Stable Diffusion as well as Midjourney as a creative outlet and it ultimately helped me understand better the importance of prompts, training models, as well as the use of LoRAs.

Obviously, what started as a hobby interest has grown, as AI has advanced, to be an intricate part of my daily work, and thought space, and ultimately impacts the strategy of the company. In my role as Co-founder and CEO of Nutrient, my entry into AI has always been about solving tangible problems for developers, businesses, and citizen-developers by unlocking the untapped potential of their data and document workflows.

Nutrient’s mission is to evolve how we humans interact and experience documents. Over the past decade, I’ve guided our evolution from a developer-focused document SDK provider to a comprehensive platform integrating AI to revolutionize the entire document lifecycle. My role is to set the vision and ensure our products align with that vision. AI is one of the key tenets in that mission, as I think it brings unprecedented productivity, insights, and efficiency gains to enterprises that are already changing how we interact with our documents.

Also Read: Predictive GenAI: Redefining ROI in the AI Revolution

We’d love the highlights on your latest AI enhancements and how it benefit end users of your products?

At Nutrient, our recent AI enhancements focus on unlocking document intelligence, automation, and ultimately a new way to interact with a document. Here are a few key highlights:
AI Document Assistant: We’ve recently shipped this as part of our Web SDK, as well as through our end-user application. Our AI Assistant allows users to interact with their documents conversationally and can be built together with OpenAI’s real-time API to talk with your voice to your document. Whether it’s extracting key data points, summarizing lengthy contracts, analyzing trends in a report, or translating a document, it empowers users to get actionable insights and accomplish work quicker.

• AI Assistant can also assist in AI-powered redaction and Anonymization: Sensitive data management is a growing challenge, and our tools use AI to identify and redact personal or confidential information at scale by understanding the context around a request—essential for industries like legal, healthcare, and finance.
• Intelligent OCR and Data Extraction: Beyond recognizing text, our AI can identify context by what you ask for. For instance, it doesn’t just extract numbers from invoices; it understands the relationships between them, such as which values correspond to taxes, subtotals, or discounts and much more.

These are just some of the first enhancements we’ve shipped that benefit users by reducing time spent on mundane tasks, improving accuracy, and delivering insights. We’re focused on much more game-changing functionality currently, especially in our workflow automation platform as well as baked-in to common document-centric use cases.

Can you share insights on the document processing landscape and how it’s been evolving over recent years?

The document processing landscape has dramatically changed in the past few years between COVID, the rise of web & mobile-based applications, and now AI. Nutrient has its roots in the mobile space. The current transition away from traditional desktop software to web and mobile applications has meant the decline of the file browser. We have a new generation that has grown up without needing to check file extensions to determine which app to use for opening them. The biggest change has been in users’ expectations around software working efficiently. These expectations, along with the death of printing, mean that documents are being used way differently than before.

As a result, users often don’t grasp why different file types are required for different purposes.COVID accelerated many long-term digital transformation initiatives that businesses had planned for the next five years, bringing them to the forefront. Remote and hybrid work significantly accelerated the adoption of digital workflows, including digital signing. While digital signing technology itself didn’t advance dramatically during this period, the necessity for businesses and governments to adopt it led to widespread acceptance. Now that it’s in place, no one wants to revert to the old ways.AI is just beginning to transform many traditional work processes. Since much of our work still happens within the confines of digital screens, this transformation is also affecting how we handle documents. Initially, we saw AI’s impact on content generation, but now we’re witnessing innovative ways to interact with, understand, and communicate through these documents. This is a groundbreaking innovation, and I believe it will be a massive accelerator in human productivity.

Also Read: From Aspirin to AI: Revolutionizing the Drug Approval Process

What features do modern workplace executives need most from their document processing tools today?

I believe not just modern executives, but also modern digital employees are navigating a world full of tight deadlines, larger workloads, dispersed teams, and increasing regulatory and security scrutiny. Here’s what they need most from document processing tools:

• Real-Time Collaboration: Teams are often global, so tools that enable seamless, real-time co-authoring, co-marking up, editing, and commenting on documents are indispensable in today’s workplace.
• AI-Driven Insights: Executives especially need more than just access to data—they need actionable insights. Whether it’s analyzing customer contracts for renewal trends or spotting discrepancies in financial reports, insights need to be delivered proactively and be right at the person’s fingertips inside of the document(s) they are reviewing, building, or approving.
• Automation and Workflow Optimization: Executives value tools that can reduce bottlenecks and manual intervention. If it saves them costs, and time, or makes their team more productive then these show up in the top and bottom lines. Automated approvals, intelligent notifications, and dynamic workflows quickly built through the help of AI ensure projects and tasks move forward without unnecessary delays along with a robust audit trail automatically made and readily available without more work.
• Security and Privacy: With sensitive documents like contracts and financial reports, tools need robust access controls, encryption, and automated redaction capabilities to ensure data is secure and privacy is guaranteed. As more documents become digital, these concerns become essential from the beginning.
• Integration Capabilities, Flexibility, and Customization: Different organizations are at different stages in their digital transformation journey and have different needs. Some executives need tools that plug into their existing CRM, ERP, or BI platforms, eliminating the cost of migration and years of investment in current platforms. Other executives want a unified platform that can handle all their document-related needs and don’t want to have separate vendors for CRM, ERP, BPO, RPA, and Workflow. Others have complex requirements and require easily integrated developer toolkits for their internal or external development teams to build out customer functionality.

A few thoughts on how end users can optimize the use of AI in using documents and extracting insights and data from it on time when they need it the most.

We’re in the beginning stages of the application layer or phase 2 of a technology trend in AI. In the future, I expect a plethora (more than people can consume) of tools promising everything under the sun with AI. The reality is that many of these solutions won’t necessarily improve or resolve current issues, particularly with documents. This is partly because, although PDFs and Word Docs are widespread, they are not all created equally. They involve complex standards that few people in today’s tech industry fully understand. Additionally, what is visible in these documents often doesn’t represent all the content they contain.

With that caveat, I would think for a majority of use cases, AI can be used to:

• Leverage Context-Aware Tools: Use AI capable of understanding the context of your documents. For instance, extracting “amount due” from invoices or “renewal dates” from contracts requires tools that understand semantics, not just keywords.
• Automate Repetitive Tasks: Invest time upfront to create AI-powered workflows for repetitive tasks like approvals, data entry, or compliance checks. The efficiency gains compound quickly.
• Finetune Training of LLMs on Relevant Data: AI is only as good as the data it learns from. Ensure your tools are exposed to your specific document types to improve performance and accuracy over time.
• Integrate AI into Collaboration: Use tools that enable collaborative AI features, such as generating summaries or pulling insights directly into a team chat or project management tool from conversations, video call transcripts, or relevant data sets.

One last reminder is to stay informed about AI ethics and the laws surrounding its use. As AI becomes more integral, it’s important to be aware of its limitations and ethical considerations, ensuring that it complements human judgment rather than fully replacing it.

Five of the latest AI innovations from around the world that have piqued your interest in the recent past and why?

The use of generative AI in healthcare is truly fascinating. Tools like DeepMind’s AlphaFold are revolutionizing drug discovery by predicting protein folding with remarkable accuracy. Additionally, the application of GPT models has empowered doctors to quickly access and analyze the latest medical research and large datasets, thereby enhancing their ability to diagnose diseases more accurately.

Further, the evolution of code generation, especially in low-code tooling like replit.com and make.com is used alongside popular LLMs to increase the accessibility of who can develop and deploy web and python apps via building AI Agents. The spending it takes to go from idea to POC is mind-blowing.

Building on this, I’m also amazed at how developer tools, specifically IDEs like Visual Code and Cursor are integrating LLMs to increase the productivity of bug fixing, script writing, and even building basic web applications from scratch. Collaborative tools like Google Colab are incredible because they facilitate seamless collaboration between non-technical users and developers, as well as between two engineers.

Image Diffusion models and incoming Video Diffusion models are revolutionary, especially now that users can easily train them in a certain style. They continue to improve and are becoming more difficult to tell the difference between something AI generated and not. While I do think there are trademark and ethical issues here, I also think the ability of designers, marketers, and creators to utilize these tools are amazing.

Lastly, I’m impressed by the improvement in LLMs and common GPTs over the past year and their ability to maintain larger amounts of context and augment it by scanning the internet or documents that you give it. What excites me the most here, though, is the coming advances in SLMs and what that means for personal, privacy-focused, and task-specific AI. If LLMs continue to exponentially grow but hit scaling issues, then the next advancement will be training and fine-tuning SLMs.

[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]