31/08/2026
Agile teams learned to treat quality, security, and accessibility as everyone's job, built into the way we work rather than bolted on at the end. Sustainability has not made that jump yet, and AI is making the gap expensive.
Every model call is a cost centre with a carbon shadow. Yet in most backlogs, "should we even use AI here?" is never a story, energy is never an acceptance criterion, and nobody in the room owns the footprint of what the team is shipping.
In this episode of AgileWoW, Dr. Niladri Choudhuri — CEO of Xellentro Consulting Services LLP and President of the Green Computing Foundation — brings sustainable AI into language the Agile community already speaks:
Sustainability as a non-functional requirement. Where it belongs in the Definition of Done, and what a green acceptance criterion actually looks like for an AI-enabled feature.
The Product Owner's newest trade-off. Value, cost, risk, and now footprint.
Why "do we need a large model for this?" is a product question before it is an engineering one, and how to make that call without stalling delivery.
Waste, in the Lean sense. Over-processing is a classic waste. An oversized model doing a job a simpler approach would handle is over-processing with an electricity bill. Muda has a new form, and it is fashionable.
Making it visible. Agile runs on feedback loops and visible information. What teams can measure across training, fine-tuning, and inference — and why inference, the part nobody watches, usually dominates over a product's life.
Who owns it? Sustainability fails the same way quality failed before Agile: as someone else's department. What shared ownership looks like across the team, the platform, and the enterprise.
The uncomfortable question for leaders. AI adoption targets are being set top-down while sustainability commitments are made in parallel, by different people, in different documents. How leaders reconcile the two before the contradiction becomes a headline.
AI for sustainability vs. sustainability of AI. Both are real. Conflating them is how organisations claim credit while avoiding accountability.
Practitioners leave with concrete additions to their Definition of Done and refinement conversations. Leaders leave with a sharper view of a risk their AI strategy has probably not priced in.
Watch more FREE learning resources - https://www.youtube.com/?sub_confirmation=1
Join the WhatsApp group for regular updates - https://chat.whatsapp.com/Gc5yep9PLnT1SLoAwzZusJ
Visit our websites for FREE learning resources -
https://aidevdayindia.org/
https://agileleadershipdayindia.org/
https://scrumdayindia.org/
https://productleadersdayindia.org/
� New to streaming or looking to level up? Check out StreamYard and get ₹740 discount! �
