Solve Knowledge Challenges with SEKA: An Ontology-Based AI Tool
The Problem

Throughout the software development lifecycle, managing and maintaining knowledge is a major challenge:
- Initial Development: 20-30% of workload
- Release & Deployment: 10-15%
- Support & Maintenance: 30-40%
- Evolution & Scaling: 10-20%
Software development generates a vast volume of interconnected knowledge about product requirements, product components, and functionalities. This knowledge is stored in documents, records, and artifacts. Over time, gaps and inconsistencies emerge, particularly between business needs and product components like APIs. This gradual degradation in product knowledge leads to poor decision-making, driving up costs, and extending development timelines.
Why Existing Solutions Fall Short
Limitations of Large Language Models (LLMs): LLMs like ChatGPT can’t process large knowledge bases efficiently due to memory constraints (e.g., a 128K token limit). These models struggle with large-scale, interconnected knowledge, creating a bottleneck in addressing gaps and inconsistencies.
Contradictions in Knowledge Management: As more people contribute to the product’s knowledge base, they create documents using varied terminology. This diversity, while reducing development time, leads to different interpretations when using LLMs—creating inconsistencies that make document verification and impact analysis difficult.

Our Solution: SEKA (Software Engineering Knowledge Assembler)

SEKA bridges the gap between vast product knowledge bases and the computational limits of modern AI. Our tool:
- Uses custom-built ontologies tailored to Software Engineering domains to markup documents, records, and artifacts.
- Automatically generates compact metadata for each item, linking it to the knowledge base.
- Assembles concise prompts for LLMs, enabling accurate and efficient document analysis.
Key Features
SEKA’s prompt templates cover a wide range of essential tasks, including:
- Verification: Identifying gaps and inconsistencies in new documents against existing knowledge.
- Traceability: Analyzing the impact of changes across the product knowledge base.
- Document Generation: Creating documents aligned with your domain-specific ontologies.
The first version of SEKA (MVP) will be available by November 30, 2024.

Custom Ontologies

SEKA uses well-established knowledge sources to build custom ontologies for different document types:
- Product Requirements Ontology: Based on a collection of templates from platforms like Figma, Intercom, Product Hunt, Amazon, and others. This ensures a thorough understanding of modern product requirement practices.
- Business Requirements Ontology: Derived from industry standards such as the Business Analysis Body of Knowledge (BABOK) and Business Technology Architecture (BTA), alongside a collection of business requirement templates.
- Functional Requirements Ontology: Grounded in both the Business Analysis Body of Knowledge (BABOK) and the Software Engineering Body of Knowledge (SWEBOK), supplemented with templates for functional requirements documentation.
- Software Architecture Ontology: Based on the Software Engineering Body of Knowledge (SWEBOK).
- Custom Ontology: Tailored to the specific document types used in your organization, providing flexibility for unique business needs.
Our Services
Our expert team can help you:
- Verify product knowledge, including requirements and functional specifications, to identify gaps and inconsistencies.
- Set up workflows for working with custom-built ontologies tailored to your knowledge management needs.
- Train your team on using SEKA and LLMs to streamline your product knowledge management processes.

Benefits for Your IT Team

- Enhanced Knowledge Quality: Eliminate gaps and inconsistencies across your product knowledge base.
- Accelerated Workflows: Radically reduce the time and effort required to create or modify product knowledge artifacts.
- Consistent Interpretation: Minimize discrepancies in understanding among contributors, even within large teams, for better knowledge alignment.
Use Cases
Here you can check some Use Cases
Already connected
Here you can try to work in Playground
Interested? Let’s Connect.
I’m
from , and I’m interested in the following:Additionally, I'm interested in setting up processes for working with ontologies and training my team (number of people:
)