Google Cloud.

The bottleneck

Optimising the right talent for impactful, complex projects.

Three large teams in a 400-person org shared one capacity bottleneck: matching the right people to each complex project.

The agent

The agent suggests. People decide.

I shipped a RAG-based agent to the org and owned it end to end: the requirements, the model and retrieval trade-offs, the rollout, and adoption tracking.

1. RequestA complex project needs people
2. RetrievalVector search finds the right context
3. ModelDrafts a recommendation
4. Human reviewA person approves every decision
5. DecisionRolled out and tracked for adoption

A person reviews every step where a decision is made. The agent does the searching; people keep the judgement.

Shipping safely

Faster releases, and every one traceable.

Agents in a regulated internal environment need guardrails. I defined and enforced the gates every agent passes before it ships.

1. BuildA new agent or update
2. Compliance reviewAutomated checks
3. ApprovalA set workflow, no side doors
4. Audit trailEvery release on record
5. ReleaseShipped

Intelligence for the field

A weekly brief for 400 sellers.

I built a fully automated pipeline that sends a 400-person field sales team a weekly, image-rich brief on competitors and the industry: technical breakthroughs and strategic wins from across the business.

1. SourcesBreakthroughs and wins across the business
2. PipelineAutomated end to end
3. Weekly briefImage-rich, every week
4. 400 sellersAcross the field team
5. InsightsEngagement data for leadership

The engagement data showed what landed, and leadership used it to tailor go-to-market.

Clean data underneath

I also owned a process-excellence push on Salesforce data quality. I triaged potential revenue opportunities and removed the blockers in how data gets captured.

Gemini for medical schools

Human anatomy, in 3D, on request.

I built an MCP for Gemini Enterprise that returns 3D structures of human anatomy. I pitched it as a frontier MCP for education, starting with medical schools across India.

1. QuestionA student asks in Gemini Enterprise
2. Anatomy MCPFinds the structure
3. 3D modelReady to explore
4. ClassFaster teaching and learning

Where AI grows next

Finding the next market for agentic AI.

I run end-to-end market and operational diagnostics on Indian markets, using deep research and data analysis to find high-impact go-to-market openings across APAC. With senior leadership, I turn them into a technology roadmap and a ranked set of initiatives for scaling agentic AI products in fast-growing, billion-dollar markets.

Case in point: banking and fintech

1. Research with buyersUser and market research across Indian BFSI and fintech buyers.
2. Map the real blockersEnterprise procurement and compliance rules, the things that stall adoption.
3. Prioritize the roadmapA ranked plan for Gemini Enterprise adoption in the sector.

Community

Sharing what works.

I keep a global forum of APAC Cloud GTM managers running, where people across time zones swap what works in account management. And at Google Cloud events, I've shown Fortune 500 leaders how to get real work done with Gemini, Veo and NotebookLM.