On 4th November, Akhilesh and I visited Waste Warriors. They’ve been working on waste management in the Himalayas for 13 years, mostly in Uttarakhand and Himachal Pradesh.
We went there to understand how their teams work on the ground, and whether tech or AI could take some load off their daily work.
We spent about half an hour with Mayank Sharma, their CEO, who shared how they think about tech in an organisation where most of the work happens on the ground. Most of our time, though, was with Angshuman, who walked us through their programs and how things run across locations, and Anjali, who showed us the applications her team has built and how they’re cutting down the paperwork for field teams.
What they do
Waste Warriors is based in Dehradun and has around 250 people, who they call warriors, across 12 locations including Dharamshala, Manali, Shimla, Uttarkashi and Corbett.
Their work involves collecting dry waste from households and businesses, who are charged around ₹30–50 per pickup. The waste goes to their recovery facilities, gets sorted into 10–12 categories and is sent for recycling. Nothing goes to landfill.
Their long-term plan for every location is to set up a system that works, hand it over to the local community or Gram Panchayat, and move on. Mayank was pretty open about the fact that none of their locations have reached that stage yet.
The Wall of Warriors at Waste Warriors
How they use tech today
In about three years they’ve gone from pen and paper to spreadsheets to proper apps. A small in-house team of four builds these on Zoho Creator: Satya for M&E (Monitoring and Evaluation) data, Saral for procurement and vendor payments, Saksham for training, and Sandesh for vehicle document and insurance reminders. There’s also a “dark spots” app where field teams log illegal dumping sites with a photo and GPS location.
The M&E setup is the part that stood out to me. Before building anything, they sat down and agreed on what each metric actually means across all locations. Something like “waste in” or “days of livelihood” can mean slightly different things in Manali and in Corbett, and if that isn’t fixed first, comparing them doesn’t mean much.
Now teams fill in data every week, it rolls up monthly into BigQuery and shows up on dashboards. Their M&E setup has become something other NGOs now look to them for advice on.
When not to bring in tech
This is the part of the conversation I keep coming back to.
Around 80% of their staff work on the ground, and most of the people they deal with, like waste collectors, households and small shops, aren’t very comfortable with tech. Their teams already know their daily work well. If you suddenly change how they do it, it usually creates confusion instead of making things easier. And if a new system means field staff spend more time filling forms and updating apps, it’s pulling them away from the actual work.
Mayank gave an example. They once brought in a platform for the fundraising team. It worked fine technically, but it didn’t help much day to day, so the team slowly went back to their own way of doing things.
Money matters too. Around 90% of their funding comes from CSR and the remaining 10% from FCRA (foreign contributions). This is also why they’re careful about spending on technology and prefer investing only when there’s a real need.
So their rule is simple. Bring in tech only when there’s a real pain point and the old way can’t keep up, and solve one problem at a time.
Where AI fits in
They’re already trying a few things with AI.
One is a chatbot where someone can ask a question about their data in plain language instead of opening a dashboard. It runs on Gemini Flash over their BigQuery data, with three agents: one writes the SQL, one pulls out insights, and one makes the graph. It does well with simple questions that need a single number, but struggles when you ask it to compare projects. Dalgo has built something similar, so we’re setting up a call with that team.
The fundraising team, led by Suraj, has connected Claude to Zoho Bigin, the CRM they use for fundraising. They had a demo planned for the next day.
They also used Glific for WhatsApp earlier, then moved to Wati about a year ago. Wati’s costs are now adding up, so they’re looking at Glific again. Since that’s what we work on, it’s something we’ll be following up on.
What I took away
Honestly, I went in thinking about what we could build for them. Most of what I came back with was about what not to build yet.
There are probably a lot of NGOs like this, doing serious work on the ground without much tech support. Helping them probably means spending time with their teams first, seeing where they actually get stuck, and fixing that without breaking what already works.
For now we’ve agreed to pick just one area to start with, whether that’s the data chatbot, WhatsApp or volunteer management. They also want to build a proper digital roadmap so future tech decisions have some direction. Looking forward to seeing where it goes.
A group picture with the Waste Warriors team during our visit.
