Active Projects
From field tension to real-world transformation.
We turn communication patterns into working systems.
Every product we build is rooted in real team dynamics, real decisions, and real outcomes.
CoManager AI — Communication-Aware Task Assistant
Context: Mid-size product team with weekly sprints and constant cross-talk between PMs and engineers.
Problem: Tasks were assigned, but conversations revealed misalignment and silent blockers.
Solution: We created an AI-driven assistant that analyzes calls, chats, and retros to highlight what's really being said — and what’s not being acted on.
Outcome: Increased velocity clarity, reduced missed expectations by 42%, improved sprint focus.
Comanager.com
X2Bike — Real-Time Logistics Control via Telegram
Context: Courier-based bike service with fragmented task flow and chaotic coordination between dispatchers and riders.
Problem: Couriers missed deliveries, lacked accountability, and dispatchers were overwhelmed with phone calls and manual updates. No clear understanding of task status or location.
Solution: We built a lightweight Telegram-based control system that:
Displays real-time courier locations
Tracks task status transitions (new → picked → delivered)
Logs every action with timestamps
Allows managers to intervene early when delays appear
LLM Chain Processor — Multi-Step AI Workflows for Real-World Use
Context: Growing demand for real LLM integrations across complex business cases — not just chatbots.
Clients wanted AI tools that could process conversations, generate outputs, and adapt based on context.
Problem: Most LLM setups are either too rigid (single prompt) or too messy (manual stitching).
No orchestration, no visibility, and no way to explain “why it said that.”
Solution:
We built a modular chain processor that:
Orchestrates multi-step LLM calls
Allows branching logic, retries, context injection
Visualizes the flow and outputs in a clear graph
Works with OpenAI, local models, or hybrid setups
Equalize Team — HR Tech for Deep Motivation
Context: 200-person company struggling with uneven management quality and unclear growth tracks.
Problem: Team members felt “invisible,” leading to attrition and disengagement.
Solution: We built a system that connects 1-on-1 feedback, self-assessment, and transparent development plans.
Outcome: +90% employee retention, visible leadership improvement in under 2 months.
Polisher — Human-Grade Copy Enhancement Engine
Context: Leaders, creators, and consultants were struggling with voice and clarity.
They had something meaningful to say — but didn’t want to sound robotic or generic.
Problem:
Existing tools either over-formalized texts or killed the emotional tone.
Clients needed a tool that could amplify their voice, not replace it.
Solution:
We built Polisher — an LLM-based enhancement engine trained on live human dialogue and conscious writing.