avianna
A research lab for AI systems, agents, and the future of software work.
Avianna studies how intelligent systems reason, act, coordinate, and become governable inside real organizations.
What we study
AI is moving from passive tools to active systems that plan, decide, coordinate, and operate across software, data, and organizations. We focus on the systems layer beneath that shift: agents, workflows, contracts, governance, human delegation, and software architecture.
Research areas
Agent governance
How autonomous systems should be authorized, observed, constrained, and reversed.
AI-native software architecture
How software changes when agents become active participants in workflows.
Enterprise AI adoption
How organizations move from AI demos to durable operating models.
Human–AI collaboration
How people delegate, supervise, evaluate, and coordinate with AI systems.
Mathematical & strategic reasoning
How formal models, systems thinking, and decision theory improve AI and business design.
Applied prototypes
Experimental products and frameworks that turn research ideas into working systems.
Featured project: Concord
Concord is a project from Avianna exploring semantic contracts for agent-initiated work. It asks a simple question: when an AI agent takes action, how should an organization know what happened, why it happened, who authorized it, and how to reverse it? Durable execution alone can't answer that — you also need what the action meant, what policy governed it, and what systems it touched.
Also from the lab: Lattice — onboarding digital labor: the enterprise control plane. A three-part series.
Advisory
Avianna works with selected startups and enterprise teams on AI strategy, product architecture, agent workflow design, and technical positioning. The work is grounded in the lab’s research — we help teams reason through where AI changes the product, workflow, architecture, and business model.
For advisory or collaboration inquiries, contact the lab · more on advisory →
Essays and notes
Research notes, frameworks, and essays from the lab.
People
Avianna is led by people working across AI systems, data strategy, software architecture, and enterprise adoption — Tushar Madan and Rishubh Khurana.