They do not yet know what to do with AI inside it. Zdopt runs practitioner-led AI programmes on campus — faculty first, then students, delivered by the people who build with these tools in production.
Thirty faculty are thirty people who go on teaching thousands of students. A student batch is one batch. That is the whole argument for the order we run in.
Students are judged on what they build
Recruiters have stopped asking whether a candidate has heard of AI. They ask what the candidate has actually shipped with it.
Faculty are asked to teach it anyway
Most AI training for colleges demonstrates the tools rather than using them, and nothing survives past the week it ran.
Nobody checks day thirty
A form signed at the end of day two tells you the session happened. It does not tell you the department kept using any of it.
Most colleges start at L0. It is free, it is once per campus, and it tells you more about whether this is worth doing than any brochure will.
Up to 150
Live build demo
We take one real problem from your department and build the solution live, start to finish, with the room watching. No slides, no cost, no commitment. This is how most colleges start with us.
A working artefact built from your own department’s problem.Up to 30 faculty
Faculty Development Programme
Faculty use AI inside their own subject rather than describe it. A department that can teach it outlasts a batch that attended it.
Three working prompts built around their own subject, not sample prompts.Up to 60 students
AI Practitioner
Students put AI into their own daily academic work rather than describe it in an interview. For arts and science batches no programming background is assumed.
Reusable projects, skills and connectors the student keeps and extends after the programme ends.Up to 40 students
AI Builder
Weekly labs, each producing something that runs, plus one final build. The line we hold is that a working prototype is not the same thing as production grade.
A multi-agent workflow with at least one MCP integration and a working RAG pipeline, deployable and production grade.Programmes are priced per batch and exclude applicable taxes. Travel and stay outside Chennai are quoted separately. You are paying for the agreed cap, not for attendance. Final pricing is confirmed in a written proposal after a scoping conversation.
Weekly labs at L2 and L3, each producing something that runs rather than something that was demonstrated.
An honest read on where your faculty and students actually are, before anybody signs anything.
- One real problem from your department, chosen with you beforehand
- The full build, start to finish, with nothing pre-prepared offscreen
- What the tools can and cannot do, shown rather than claimed
- Where this fits in your syllabus and your placement calendar
- Open questions from the floor, faculty and students together
Faculty leave able to write a prompt that returns a usable first draft rather than a generic one, judge what may and may not be pasted into a tool, set up one project for their own recurring work, and recognise a fabricated answer.
- What changed and why now, pitched at a room that includes sceptics
- Prompt engineering fundamentals
- Safe and compliant use: what you share and what you never share
- Tokenomics and model selection
- Working with documents and files
- Extended thinking and research workflows
- Projects, and setting up their first three
Claude and Microsoft Copilot are covered side by side, so the programme works on whichever stack the college already runs.
Students finish able to run their own research, drafting and document work through AI, and to say what they may and may not paste into a tool.
- Prompt engineering fundamentals
- Safe and compliant use, and what never gets pasted into a tool
- Model selection and token economics
- Document and file workflows, document intelligence
- Extended research and market research workflows
- Chat layer end to end, projects at second-brain level
- Skills, connectors and no-code agents in Copilot
Taught on Microsoft Copilot, the stack most colleges already hold through M365. Runs entirely on a free tier.
The final build is designed to stand as the student’s mini project or final year project submission.
- What generative AI is and how agentic AI differs
- Tools and protocols: MCP, A2A, skills, plugins
- RAG and context management
- Traceability, watermarks, GDPR and compliance
- Observability and guardrails
- Setting up an AI IDE
- Choosing a model: capability, tokenomics, reading model cards
- Embedded document intelligence, and code to documentation
Taught on Claude. The agentic stack this level is built on has no production-grade equivalent in Copilot. Entry is gated by a short screener so the room stays at one level.
Demo
L0 first. Ninety minutes, one real problem from your department, built live in front of the room.
Faculty
L1 runs across two consecutive days, before any student batch is scheduled.
Students
L2 and L3 run as an intensive block or across weekends, whichever fits the semester.
Measure
We check who is still using the tools thirty days after the session, not on the last day.
Student windows and faculty windows are counter-cyclical — students in Aug–Oct and Jan–Mar, faculty in Dec and May–Jun. That is not a scheduling trick, it is why a department can run something with us in most months instead of fighting for one crowded term.
Even semester opens. L2 and L3 run.
- Student batchesL2 and L3
- Faculty developmentL1, counter-cyclical
- ExamsNothing delivers
- ScopingPlan the year
Running an external programme should not create work for your department. Every programme ships with the documentation your accreditation file needs, as standard.
- Digitised attendance for the full batch
- CO-PO mapping in NBA format
- Co-branded certificates with a verifiable ID, checkable on zdopt.com
- Feedback summary and a day-thirty usage check
- Photo and report pack for your NAAC file
- A named coordinator from the department for the run-up and the day itself.
- One real problem from the department, so the live build is yours rather than ours.
- Confirmed batch list at least one week ahead.
- A room with a projector and reliable power, seating the agreed batch size.
- Wifi that holds the full batch online at once. This is the single commonest cause of a session underdelivering.
- Participants on their own laptops wherever possible.
Some of what we teach at L1 and above needs a paid AI plan rather than a free account. We will tell you exactly which parts, and what it costs, before you commit to anything. We would rather raise it now than have half the agenda become a demonstration on the day.
Ninety minutes, one real problem from your department, built live in front of the room. No cost, no commitment, and it tells you more about whether this is worth doing than any brochure will.
Pricing is confirmed in writing after a scoping conversation. Nothing quoted on this page is a commitment until then.