Start a conversation

Research

Writing on AI agents, enterprise systems, and the work we do.

Futuryntix research — enterprise AI orchestration

Product10 April 20267 min read

Why we don't build general-purpose agents

The AI industry keeps reaching for one large, capable assistant. Enterprise software has never worked that way — and neither will enterprise AI.

Read the article

All writing·ProductEngineeringIndustryPolicy

Engineering2 April 2026 · 9 min
How we built Smart Warehouse's memory architecture
A production AI agent without memory is a very expensive chatbot. Here's the three-layer model we use to make Smart Warehouse feel like a colleague who has been on the floor for years.
Industry24 March 2026 · 10 min
The state of warehouse AI in 2026
After a year of demos, a much smaller list of real deployments, and the quiet rise of the category nobody is talking about.
Policy15 March 2026 · 8 min
What India's 2026 EPR tightening means for enterprises
The new plastic and e-waste EPR rules are a once-in-a-decade compliance shift. Here's what changes, who is on the hook, and how AI changes the economics of keeping up.
Product28 February 2026 · 8 min
From SAP consultant to AI founder: the pivot nobody saw coming
Twenty years inside enterprise software taught me where AI will land hardest — and why the people building the future of enterprise AI are not who most expect.
Engineering14 February 2026 · 7 min
MCP is the best thing to happen to enterprise AI in a decade
The Model Context Protocol quietly solves the problem every enterprise AI project kept re-solving — and it will reshape how agents reach the systems of record.
Engineering30 January 2026 · 9 min
Evaluating enterprise agents: the eval harness nobody talks about
Agent evals for enterprise are a different beast from academic benchmarks. Here's the harness we use — and the one mistake that kills most internal agent programs.