💬 Ask anything about this report
Thinking...
Technology Team · Annual Report

A Year of
Turning Time Into
Magic

Across three areas — IT Infrastructure, Digital Transformation, and AI at Teach For India — the Technology Team built the systems, tools, and capabilities that gave our people more time for what matters most.

770+
Hrs Saved
via AI
91%
IT Issues
within SLA
24
AI Projects
Proposed
97,756
Records
Automated
🖥️
3,280
Devices in active inventory
🤖
$20K
Anthropic AI credits secured
🎓
1,200+
Students surveyed, 8 cities
✅
7 Live
AI projects in production
1
IT Infrastructure

Keeping the Foundations Strong

The behind-the-scenes work that keeps every laptop, device, and data point running reliably

🎫
Helpdesk Support
91%
Of 800 issues resolved within SLA this year.
1.92 hrs
Avg TAT
vs 2-day target
1.5 hrs
Avg First Response
vs 8-hr target
💻
Device Inventory & Management
95.4%
Of devices in inventory are fully functional. The Blended Learning device pool was right-sized from 4,000+ to 2,200 — keeping only what's actively needed.
1,091
Laptops
2,189
BL Devices
3,280
Total Active
🎣
Cybersecurity: Mock Phishing Attack
15%
Of Staff clicked on suspicious links in a mock phishing exercise conducted across the entire Staff base — highlighting a real and present security risk we need to address together.
⚠️
Why this matters: Phishing is the #1 entry point for data breaches. Awareness training is now being prioritised.
📁
Google Drive Audit
>3 TB
Of organisational data sits in individual "My Drive" folders — not in shared drives. This creates real risk: data becomes undiscoverable, and gets lost when people leave the organisation.
💡
Next step: Migrate individual My Drive data to shared drives with proper folder structure and ownership.
✦
Experiments
🧪
In progress · Claude Code
Self-Service IT Tools for Staff & Fellows
We're experimenting with Claude Code to build lightweight tools that let Staff and Fellows self-audit their devices and run Level 1 troubleshooting on their own — before reaching out to our engineers for support. The goal is to put simple diagnostics in the hands of every person in the organisation, reduce wait times, and free up our team for the issues that genuinely need expert attention.
🔍
Device Self-Audit
Check device health independently
🛠️
Level 1 Troubleshooting
Resolve common issues without tickets
⚡
Faster Resolution
Engineers focus on complex problems
2
Digital Transformation

Technology That Did the Work

Six projects that saved time, improved quality, and scaled impact across the organisation

🎯
AI in Fellowship Selection
770+ hrs saved 4,753 apps reviewed
📹 Video Purpose Essays added to the application — letting the Selection team assess verbal communication and grammar early in the funnel, before Assessment Centre stage.
🎙️ AI Transcription at 90–95% accuracy means Selectors read a structured transcript instead of watching each video in full — saving ~1 minute per application, ~330 person hours in total.
✍️ Automated Written Communication Scoring via OpenAI + Salesforce reviewed 4,753 applications at 75% accuracy — saving ~3 minutes per reviewer per application, ~230 person hours in total.
⏱️ Automated Late AC Signup & Reschedule approvals eliminated 10–15 minutes of manual handling per applicant across 140 ACs, saving ~100 person hours.
🔒 Enhanced Proctoring detects whether the same applicant is giving the test and whether they're doing it alone.
🤖
Community Chatbots
4 bots deployed 500+ users
👥 Alumni Discovery Bot: 220 Fellowship Alumni and 38 Y2 Fellows used it to find each other by location, sector, and designation — self-serve, no team involved.
🎓 Student Alumni Bot: 245 students engaged, 80 updated their data, 273 opportunity searches made.
🏢 Career Fair Portal Bot: Answered all FAQs for Fellows and Entrepreneurs automatically on the Career Fair portal.
🙋 Volunteer Assistant Bot: 51 Fellows accepted or rejected volunteer applicants at a click, with automated onboarding reminders sent to volunteers.
📊
Student Holistic Survey (SHS) Pilot
1,200+ students 90%+ completion
🎨 Child-friendly UI with multiple modes of data upload made the survey intuitive and accessible across age groups and all 8 cities.
🎯 Exceeded the 1,000-response target — reaching 1,200+ students with a 90%+ completion rate maintained across every cycle.
📈 Analytics on our SVS give Fellows and PMs real-time insights into student growth — at the student, classroom, LC, and city level.
✅ Zero technical issues reported by Fellows during the entire rollout across 8 cities.
⚡
Smart Automations
100+ hrs saved ~15 person days
📧 Automated Campaign Emails & WhatsApp for 400+ events — saving ~67 person hours for the Recruitment team over the year.
📋 Zoom–Salesforce Attendance Capture for 600+ online events — ~100 person hours saved, with no manual tracking during live sessions.
🔗 Alumni LinkedIn Profile Scraping via UIPath: 4,000+ records updated — saving ~15 person days of manual effort.
⏰ Automated AC Reminder Emails to Selectors — saving 20–25 minutes per week and eliminating potential human error.
🏛️
HRMS Migration & Automations
1,100+ profiles 97,756 records
🔄 Migrated from Keka to Darwinbox with full Salesforce integration — 400+ Fellow profiles auto-created during the 2025 Institute and 700+ Fellow profiles auto-created during the 2026 Institute — saving ~34 person hours with zero manual action.
🗂️ 97,756 Keka receipts exported automatically for backup. Doing this manually at 2 minutes each would have taken a person 1.5 person years working 8 hours a day.
🌐 BTCP Portal Revamp: Updated the Beyond the Classroom Programme portal for improved usability and data capture.
🖥️ Teach For India Website Revamp: Refreshed the public-facing website for better navigation and storytelling.
🚀
AI in Movement Building
800+ essays ~67 hrs saved
📝 KER Essay AI: 800+ KER applicant essays evaluated with AI-generated scores and rationale. Each evaluation took ~5 minutes manually — saving ~67 person hours in one cycle.
🏫 TFIx AI Site Visit Assistant: An AI agent trained on site context and the TFIx rubric. Helped Selectors prepare the right questions before visiting a school — reducing ramp-up time and improving evaluation consistency.
3
AI at TFI

Building Our AI Capability — Inside and Out

From literacy sessions to live partnerships — how TFI is thinking about and acting on AI

A
AI Literacy
🏫
Sessions Across Teach For India
AI awareness and hands-on sessions were run with 3 Teach For India clusters, Firki Teachers, and TFIx Fellow Alums — covering AI fundamentals, practical tools, responsible use, and real examples from within the organisation.
🎓
AI Depth Track — 8 Sessions with SPOCs
An 8-session deep-dive programme run with SPOCs across functions — moving from concepts to hands-on project design. Each SPOC left with an AI project scoped for their function.
24 projects proposed · 7 in production · 1 pilot · 11 in ideation
B
Partnerships
Anthropic
A 90% discount secured — giving every Fellow and Staff member access to Enterprise AI
Last year, Anthropic visited our classrooms and offices in Delhi, engaging with the purpose and philosophy of our movement. We received 15 Claude Enterprise licenses and $20K API credits pro-bono to evaluate Anthropic's tools for 7 months. We piloted these across 15 verticals successfully — and are getting Claude Enterprise licenses for all Staff and Y2 Fellows at a 90% discount.
Salesforce · Agentic AI Accelerator
Building an AI Teaching Assistant for Firki users
Selected for Salesforce's Agentic AI Accelerator — Salesforce is helping us build an AI-powered Teaching Assistant for Fellows and educators on the Firki platform. This moves us from passive content delivery to active, personalised learning support at scale.
Polymath
Pushing our Sense of Possibility with AI
Polymath ran the AI Depth Track with our SPOCs across functions, and pushed the boundaries of what the Technology Team can build with AI. A session with the Polymath team pushed our Sense of Possibility with the use of Claude Code for end-to-end feature development — showing us what was possible when we stopped thinking about AI as a chatbot and started thinking about it as a builder.
Gooey.AI Accelerator · Rockefeller Foundation
Key learning: off-the-shelf platforms have ceilings
Good exposure to building a Fellow-facing chatbot grounded in our knowledge. Ultimately limited to chatbot functionality — could not leverage the agentic capabilities of tools like Claude. Significantly more expensive beyond the accelerator period.
C
Outside-In Learnings

At the India AI Summit, three organisations stood out for deploying focused intelligence where it matters most — and their learnings are directly applicable to Teach For India.

Cash Transfers
GiveDirectly
AI-powered disaster forecasting enables emergency cash delivery by analysing real-time mobility data — a timely alert paired with instant, no-strings-attached transfers.
Legal Access
Adalat AI
AI legal assistance for district court judges and clerks — clerical support only, intentionally never influencing decisions. Blockchain with private keys on judges's devices.
Early Learning
Rocket Learning
WhatsApp-based programme for Anganwadi workers and parents. AI for personalisation. Meets families where they already are.
1.5M users · 300K Anganwadi workers trained · 80% picture recognition vs 50% national avg
"We're past the age of mass instruction. We learn best when taught individually — we'll move into an age of highly personalised learning for each child, where they learn more and enjoy it much more."
Reed Hastings, Co-Founder, Netflix · Shared at a Teach For All network session

Ben Gomes, Head of Learning at Google: While AI can customise data and organise lessons, it lacks the ability to solve the human challenge of student motivation. The teacher remains irreplaceable.

Anthropic's own research on AI use in software engineering investigates how it affects long-term skill formation — balancing immediate productivity with ensuring humans stay capable of overseeing AI output. This directly shapes our approach to AI literacy.

Sharpest lesson for non-profits — India AI Summit
  • Do not build end-to-end AI solutions. Targeted, focused intelligence outperforms grand platforms.
  • Build optionality into your architecture. Your building blocks should be swappable as the technology improves.
  • Focus on foundations — the data layer and curated organisational knowledge are the highest-value investments.
  • Stay nimble. The organisations winning are those who prepared their foundations best, not those who built the most.
D
Way Forward

Two major directions for 2026–27: making AI accessible to every Staff member and Fellow, and building the knowledge foundation that makes our AI uniquely powerful.

Claude Enterprise — Organisation-Wide Rollout
Giving every Staff member and Y2 Fellow a capable AI tool — at a 90% discount
Given the Anthropic partnership and 90% discount secured, we plan to roll out Claude Enterprise across the organisation. Every Fellow and Staff member gets access to AI for planning, lesson design, reflection, and knowledge work. This will be paired with a structured AI literacy programme — not just tool access.
Ed Equity Hub

The Ed Equity Hub connects Claude to Teach For India's actual institutional knowledge — 15 years of documents, org data, and the Alumni graph mapping our network across cities and cohorts. A knowledge commons for our entire movement, built in partnership with McKinsey. We have also been selected for the AgentForce Accelerator (Salesforce Foundation) for the public-facing WhatsApp chatbot that connects to this knowledge layer.

Ed Equity Hub — Architecture Overview
Tier 1
Claude Enterprise
External Chatbot
Public MCP Connector
↓
Gateway
MCP Server on Google Cloud Run
↓
Data
BigQuery (Org Data)
Vector Store (Org Knowledge)
Spanner Graph (Alumni Network)
4
Our Team

The People Behind the Magic

A team that builds, reflects, and grows together

📸
The Team
The Technology Team
⭐
Team EFR Scores
3-year consistent track record
EFR Score
4.4
Range 4.4–4.5 for 3 years
Team NPS
67
Range 60–80, 0 detractors · 3 years
🚀
IJP Movements
This year, 3 team members grew into new roles — 2 stepping up vertically, 1 moving laterally to broaden their impact across Teach For India.
🏫
Nagpur Stepback · February 2026

In February 2026, the team came together in Nagpur to reflect on the year — a lot of what you see in this report came from those reflections. We spent time with four organisations that deepened our thinking about purpose, technology, learning, and impact.

Polymath session
💡
Polymath
A session that pushed our Sense of Possibility with the use of Claude Code for end-to-end feature development.
Learning Companion visit
📖
Learning Companion (TFIx)
Focus on the FOUNDATIONAL skills
Just like Learning Companion focuses on foundational skills to enable children to learn independently — taking the burden of teaching everything away from Fellows.
If we focus on:
  1. Research and planning
  2. Context engineering
It enables Claude Code to build the entire solution on its own — removing the burden to code, test and debug from us.
Slum Soccer visit
⚽
Slum Soccer (TFIx)
Constraints can lead to CREATIVITY.
Just like children who could not hear were able to use their other senses to the fullest — if we have limited AI tokens, we can:
  1. Define our requirements thoroughly
  2. Research & plan well
  3. Write our context completely — then let AI do the coding in one shot
Enabling us to ship thousands of lines of code without writing a single line ourselves.
🏛️
Akanksha Foundation
Visited classrooms from Pre-KG to Grade 4 — a window into the daily experience of Fellows and the students they serve.
View reflections ↓
A
Appendix
Teach For India — Technology Team
Gopalraoji Ramji Motghare
NMC English Medium School
Grade 4 Classroom Visits
February 26, 2026

What We Saw at Akanksha

Round tables on green mat
🖼️
Round tables on green mat — no rows, only conversation
No one is too small to make a difference
🖼️
"No one is too small to make a difference"
Imagine how far you can go
🖼️
"Imagine how far you can go…"
"Great teachers guide you to look — but don't tell you what to see."
— Navoneel, on the teaching he observed
Recurring Themes Across All Observers
Calm Classroom Culture Creativity & Expression Collaboration at Every Age Patience as Pedagogy Student Ownership Embracing Imperfection
🏫
The Environment
What We Walked Into
  • –Walls covered in student work — imperfect and celebrated
  • –Creativity across the school: murals, decorated doors, corridors
  • –Every classroom had a different seating arrangement — no rows
  • –Responsibility charts in office: student drop/pickup, food testing
  • –Learning materials: abacuses, beads, blocks, number tiles — not tablets
  • –Shoe racks at classroom doors; stars drawn on hands as appreciation
🧑‍🏫
The Teachers
What Struck Us Most
  • –Not a single teacher raised their voice — even with 30–40 kids
  • –Calmly redirected attention — never accused students of talking
  • –Paused mid-lesson to ask word meanings — Hindi explanation welcomed
  • –Students who struggled weren't singled out; teacher looped back gracefully
  • –Structures and preparation likely enabling calm and consistency
👧
The Students
What They Were Doing
  • –Small round tables/shifted benches, group sitting enabling collaboration
  • –When a classmate made a reading error, no one laughed — they kindly corrected
  • –Walked between classes with hands behind their backs — self-regulated
  • –Family tree: students drew money plants, bushes — not just literal trees
📚
The Teaching
Pedagogical Moments
  • –Maths division taught pictorially — repeated subtraction, division chart
  • –Not all students at same pace — teacher adjusted without bias
  • –Reflective questions: "Are family members who share food and money alike?"
  • –Class 4 was learning "cartographer" (map makers). I learnt something new too
  • –Teacher modelled multiple approaches to one problem — process over answer
  • –Students wrote on board, counted attendance — ownership and involvement in operational tasks
A tray of abacuses
🖼️
A tray of abacuses — simple tools, real learning
Simple tools.
No tablets needed.
One of the first things that struck our team was what wasn't there. No screens, no apps — just wooden abacuses, counting beads, number placement tiles, and a whiteboard. The structured simplicity matched the structured simplicity of the teaching. As Hitesh noted: "Inspired by simple use of tools — learning beads & blocks. Even the whiteboard was being used as a very effective tool with the structure they followed"
The walls speak
before the teacher does.
Every corridor carried a message. From "No one is too small to make a difference" to student-made anti-corruption posters to a butterfly mural declaring "Imagine how far you can go" — the school's values weren't in a mission statement. They were on every surface children walk past every day. Shanya summed it up: "You be you, we'll learn through it. Visual learning from every wall."
Student-made poster: Vigilant India
🖼️
Student-made poster: "Vigilant India — Say No to Corruption"
Voices from Our Team
Sfoorti
"The family tree — I would've just drawn a tree. These kids drew money plants and bushes. That's creativity no one forced."
Prerna
"In my school, whoever read well got to read more. This teacher gave every child a turn — enthusiastic or shy. She wasn't biased."
Dimple
"We were never allowed to move chairs around in school. Here, every room was arranged for conversation."
Shanya
"Imperfect — but they embrace it. 'You be you, we'll learn through it.' You can see it on every wall."
Akshat
"We came with typical expectations of a govt school. What we saw: creativity, commitment, patience, compassion — the teacher greeted all 40 of us."
Hitesh
"The calmness was structural. Centrally given lesson plans, responsibility charts. Systems created the peace."
Nitesh
"A grade 1 student had the courage to point out the teacher's mistake. That's the kind of classroom safety we all want."
Vinayak
"Every class was engaging. Student work embraced throughout — even with mistakes. Kudos to these teachers."
Manorma
"The staff felt like a small family. All our observations came back to the same four words: creativity, commitment, patience, compassion."
What We're Taking Back
Patience as Practice
Multiple observers named patience as their single biggest takeaway — the way teachers taught, not just what they taught.
Structure Enables Freedom
The calm wasn't magic — it came from systems: centrally designed lesson plans, co-teaching, clear routines. Structure gave teachers permission to be patient.
Embrace the Imperfect
Mistakes on walls, Hindi mid-English lesson, slow readers given equal turns — imperfection was the learning, not an obstacle to it.
Teach For India · Technology Team
School Visit Reflections · February 26, 2026
Navoneel · Shanya · Dimple · Amit · Sfoorti · Prerna · Manorma · Naveen · Vinayak · Akshat · Nitesh · Hitesh