Problems

The questions I keep returning to.

  • When AI systems act and transact on behalf of humans, who is accountable when they err - and what infrastructure catches that before damage is done? Pursuing it through Commerce Suite and Compass GovernReady - governance before deployment, accountability after.
  • How do organisations build meaningful AI governance before deployment, not compliance frameworks retrofitted after the fact? Pursuing it through Compass Suite - four assessments covering the full governance arc from readiness to continuous compliance monitoring.
  • How do you govern a resource that grows exponentially while remaining almost invisible to the people paying for it? Pursuing it through Steward - token-level visibility and control of enterprise AI spend.
  • What would it take for public institutions to trust AI enough to depend on it for consequential decisions - and who validates that trust? Pursuing it through NitiScope and AI Envisioning work with 583 civil servants and PSU executives across government and state enterprises.
  • What does it take to reliably separate signal from noise when information is abundant and attention is scarce? Pursuing it through getHeard - media intelligence and stakeholder research.
  • If one person can work at the leverage of a team, what new responsibilities come with that power? Living it daily - Dendrons is built and run with the agents it ships.
  • Can a machine tell us something true about ourselves that conventional assessments miss? Pursuing it through MirrorMirror.life - personality assessment built on narrative psychology.

Products

From public procurement intelligence to enterprise AI governance. Built at Dendrons.ai.

Procurement intelligence

NITI Scope

What does India's public-procurement record actually reveal?

by Dendrons.ai
1 in 8 government awards drew only a single bidder
3x less competitive - central procurement vs. states
34% of awards land on exactly 3 bids - the minimum competitive threshold
Explore more at nitiscope.com →

Personal Agents

Tools built for my own workflows. They work. I use them daily. Not yet productized - but that may change.

Verne built on NanoClaw

Personal AI assistant built on NanoClaw, an open-source framework I forked and extended. The customisation layer - memory management, WhatsApp and web interfaces, and my specific daily workflows - is what became Verne.

Named after Jules Verne - who imagined worlds that didn't exist yet, and wrote them into being anyway. It felt like the right name for a personal AI built from scratch.

Know more →

Arsène

Task management agent. I forward emails to it and it builds and maintains a Kanban board of what needs doing. Keeps incoming chaos organised without manual triage.

Named after Arsène Lupin - the fictional gentleman who was always three steps ahead, never lost track of anything, and made it look effortless.

Know more →

Thoth AI

Form-filling agent. Handles repetitive documentation and application workflows.

Named after Thoth - the Egyptian god of writing, knowledge, and record-keeping. The divine scribe felt like the right name for an agent that exists to handle paperwork.

Know more →

Dendrons Suites

Intelligence Suite

getHeard

Live

Know what's being said. Know what people actually think.

Media intelligence and stakeholder research for government, public affairs, and research teams - from signal extraction to structured qualitative insight.

Explore suite →
Conversion Suite

getBookings + Newton

Live

Turn high-intent searches into confirmed visits.

Two AI agents that close the gap between online interest and in-person commitment. getBookings hunts for high-intent visitors; Newton converts them into booked appointments.

Explore suite →
Governance Suite

Compass

Live

AI ready. Data ready. Regulator ready.

Four assessments for the full governance arc - from AI readiness strategy to continuous compliance monitoring.

Explore suite →
AI Operations

Steward

Pare live · Rein in private access

Every token. Every dollar. Accounted for.

Rein governs AI spend at the enterprise level - metering, budgets, and an audit trail. Pare shows individual developers what their AI usage actually costs, privately and locally.

Rein Pare
Explore Steward →
Agentic Commerce

Commerce Suite

Coming soon

Rank higher. Transact smarter. Stay secure.

Be found in AI search, ready for agentic buyers, and trusted by regulators. Three products for businesses entering the agentic commerce era.

RankReady TransactReady TrustReady
Learn more →
Standalone

MirrorMirror.life

Live

Personality assessment built on narrative psychology. Tells you something true about yourself that standard tests miss.

Visit →