AI Insights: Ivan Ilin on Transforming Information Intelligence

Have you ever imagined how AI can create your personalized wiki, seamlessly bringing all your essential information into one intelligent interface? Meet Ivan Ilin, co-founder of IKI, who is making this vision a reality.

Welcome to another edition of our exclusive interview series on AllAboutAI.com. Today, we are thrilled to present insights from Ivan, a seasoned machine learning expert with a PhD and over eight years of experience in NLP.

With a rich background in applied mathematics and extensive experience in search and information retrieval, Ivan has crafted IKI AI to revolutionize knowledge management for professionals and teams.

Join us as we explore Ivan’s journey, the groundbreaking features of IKI AI, and AI’s transformative impact on the future of knowledge management.

Here’s the complete interview!

IKI AI – More Than a Personalized, Efficient WIKI

Explore how IKI AI revolutionizes knowledge management with an intelligent interface designed to organize, search, and utilize information effectively. In this interview, co-founder Ivan Ilin shares insights on the unique features of IKI AI, its impact on productivity, and its vision for the future of AI in knowledge management.

How did your background and experiences lead you into the AI industry and, ultimately, the creation of IKI AI?

Ivan Ilin: I was doing a PhD in Applied Mathematics, focusing on trajectories designed for scientific spacecraft for space telescopes. This field naturally led me to machine learning. Around 2015, I explored online trends, tried quantitative trading, and ventured into time series prediction and forecasting. Eventually, I discovered the fascinating world of natural language understanding, which drove me to apply this technology to help people gain professional knowledge and navigate their careers.

Did you conduct market research before building IKI AI? 

Ivan Ilin: Yes. We understood that there’s a significant need for a personal wiki for professionals because we are overloaded with information. Some bookkeepers, bookmarks, and note-keeping tools like Evernote are on the market, but they don’t have AI under the hood or an excellent full-text search implemented. You could save stuff, but navigating it was challenging. We focused on solving this problem by building a visual, self-organizing library with good UI and UX for day-to-day usage.

What are IKI AI’s main objectives, and does IKI aim to transform knowledge management?

Ivan Ilin: All your knowledge is just a question away, ultimately helping you gain knowledge practically because most current systems are reactive. You have to query things; you have to query Google to come up with new ideas and solutions. So you have to start this process on your own. However, the next generation of tools should be proactive, not reactive. For example, I don’t have to query everything myself if I’m researching deploying a large language model on a chip; I just need an agent to gather this information and develop a solution.

How does IKI AI differ from traditional knowledge management tools like Microsoft OneNote and Evernote?

Ivan Ilin: We have a few key differences. One is that we grab the full text of any saved source, making us quite source-agnostic. We read any web page saved to Iki. We aim for around 90% accuracy in fully extracting text to enable content discussion. We also support YouTube videos via Boosted Transcription. But these are technical differences. The UI is vastly different in its ease of use and visual appeal. Many products claim to have a Copilot but don’t extract information from your links, making it unclear where it comes from. We simplify this and offer a community and index of users’ public resources. You can search this curated index, benefiting from your library and your peers’ collective knowledge.

How do IKI’s capabilities improve productivity compared to conventional tools?

Ivan Ilin: For example, when researching agents and models, I found numerous relevant resources scattered across web pages. Navigating these resources was only possible by having them centralized, so I created about 50 web pages describing various algorithms and architectures. With everything in one place, I could easily query the collection with questions like, “What’s the difference between self-reflection and React architecture? What are the benefits and downsides? How are they implemented in the LAMA index?” This approach allowed me to get comprehensive answers without manually comparing the information from different sources.

How is IKI different from competitors like Brandfolder and MediaValet?

Ivan Ilin: First of all, we have a community aspect. The previous question partially answers this. We aim to read everything people save and interact with, which is crucial for compiler quality. Regarding UI/UX, we designed the product to be easy for Dutch legal professionals, who differ from Silicon Valley tech geeks. This differentiation is significant. I know good products and projects that started from GitHub repositories, which are sophisticated but challenging for non-tech people to adopt.

< class=”table table-sm table-striped table-bordered small”>

What are the key Features of IKI AI?

Ivan Ilin: So the key features are, first of all, your personal library, where you can drop stuff, and then you get nice cards. You can open any source inside Iki and read it inside without going to the webpage. Then you get compilers in any context. So you can chat with a single document or material, chat with your whole library, chat with a collection, and share a collection with a copilot based on it. So, it’s a super viral and easy way to get and share media knowledge on social media. Then there is search, Copilot in different contexts, the Google IKI Index, web search, and your library. Then, there is a community part where you can share your collections with the people on the platform and check what your peers save to their libraries if they make it public because you can always choose if your material is public or private.

<class=”table table-sm table-striped table-bordered small”>

How does IKI AI handle the diversity in data formats?

Ivan Ilin: We support PDFs, web pages, and soon DocX. We also support TXTs, YouTube videos, and note-taking. Web pages include resources like Medium and GitHub. While we can’t parse everything due to some protections, we can save LinkedIn posts well. We prioritize knowledge sources familiar to professionals in machine learning and tech.

How does IKI AI ensure the accuracy and relevance of AI-generated recommendations?

Ivan Ilin: We initially had a recommendation engine and personal feed of recommendations, but it didn’t solve the information overload problem; it created another unnecessary information flow. So, we deprecated that functionality. Now, we have “Related Materials,” which appear on the right side of the screen based on title similarity when you save something. It’s a good discovery pattern. Additionally, augmenting your collection with relevant materials is an option that works well for researching topics. You just hit a button, and related knowledge from Iki on that topic is automatically added.

Does IKI offer information retrieval and knowledge discovery? 

Ivan Ilin: The answer is NO, but we’re moving towards an agentic approach. If you’re researching a topic, IKI AI gathers knowledge from top sources like LinkedIn and Medium, then sends you a relevant digest via notification—a mechanism researchers will appreciate.

What integrations does IKI AI offer?

Ivan Ilin: We have prioritized integrations with Notion, Obsidian, and Google Drive, which are coming in late summer. These integrations are based on user demand, allowing them to upload their knowledge to Iki during onboarding seamlessly. Next, we will focus on Slack and link collector integrations. We’re unsure about integrating with the Microsoft ecosystem, as it operates independently.

How does IKI manage the cross-availability of digital libraries?

Ivan Ilin: Right now, we have sharing options for collections. You can share them online through a web link. We also offer collaboration on collections and will introduce team spaces soon. This feature will allow you to drag and drop knowledge from your private library to a team space and back. Multiple accounts can collaborate within these team spaces, which is essential for small and medium businesses working on a company knowledge base. Beyond this, we focus on integrations to enhance functionality.

How does IKI AI track permissions and access?

Ivan Ilin: If you collaborate on a company’s knowledge, you can keep it private so only your team members can see the activity within the space. If you share materials publicly, you can create public notes with features like likes and comments. You can see interactions with your material but not extensive statistics. While detailed analytics aren’t available yet, we plan to include views and other metrics soon.

Does IKI protect user data?

Ivan Ilin: Yes. We leverage AWS infrastructure to address issues within their ecosystem effectively. Our operations are SOC 2 compliant, ensuring robust security and compliance standards.

Can you share some success stories or case studies where IKI AI significantly improved a client’s knowledge management processes?

Ivan Ilin: First, Dutch lawyers could not search within many PDFs containing public court decisions they referred to when working on new cases. We helped them by creating a data pack of these public court decisions, allowing them to augment it with their documents and cases. This made it easy for them to describe their case, ask questions, and get drafts of their future solutions, saving time. We also have companies collaborating on technical projects, where sharing tech knowledge within their team is made easy by creating and sharing collections. This ensures that team members are aware of the most essential sources. We also have requests from larger companies and a network of universities interested in using our solution for their students. This resonates with academia and helps shorten the path to knowledge discovery and sharing.

How do you see the role of AI evolving in the knowledge management industry over the next five years?

Ivan Ilin: Oh, it’s a super hard question because, you know, the pace is so fast now that it’s challenging to predict even where we’ll be within a year. We see moves of Gen AI coming to consumer apps. Apple made a massive move in the area, and Google will follow. So, horizontal solutions will be owned by tech and humans. There’s a lot of opportunity in the niche space, especially in the B2B area, with many different solutions. Even industries and many cases will be automated by just taking the existing workflow and automating it with a genetic workflow or augmenting it with knowledge injections at different points in the process because many knowledge workers are doing research manually and are involved in routine pipelines that they must execute manually. That’s all going to be automated for sure. But right now, the problem is that you’ve probably seen this article about this mismatch between Gen-AI investments and revenues, a recent publication by Sequoia. There is a gap right now because investments have spiked, but the returns are still few companies are making good returns except OpenAI, and they’re yet to come.

Can Artificial Intelligence Replace Humans?

Ivan Ilin: No. We’ll see the market eventually balance out. Solutions that generate revenue will gain more traction, while some visionary concepts might diminish, leading to less hype around Gen AI and more practical applications. AI will balance out the market with revenue-generating solutions gaining traction. Next-gen models may replace many general-purpose knowledge workers but not experts. This shift will impact the labour market, requiring effective mitigation strategies. However, next-generation models could be game-changers, potentially reaching a higher level of reasoning. This advancement could replace many general-purpose knowledge workers, though not the experts. Such a shift would significantly impact the labour market, posing challenges in finding effective mitigation strategies. The evolution of these models will bring substantial changes, and we’ll have to wait and see how it unfolds.

Any advice for young entrepreneurs entering the AI-driven tech industry

Ivan Ilin: You’re in the most thrilling era of AI. Move quickly, identify client needs, and evaluate the value before building anything. Avoid wasting time on unnecessary products; follow market trends and your passion. You’re in the most thrilling era of AI right now. Move quickly, identify what resonates with potential clients, and don’t build anything before evaluating its value. Make sure not to waste time on products people don’t need. Instead, follow market trends and your passion.

Key Takeaways

In our interview with Ivan Ilin, co-founder of IKI AI, he shares how his background in applied mathematics and machine learning led to the creation of IKI AI, an intelligent knowledge interface designed to manage information overload.

IKI differentiates itself by offering proactive AI-driven suggestions, visual UI/UX, and a community-based public index for collaborative knowledge sharing. The platform supports diverse data formats and aims to enhance productivity by simplifying access to complex information.

Ivan anticipates significant automation in knowledge management through AI, transforming how professionals utilize information. He advises aspiring entrepreneurs to validate their ideas with potential clients and focus on building market-needed solutions while following their passion.

Enjoyed this insightful conversation with Ivan Ilin from IKI AI. Dive into more fascinating discussions with AI pioneers by exploring our exclusive interview series on AllAboutAI.com. Stay updated with the latest AI innovations and breakthroughs, gain industry experts’ valuable knowledge, and be inspired by their journeys. Visit our Interview Series to read more and join our community of AI enthusiasts!