Teklens is an AI product manager for software teams that connects product decisions, Jira tickets, and actual code into a persistent context graph. It keeps humans and AI coding agents aligned from decision to release, providing code-grounded findings in under 10 minutes. The platform integrates with Jira, Confluence, and GitHub without replacing them, and offers model-agnostic AI with EU/CH hosting.
Funding
Funding not disclosed
Founders
Product
Problem
Software development has become fast with AI coding agents, but coordination has not. Product decisions often happen outside of Jira, every agent needs someone to brief and review it, and the more work runs in parallel, the more someone has to manually hold everything together—from decisions to specs, tickets, reviews, and stakeholder updates.
Solution
Teklens provides an AI product manager that connects product decisions, Jira, and actual code into a persistent context graph. Its Context Engine links conversations, tickets, specs, and code into a living graph of requirements, decisions, and people, so every idea is tied to affected components and every estimate is grounded in the repository. The platform works across the entire product cycle—Discover, Define, Build, Operate—by clustering signals, prioritizing based on value versus effort and risk, generating code-grounded PRDs and acceptance criteria, and feeding operational learnings back into planning. It integrates with existing tools like Jira, Confluence, and GitHub rather than replacing them, and delivers answers directly in the tools where teams already work.
Target Audience
Primary customers are software product and engineering teams that use AI coding agents and need to maintain alignment between human decision-making and automated development work, particularly those using Jira, Confluence, and GitHub.
Features
- Context Engine that builds a living graph connecting code, requirements, decisions, and people, avoiding copy-paste context into prompts
- Code-grounded prioritization with rationale based on value, effort, risk, and complexity, grounded in the actual repository
- Master playbook that keeps the canonical state per initiative and makes decisions traceable
- Model-agnostic AI with per-project model choice (Claude, GPT, Gemini) and CH/EU hosting with no training on customer data
- Live agent team per initiative (PM, PO, Engineering) working in the same context along conversations and artifacts
- Risk analysis and edge case identification from real code paths before the build phase
- Integration with Jira, Confluence, and GitHub, with a web app and meeting bot for the same context across tools