Field notes
Ideas on AI, automation & building well
Practical thinking on applied AI, automation, and engineering done well — cutting through the hype to what actually moves the needle.
More field notes
When the internet doesn't cooperate: offline-first systems for far-west Nepal
Lessons from delivering education technology into Kailali and Kanchanpur — why offline-first is an architecture, not a feature you bolt on, and why sync is where the honesty is.
Where agentic AI pays off first: an honest ROI map for 2026
Only 23% of organisations are scaling AI agents, and no more than 10% in any single function. The teams seeing real return aren't chasing autonomy everywhere — they're picking the few workflows where agents genuinely pay. Here's how to find yours.
Shipping AI agents to production in 2026: the playbook behind the 40% that get cancelled
Gartner says 40% of enterprise apps will ship a task-specific agent by end of 2026 — and that over 40% of agentic projects will be cancelled by 2027. Here is the engineering that decides which side you land on.
What AI actually changes for a Nepali SME
Where applied AI genuinely helps small and mid-sized businesses in Nepal — the busywork it removes, the real hours saved, and the parts that are still hype.
How automation ROI actually works (with the math)
Where the time and money really go, what to automate first, the payback math, and the pitfalls — a practical guide to automation ROI for a Nepali business.
Nepal's fintech at scale: how a QR tap got instant
Nepal moved Rs 98 trillion digitally in FY 2024/25, with 24 million mobile-banking users. Here is how the wallets, rails and QR came together, and what comes next.
What makes software world-class, built from Kathmandu
World-class software isn't a postcode — it's a set of engineering standards. Here are the five we hold every build to, and the evidence behind them.
Agentic AI in 2026: past the hype, into your workflows
AI agents went mainstream in 2026 — Gartner sees 40% of enterprise apps shipping one, yet 40% of projects will be scrapped. The playbook for a Nepali team.
The AI adoption gap: redesign the work, not just the tool
Almost every company now uses AI, yet few see real gains. The difference is not a better model. It is redesigning the workflow around it. Here is what that takes.
Nepal is climbing from outsourcing to its own products
Nepal's IT exports near $1 billion and startups have raised $258M. The next move is up the value chain — from selling hours to selling software. Here's the shift.
Frontier vs efficient AI: why the cheaper model wins
The biggest AI models grab the headlines, but for most everyday business work a small, cheap model does the job for a fraction of the cost. Here is how to choose.
Nepal's 2083/84 budget: a 50% IT-export rebate and a weightless digital economy
Nepal's 2083/84 budget decoded for founders: a 50% IT-export tax rebate, a 100% sweat-equity exemption, a sovereign AI compute centre, and startup support.
How a Nepali dev team competes globally in 2026
Low cost is not the whole story anymore. A Nepal team wins on AI-augmented output, a rare time-zone offset, and real quality. Here is the honest playbook.
From AI tools to AI teammates: what orchestration means for your team
The jump in 2026 is not smarter chatbots. It is many small agents working together on a whole job. Here is what that shift looks like for a Nepali team.
UPI meets NPI: India-Nepal money now moves in seconds
India-Nepal cross-border remittance went live on the UPI-NPI rail. Here is how it works, the real limits and fees, and what it means for Nepali builders.
Google now answers instead of listing. How Nepali businesses stay visible
Google AI Mode writes an answer instead of showing ten links. If your business is not named in that answer, you are invisible. Here is how to get cited.
Nepal's AI moment in 2026: the policy, the money, and what it means for builders
A national AI policy is now on the books and real capital is flowing into Nepali startups. What's new in 2026 — and what a Kathmandu builder should do about it.
Why Nepal building its own AI compute matters more than it sounds
Nepal's 2026 budget announced a national AI compute center in Kathmandu. Behind the jargon is a simple question: whose machines, and whose law, run our data?
Nepal just cracked the global outsourcing top 20. Now the hard part.
Nepal ranked 19th in the 2026 Global Outsourcing Talent Index — top 10% worldwide. What the ranking rewards, the gaps it exposes, and how to turn it into work.
Your company data is the moat: build a context layer for AI
Anyone can rent the same AI models you can. What they cannot copy is your own data, joined up. Here is why a context layer is the real advantage.
➜ ~ ready to ship
Ready to build what's next?
Tell us about your project — we'll reply within one business day with a clear plan and a straight answer on fit.