About
I am Chinmaya Dash. Thirteen years leading data, analytics and AI at enterprise scale, most recently running the function for a multi billion dollar commercial and operations business.
I now run The Modern Analytics, consulting across data modernisation, analytics and BI, AI and agentic transformation, legacy modernisation, and the governance that decides whether any of it is allowed to go live.
Two engagements are running as I write this. A healthcare marketing and media agency, where I lead the whole chain from the data foundation through to agentic workflows in a regulated content environment. And a digital trust and AI platform company, where I have taken two platforms from definition to release with a third in build. Both are described here.
I do not sell development, and I am not going to. Having built things is useful to a client. Selling the build alongside the advice on whether to build is not, and the reasoning is here.
The numbers behind that
- $25M+ verified business impact across revenue growth, cost reduction and customer experience, from 20+ AI and analytics products.
- Analytics maturity from 46% to 88% in one year through operating model redesign.
- 50+ person cross functional team built and led across India, the US and Europe. Data engineering, analytics engineering, data science, product and BI, on an agile pod model.
- 25,000 users on the enterprise BI platform I owned at Novartis, against a $12M+ analytics portfolio.
- 10,000 users on a RAG based conversational AI platform, taking insight turnaround from days to minutes.
- $1.1M+ saved rationalising legacy platforms, and 1,500+ manual hours a month eliminated.
- 60%+ reduction in manual monitoring from agentic workflows for customer service and sales operations.
Where I have done it
TE Connectivity, leading data, analytics and AI for the commercial and operations business. Silicon Valley Bank, building the enterprise analytics function from zero. Novartis, owning the global BI platform and the drug development analytics portfolio. Earlier at JP Morgan Chase, KPMG and TCS, across risk, finance, retail, banking and manufacturing.
Life sciences and healthcare, industrial manufacturing, banking and financial services, commercial and operations. Regulated environments, mostly, which is where governance stops being a slide and starts being the thing that decides whether you go live at all.
The healthcare thread runs longer than most of it. The global BI platform and drug development analytics portfolio at Novartis, and now the agency work, where the constraint is not whether an automated step is right but whether the reasoning behind it can be reconstructed for a regulator months later.
What I actually do well
- Turning investment into evidence. Every product I have delivered carries a measured baseline and a named owner, which is why the impact figure survives scrutiny.
- Building the function, not just the thing. I have taken data and AI organisations from zero to enterprise maturity twice, including the operating model, the hiring and the succession pipeline.
- Governing AI that acts. Model governance across eight production models with versioning, lineage, drift detection and audit, plus authority boundaries for agentic workflows.
- Killing what should not run. Demand governance that eliminated 30%+ of low value initiatives from a $17M+ portfolio. Saying no is most of the value.
What I will tell you that others will not
Sometimes the answer is a licence you already own. Sometimes the process you want to automate should not exist. Sometimes your data team is right to say no. I have said all three and lost work for it, and I would rather that than the alternative.
Qualifications
- Completed the Chief Digital and AI Officer programme, Indian School of Business, 2026
- Postgraduate programme in AI for Leaders, Texas McCombs, 2024
- B.Tech Computer Science, KIIT University
- Databricks: Data Engineer Associate, Generative AI Engineer, Machine Learning Associate
- AWS Cloud Practitioner, Alteryx Designer, SAFe Agilist 5.0, Certified Scrum Product Owner
The full record is on LinkedIn, where I also write about data trust, AI adoption and proving value.
Let's talk.
Forty five minutes. Bring one thing that has not moved the way you expected, and I will tell you which gate it is failing at.