Fiverr's Shift Towards AI and the Role of Knowledge Graphs

Fiverr Platform: Fiverr, the leading platform in the gig economy, plans to lay off 250 employees, equivalent to 30% of its workforce, as part of its strategic transformation to become an «AI-first company». The company's CEO, Micha Kaufman, announced this decision via an article published on platform X.

Company Trend: This trend comes as part of a growing wave among technology companies in 2025, with other companies, such as Duolingo, adopting similar plans to transition to an «AI-first» model. Kaufman described this process as a return to «startup mode,» aiming to make Fiverr more agile and faster, with a modern tech infrastructure focused on Artificial Intelligence, a smaller team with significantly higher productivity, and fewer management layers.

Justifications for Layoffs: Justifying that Fiverr no longer needs the same number of employees to operate its current business, Kaufman noted that the company has already integrated Artificial Intelligence into its customer support and fraud detection programs.

Previous Advice: It is worth noting that Kaufman had previously pointed out the risk of technology to employees in a May 2025 interview with CBS News, advising them then to «100% automate» their tasks using Artificial Intelligence, claiming that this would not make them replaceable because they still possess the ability for «nonlinear thinking» and «making difficult decisions». However, this advice ultimately did not seem beneficial to Fiverr's employees themselves.

Comparison of Layoffs: Although the layoffs at Fiverr are fewer in number than those announced by larger companies like Workday, which planned to eliminate 1,750 jobs in February 2025, the impact of these decisions remains the same: fewer employees bearing a heavier workload.

What is a Knowledge Graph?

Knowledge Graph: Also known as a semantic network, it is a structured representation of real-world entities such as people, places, events, and concepts, illustrating the relationships between them. This information is typically stored in a graph database and visualized as a graphical structure. Knowledge graphs contextualize data through linking and semantic metadata, providing a framework for integrating, unifying, analyzing, and sharing data. These graphs are used to store interconnected descriptions of entities with encoded semantics or free-form relationships that underpin these entities. They are an effective tool for organizing information to facilitate access and understanding, allowing question-answering systems and search engines to retrieve and reuse comprehensive answers to specific queries.

Key Components of a Knowledge Graph

Key Components: A knowledge graph typically consists of three main components: Nodes, Relationships (or Edges), and Organizing Principles (or Schemas).

• Nodes: Represent real-world entities such as people or places or things or events or concepts. Each node typically has a label (or multiple labels) to identify the type of node, and may optionally contain one or more properties (attributes).
• Relationships (Relationships/Edges): Connect two nodes to each other and illustrate how entities are related. Just like nodes, each relationship has a label that defines the type of relationship and may optionally contain one or more properties.
• Organizing Principles (Organizing Principles) or Schema (Schema) or Ontology (Ontology): Are a framework or blueprint that organizes nodes and relationships according to the fundamental concepts necessary for the respective use cases. An organizing principle can be thought of as a conceptual map or an overlaying metadata layer on top of the data and relationships in the graph. Definitions of knowledge graphs vary, but they generally agree that the basic components are nodes and relationships that describe entities and how they are related.

Benefits of Using Knowledge Graphs

Numerous Benefits: Knowledge graphs offer numerous benefits that make them a powerful tool for transforming data into actionable insights. These graphs help standardize data access, provide flexible data integration, and automate data management, fostering a deeper understanding of complex relationships between different data points.

Key Benefits: Among the key benefits are:

• Improved Search Results: Knowledge graphs are transforming how we search for and find information across the web, organizing facts about people, places, and things into a network of entities. When performing a search, the connections between entities are used to display the most contextually relevant results, such as what appears in Google's Knowledge Panel.
• Streamlined Data Integration: Knowledge graphs provide a framework for linking diverse and heterogeneous data sources, allowing information from different parts of a company to be integrated, even from "traditional data silos." This enhances information sharing and reuse.
• Revealing Hidden Patterns and Relationships: By analyzing relationships and patterns in a knowledge graph, it becomes easier to detect potential fraud or security threats, or to identify distant relationships between entities that may not be apparent in traditional data.
• Supporting AI and Machine Learning Systems: Knowledge graphs are used to ground Large Language Models (LLMs) with domain-specific or company-specific information (GraphRAG technology), which increases response accuracy and improves interpretability through the context provided by data relationships. They are also useful in drug discovery in healthcare, where they organize relationships within medical research to support diagnoses and treatment plans.
• Enhanced Operational Efficiency: In supply chain management, a knowledge graph can represent the network of suppliers, raw materials, products, and logistics, enabling comprehensive visibility, identification of weaknesses, and prediction of disruptions. They also support Master Data Management (MDM) systems for a unified view of customers.
• Knowledge Enrichment and Generation of New Knowledge: Data integration efforts around knowledge graphs can support the creation of new knowledge by establishing connections between data points that might not have been previously recognized.

Use Cases of Knowledge Graphs

Versatile Tools: Knowledge graphs are versatile tools that can be applied across a wide range of industries and tasks, helping organizations gain a deeper understanding of their complex data and extract valuable insights. Common use cases for knowledge graphs include:

• Search Engines and Recommendation Systems: Major companies like Google, Facebook, and Amazon use knowledge graphs to improve the accuracy of search results and personalize product or content recommendations for users based on their interactions and behavior.
• Fraud Detection and Financial Analytics: In the financial services sector, knowledge graphs are used to represent networks of transactions and their participants, enabling the identification of suspicious activities, investigation of potential fraud, and combating money laundering by understanding the flow of funds and relationships between customers.
• Master Data Management (MDM): Knowledge graphs provide a structured and comprehensive database of customer data and interactions, ensuring a unified and accurate view of customers, especially for companies with multiple departments or applications dealing with clients.
• Supply Chain Management: Knowledge graphs help represent the network of suppliers, raw materials, products, and logistics. This comprehensive view of the supply chain enables managers to identify weaknesses, predict disruptions, and optimize logistics routes in real time.
• Investigative Journalism: They are used to link key entities (companies, people, bank accounts) and activities under investigation. Organizing these entities allows for the discovery of hidden patterns, such as distant relationships between entities that may not be clearly apparent.
• Drug Discovery and Medical Research: In healthcare, knowledge graphs store and organize relationships within medical research, such as protein and genome sequences with environmental and chemical data, which helps verify diagnoses and determine individualized treatment plans.
• Generative AI Applications: Knowledge graphs are used to organize domain-specific information and serve as a foundation for Generative AI applications that use private data (RAG applications), which enhances response accuracy and improves interpretability.

Future Trends in Knowledge Graphs

Continuous Developments: Knowledge graphs are experiencing continuous and promising developments, driven by advancements in Artificial Intelligence and Machine Learning. Their applications are expected to expand and their efficiency to increase, making them a fundamental component in many intelligent systems.

Key Trends: Key future trends include:

• Machine Learning and Knowledge Graph Embeddings: Knowledge graphs will continue to play a central role in representing information extracted using natural language processing and computer vision. Domain-specific knowledge expressed in knowledge graphs will be fed into machine learning models to produce better predictions.
• Entity Alignment Enhanced by Large Language Models (LLMs): As the amount of data stored in knowledge graphs grows, developing reliable methods for knowledge graph entity alignment has become an increasingly critical step in integrating and cohering knowledge graph data. Recent successes of Large Language Models, particularly their effectiveness in producing syntactically meaningful embeddings, have propelled the use of LLMs in the task of entity alignment.
• Virtual Knowledge Graphs: Unlike traditional knowledge graphs that store information in specialized databases, virtual knowledge graphs will increasingly rely on underlying relational databases or data lakes to answer queries on the graph, providing greater flexibility and efficiency in data management.
• Integration with Generative AI: The role of knowledge graphs will become more crucial in providing reliable and accurate context for Generative AI models, especially in applications such as chatbots and question-answering systems, ensuring more informative and precise responses.
• Automation of Knowledge Graph Creation: Future developments aim to automate the creation of knowledge graphs and deepen the integration of AI models to achieve more trustworthy and intelligent systems.