In a world where businesses generate mountains of documentation but still rely heavily on manual processes, Dodocs.ai stands at the intersection of AI innovation and real-world operational efficiency. At its core is Dan Gudkov, an entrepreneur with a sharp eye for inefficiencies and a passion for building practical, scalable solutions. With experience ranging from tech startups to running a restaurant, Dan has seen firsthand how businesses—large and small—struggle with document workflows. This insight became the seed for Dodocs.ai, an AI-driven platform that automates and intelligently processes enterprise documents using a unique multi-microservice architecture and multiple open-source LLMs.
Dodocs.ai is an enterprise-grade SaaS platform focused on AI-powered document automation. With a foundation built on microservice architecture and open-source large language models (LLMs), Dodocs.ai transforms how companies manage, extract and utilize data from documents across diverse industries. From invoice processing to ERP integrations, the platform provides a robust, scalable and intelligent solution to reduce manual effort, increase accuracy and create value in operations where documentation plays a crucial role. Dodocs.ai is at the forefront of what its founder calls “Tech Revolution 5.0” where AI becomes the cognitive layer behind enterprise workflows.
TFS: Hi Dan, thank you so much for taking the time to speak with us today. It’s great to have you here!
Dan Gudkov: Thank you! I’m really glad to be part of this conversation. It’s always a pleasure to share what we’re building with Dodocs.ai and to talk about the real impact AI can have when it’s implemented thoughtfully.
TFS: Absolutely—and there’s so much to dive into. Let’s start from the beginning. What inspired you to create Dodocs.ai and what gap in the market were you aiming to fill?
Dan Gudkov: The inspiration for Dodocs.ai came from years of grappling with documentation inefficiencies in multiple industries. I’ve been involved in both tech-heavy environments and traditional businesses like the restaurant I used to run. Across these varied experiences, I consistently noticed that teams were overwhelmed by documentation tasks—whether it was procurement paperwork, inventory sheets or financial reports. These were manual, repetitive and error-prone processes that drained time and resources.
I realized that while data was being generated and stored, it was rarely used effectively because the access layer—the document workflows—was broken or archaic. That’s where Dodocs.ai was born. I wanted to build a tool that could not only automate but intelligently manage documents by using AI to interpret, extract and process information seamlessly across industries. The goal wasn’t just efficiency—it was about empowering companies to truly harness the value hidden inside their documents.
TFS: The concept of AI-driven document automation has been explored by various platforms. What makesDodocs.ai truly unique?
Dan Gudkov: You’re right—document automation isn’t a new concept. What makes Dodocs.ai stand out is our architectural philosophy and our hyper-focus on real-world applicability. We designed Dodocs.ai using a multi-microservice infrastructure that allows us to deploy different open-source LLMs for different types of tasks. This enables us to be much more flexible and scalable than traditional monolithic systems.
Moreover, we’ve zeroed in on operations that are universal across verticals—like invoice processing, logistics documentation and POS-related data flows. These are processes that occur in almost every business, regardless of the industry and we’ve designed our tools to be exceptional at automating those workflows with minimal customization. That gives us a strategic advantage—we’re not trying to be everything to everyone. We aim to dominate in very specific, high-volume use cases and be the best at it.
TFS: How do you see the role of AI evolving in legal and enterprise documentation over the next five years?
Dan Gudkov: In the next five years, I believe AI will become an inseparable layer of enterprise software, especially in documentation-heavy departments like legal, finance and compliance. Most of the friction in document workflows comes from manual interpretation and data entry—two areas where AI is already showing incredible results.
At Dodocs.ai, we’re focusing on what we call the “grey zones”—the spaces between existing automation systems where documents still need human intervention. For instance, a manager might spend hours reading contracts or transferring invoice data from PDFs into ERP systems. Our system already reduces that burden by up to 97% for specific tasks like data entry from original documents. This is just the beginning. We envision AI enabling a fully interconnected document ecosystem—where every document is automatically understood, categorized and acted upon by intelligent systems, allowing humans to focus on strategic decision-making.
TFS: Dodocs.ai integrates large language models (LLMs) to process and draft documents. Can you share insights into the training process and how you ensure accuracy?
Dan Gudkov: We use a robust system involving four open-source LLMs, each running in parallel within a modular architecture that we’ve built in-house. This approach allows us to compare, validate and cross-reference outputs in real-time. One model might be better at interpreting legal clauses, while another excels at formatting financial tables. By running them together and analysing discrepancies, we can achieve higher accuracy and reliability.
Beyond that, we’ve built a self-improving document catalogue system that grows more intelligent every day. Thousands of documents flow through our system regularly and this exposure helps refine our models through continuous learning and pattern recognition. It’s like a neural network that gains experience just as a human would—learning what to expect, how to interpret context and how to structure output more efficiently over time.
TFS: AI-generated documents often raise concerns about reliability. How does Dodocs.ai balance automation with human oversight?
Dan Gudkov: We understand that trust is everything when it comes to automation, especially in enterprise environments. That’s why Dodocs.ai is built with configurable human-in-the-loop systems. Organizations can set checkpoints, approval flows and customizable review protocols before any AI-generated data is accepted into critical systems.
This flexibility ensures our users can maintain control without sacrificing efficiency. Supervisors can verify data, choose export formats or decide whether the output gets pushed directly into their ERP, CRM or document management systems. It’s about enhancing human expertise, not replacing it. By shifting tedious tasks to AI and allowing humans to make final decisions, we create an ideal partnership between automation and accountability.
Q6. With capabilities like video call transcription and chatbot automation, what are the most challenging technical problems you’ve had to solve?
Dan Gudkov: One of the biggest challenges we’ve faced is handling unstructured and poor-quality data. For example, documents that are handwritten, scanned with smudges or partially erased are extremely difficult to process accurately. Traditional OCR (Optical Character Recognition) tools often fail or return unreliable results in such cases. We had to build custom preprocessing pipelines that include noise reduction, smart segmentation and handwriting analysis powered by AI to extract meaningful data from even the most compromised inputs.
Another major challenge was integrating Dodocs.ai into the wide variety of ERP, POS and inventory systems used by different clients. Each system has its own API structure, security protocols and data formats. We overcame this by designing modular APIs and adaptive connectors that allow seamless integration, regardless of the backend infrastructure. Our approach was to build once and scale infinitely—so even if the systems are radically different, Dodocs.ai adapts without requiring bespoke development for every client.