Welcome to Analytics in the AI Era — a weekly newsletter for the people doing the actual work.
Every week brings a new model, a new framework, a new set of claims about what AI can now do that it could not do before. Some of it matters. Most of it does not. The hard part is knowing which is which, and that takes time most people working in data and analytics (D&A) simply do not have.
That tension is what this publication is about.
The function itself is changing
Generative AI has not just added new tools to the D&A stack. It is putting pressure on the structure of the function itself. The teams that exist today may be smaller, differently composed, or organised around entirely different problems within a few years. Processes that are currently human-dependent are being automated in some places and proving stubbornly resistant to automation in others. The skills that get someone hired today may not be the skills that make them valuable in three years.
Most of this transition is happening without a clear blueprint. That is not a failure of planning. It is what happens when the pace of change outstrips the time it takes for best practices to emerge.
What is worth naming clearly: this is not a tools story. It is a function story. The question is not just "which AI capabilities should my team use?" It is "what does an effective D&A function look like when these capabilities are assumed?" That is a harder question, and most organisations are not asking it yet.
Nobody actually knows what success looks like
The organisations generating the most confident AI content are often not the ones with the most rigorous answers. The playbooks are being written in real time, and a lot of what gets presented as settled wisdom is confident speculation dressed in case study clothing.
In that environment, a well-reasoned framework for how to approach a problem is probably worth more than a perfectly engineered AI solution. The organisations getting traction are not the ones who have figured out the right answer. They are the ones making thoughtful bets, learning fast, and adjusting before the cost becomes significant.
That distinction matters for how D&A leaders should be spending their time and attention right now. Not chasing every tool announcement. Not deferring until the market settles. Building clearer ways of thinking about which changes actually matter and why.
We do not need to start from scratch
Many of the challenges D&A teams are navigating right now are not new problems. They are familiar problems in unfamiliar form.
How do you build trust in a capability that decision makers do not fully understand? How do you scale something that worked as a pilot but falls apart in production? How do you upskill a team while still delivering on existing commitments? How do you make a case for investment when the outcomes are genuinely uncertain?
These questions have been answered, imperfectly but usefully, across product management, organisational design, change management, and software engineering. There is no reason the D&A community needs to discover every answer independently. Borrowing and adapting thinking from adjacent disciplines is underused in this space, and I want to change that.
Why this publication exists
There is no shortage of AI commentary right now. What is genuinely hard to find is writing aimed specifically at the people running D&A functions: leaders deciding how to restructure their teams, practitioners figuring out which skills are worth developing, analysts trying to understand what their role looks like in two years.
Most AI content is written either for a general audience or for software engineers. The D&A practitioner is largely being asked to translate generic guidance into function-specific decisions without much help. That is the gap this publication is trying to fill.
The best answers will not come from any single voice. They will come from practitioners sharing what they are actually seeing, pushing back on frameworks that do not hold up in practice, and building a more honest picture of what is working and what is not. That is the conversation I want this publication to be part of.
Who I am
I am Abhinav. I have spent the last decade implementing analytics and AI in large enterprises across financial services, loyalty, and retail. I have started ventures both independently and within existing businesses. I run side projects with emerging tools because hands-on experience with what is actually being built is the only reliable way to separate genuine capability from well-marketed noise.
My most recent side projects have involved building agentic AI applications: a text-to-SQL-to-visualisation engine, an analytics concierge system, and a personalised content generation application. Building is how I develop a view worth sharing.
Who this is for
This publication is written for D&A leaders and senior practitioners: people who are responsible for how their organisations use data and AI, and who want a grounded, peer-level perspective on what is actually worth paying attention to.
You do not need to have it figured out. Neither do I.
What to expect
Every edition will be one of three things.
AI Watch. The D&A space produces more signal than anyone can reasonably track. I will filter what is worth your attention, translate new developments through a practitioner lens, and say clearly what I think it means for the work we do.
Point of View. My perspective, argument, or framework on a topic. Not a summary of what others think. A specific position, held with conviction and backed by reasoning. The goal is that you walk away thinking differently about something you thought you understood.
Practitioner Playbook. Instructional content where the primary value is the method. You should be able to implement or apply something after reading. No filler.
If you found your way here, I would genuinely like to know what you are navigating right now. What is the question your organisation has not answered well yet? What feels most uncertain? Reply to this email or leave a comment below.
The best thinking in this newsletter will come from the people reading it.
See you next week.
Abhinav