The Fifth Elephant 2026 Winter Edition

Call for Data Problems and Proposals

Kruthika Kumar Muralidharan

Let AI Do the Work. Keep the Analyst in Charge.

Submitted Oct 11, 2026

Submission type: Showcase - new ideas - 20 mins

An analyst asks why revenue fell. An AI assistant can find the data, write SQL, draw charts, and produce a convincing explanation in minutes. But what if it uses the wrong definition of revenue, compares the wrong periods, or mistakes incomplete data for a real decline? Checking only the final answer is too late. But asking the analyst to approve every step defeats much of the point of using AI. I’ve been building Neri, an open-source tool that takes a different approach. It helps analysts do their work with AI, while keeping them in charge of the choices that matter.

In this showcase, I’ll take one business question through Neri, from the first question to a report. I’ll show how the analyst can guide the query, check the data, challenge an assumption, and correct the analysis before drawing a conclusion. The goal is to let AI handle more of the work without making the analyst a passive observer or forcing them to check everything. Neri is still under development, and I plan to demonstrate a working part of the product, including what it cannot yet do. The question I’m trying to answer is simple: can AI make analysts much more productive without taking away their judgment?

Takeaways

  1. Know when to step in. The definition of a metric, the completeness of the data, and the choice of comparison can all change an answer. I’ll show how Neri brings these choices to the analyst’s attention while AI does much of the work.

  2. Being in control shouldn’t mean approving everything. Once an analyst has made a choice, the tool should be able to reuse it and only ask again when something important changes. I’ll explain how I’m approaching this in Neri, what works, and what still needs work.

Which audiences will benefit most?
Analysts and data scientists who want to use AI in their daily work, analytics leaders thinking about how their teams should work with AI, and people building AI-powered analytics tools. The showcase focuses on everyday analytical work, not model architecture. No AI engineering background is needed.

Bio
Kruthika Kumar has spent 14 years working in data and analytics, most recently as Director of Analytics at Razorpay. His work has covered product/business analytics, experimentation, data platforms, and AI-assisted analytics. He is now working on multiple projects, including building Neri, an open-source Analyst Control Centre that helps analysts work with AI while keeping control of their methods and decisions, and Chowk, the fictional marketplace used to test it.

Links
Not yet ready.

What I don’t know yet / what I want help with

  1. How much control is enough? Too little, and analysts may trust answers they shouldn’t. Too much, and the tool becomes slower and more frustrating than doing the work yourself. I’m testing when AI should act on its own and when it should ask. I’d welcome feedback from analysts on which interruptions are useful and which just get in the way.

  2. Will this work on real company data? A synthetic dataset lets me test specific mistakes, but real companies have inconsistent definitions, missing context, and different ways of working. I’d like to test Neri on real analytical workflows and learn where it helps, where it slows people down, and what it misses.

Matchmaking tags
I can help with: ideas, experience, technique, critique
I need help with: experience, critique, collaborator
I’d like to meet: practitioners, domain experts, tool builders

Topic/domain tags
AI-assisted analytics, business analytics, human-AI collaboration

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

Jumpstart better data engineering and AI futures