Taskflo AI
A platform that lets companies run AI agent workflows inside their own cloud, with every action checked and logged. Built for teams whose security people ask good questions.
Explore Taskflo AIPractical AI strategy for founders & leaders
AI strategy that survives Monday morning.
I help founders and leadership teams work out where AI actually fits, and turn AI pilots into workflows their teams use every day.
What brings you here?
Building software since 2013. Founder at 19.Founder, Taskflo AIPhD researcher, IIT Jodhpur
Everyone has an AI strategy deck. Few have AI that works.
- Too many ideas. No clear priority.
- Twenty possible use cases, but no shared answer to which one matters first.
- Permissions are an afterthought.
- Nobody can say where the data goes or what the AI is allowed to touch.
- The pilot never becomes a habit.
- It worked in the demo. It didn't fit the people and process doing the work.
Diagnose. Decide. Build.
Find the blocker, choose the trade-off, then build something your team can own. The output is a useful next move, not a giant report.
Diagnose
Find the real blocker.
Trace the work, the people and the data. Separate the symptom from the constraint.
A workflow map- Current workflow
- Data & permissions
- One clear problem
Decide
Make the trade-off explicit.
Put value, risk, cost and team capability on the same page before choosing the build.
A decision sheet- Value & feasibility
- Risk & ownership
- A scoped next move
Build
Leave a path forward.
Turn the decision into an architecture, a delivery plan or the first focused workflow.
A working next step- Human checkpoints
- Observable actions
- Team handover
One question.
Different scales.
What does it take for a system to stay useful? I've followed that question from my first software at 15 to enterprise workflows and multi-agent research.
The longer storyFrom architecture to the workflow.
Enterprise platforms, AI integrations and the practical decisions around permissions, ownership and delivery.
What that looks like in an engagementHow should agents divide the work?
PhD research at IIT Jodhpur in multi-agent systems, fair division and allocation. Expected completion: 2028.
Explore the researchStart with the right questions
Before you build another pilot.
A short, practical guide to working out where AI fits in your business. For founders & leaders.
- Five questions to ask before an AI project.
- How to choose your first three workflows.
- The mistakes that stop a pilot becoming useful.
Free PDF + monthly letter
Get the AI Strategy Starter.
Confirm your email to get the PDF.
The inline signup is unavailable. A content blocker or connection issue may be preventing it from loading. You can subscribe directly on Kit instead.
Open signup on KitPlus Practical AI, one email a month. Unsubscribe anytime. Privacy.
Here for yourself? Get the free kit
Practical AI
From the monthly letter.
Practical AI for your business and your life: what works, what stalls, and how to make it stick. One long letter a month, no fluff.
Writing & the monthly letterbusiness
Which Tasks Should You Automate with AI? A Practical Framework
Decide which tasks go to AI and which stay with people. A practical framework for task allocation and human review, borrowed from fair division.
business
Why AI Pilots Fail (It's Rarely the Model): 5 Reasons and How to Fix Them
AI pilots stall on ownership, messy data, security and missing metrics. Learn the five reasons AI projects fail and how to move from pilot to production.
Got an AI idea that's stuck?
If you're trying to work out where AI fits in your team or your company, I'm always happy to talk. No pitch deck needed.