None of them is the model. Zdopt runs practitioner-led AI adoption programmes for enterprise, global capability centre and mid-market teams. Sponsors first, then the functions that have to use it.
We have not yet met a stalled rollout whose root cause was the model. It is almost always one of these three, and each one is closed in a room rather than in a tool.
No sponsor
Nobody senior owns the licence decision, the rollout order, or the measurement. Training without a sponsor does not survive contact with a budget cycle.
No licence
Half the capability taught needs a paid or enterprise tier. On free accounts the agenda quietly becomes a demonstration.
No measurement
Nobody agrees what success looks like before the room opens, so nothing gets checked after it closes. Week five arrives and there is no baseline to compare against.
Leadership runs first. It is the room where licence tier, first functions and measurement get decided — and a programme that skips it does not survive contact with a budget cycle.
Up to 15
Leadership
For the people who approve the budget, with zero hands-on expectation. This is the room where licence tier, first functions and measurement get decided.
A one-page adoption plan they own, not one we wrote for them.Up to 30
Fluency
For mixed-function rooms with little or no prior use, including the people who are sceptical about it.
Three working prompts for their own actual job, not sample prompts.Up to 25
Function Practitioner
A single function goes deep on the platform you already run. We scope against your instance during discovery, so the two days are spent on your workflows rather than generic examples.
Reusable projects, skills and connectors the function keeps and extends after we leave.Up to 25
Practitioner, your stack
The Practitioner track rebuilt around the platform your teams already live in, so the two days are spent inside your own workflows rather than a generic environment. Scoped against your setup during discovery.
Accelerators that run against your own systems, not a sandbox.Up to 20
AI Builder
For engineering teams. The line we hold is that a working prototype is not the same thing as production grade, and the four days are spent on the difference.
A deployable artefact, not a demo that only runs on the trainer’s laptop.Programmes are priced per cohort and exclude applicable taxes. Travel and stay outside Chennai are quoted separately. Final pricing is confirmed in a written proposal after a scoping call.
The same routing we would do on a scoping call. It will tell you to start with Leadership if that is the honest answer, even though it is the smallest thing we sell.
Does someone senior already own the licence decision and the rollout order?
Every track ends with a live build on one of your real workflows, not a worked example from a deck.
Sponsors leave able to decide a licence tier against actual seat maths, name the first three functions to roll out, approve a data-sharing policy, and sign off on how adoption will be measured.
- What actually changed in the last twelve months, and what did not
- Licence tier economics: what free, paid and enterprise buy, with seat maths
- The three adoption failure modes and how to close each one
- Data boundaries: what leaves the building and what does not
- Build versus buy versus adopt, as a decision framework
- How to measure: baseline, day N, day thirty
- A forty-five minute live build on one of your real workflows
Participants leave able to write a prompt that returns a usable first draft, judge what they may and may not paste in without asking anyone, 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 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
The function finishes with accelerators built for its own recurring work, on the stack it already uses.
- Prompt engineering, safe use and compliance in depth
- Tokenomics and model selection against real workloads
- Document and file operations, document intelligence
- Market research and extended thinking for management and finance functions
- Chat layer end to end, projects at second-brain level, skills, connectors and plugins
- No-code agents in Copilot
- Augmented versus embedded: where AI sits relative to your systems
Scoped during discovery, so the room works on your own configuration rather than a demo one.
- Everything in the Function Practitioner track, rebuilt on your platform
- Discovery against your own environment before the session date
- Where AI sits relative to your systems: augmented versus embedded
- What is worth automating in your workflows, and what is not
Teams ship a multi-agent workflow with at least one MCP integration and a working RAG pipeline, deployable rather than demonstrable.
- 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
Scope
Thirty minutes on what you have already tried, which functions are asking for this, and where the licences sit.
Sponsor
Leadership runs first. Licence tier, first functions and measurement get decided in that room.
Deliver
Hands on keyboard, not slides. The live build in every track uses one of your real workflows.
Measure
Baseline before, usage checked at day thirty. The target is sixty percent still using it.
By the end of a Practitioner or Builder cohort we have spent days inside your actual workflows. We have seen where the friction sits, which steps are worth automating, and which are not. That is the point at which an implementation conversation stops being speculative.
Workflow automation
Scoped build of the processes surfaced during the cohort, taken to production rather than left as a prototype.
Advisory retainer
Ongoing review as your licence position, model choices and internal capability change.
We will tell you what is worth building, and we will tell you when the answer is nothing. Consulting is a separate engagement, scoped and quoted separately — nothing in the training is contingent on it.
Short list, but we mean all of it. Each item on it is something we have watched a session underdeliver without.
- A named executive sponsor for the Leadership session. This is the one dependency we will not scope around.
- Two or three real workflows shared ahead of time, so the build is yours rather than ours.
- Confirmed attendee list one week ahead.
- A room seating the agreed cap, with a projector and network access for every participant.
- Participants on their own machines, with licences confirmed before the session date.
- Network that holds the full cohort online at once. This is the commonest cause of a session underdelivering.
Several capabilities we teach need a paid or enterprise tier rather than a free account. We will tell you exactly which parts and what they cost during scoping. We would rather raise it early than have half the agenda become a demonstration on the day.
Thirty minutes on what you have already tried, which functions are asking for this, and where the licences currently sit. We will tell you which track fits and which one does not.
Pricing is confirmed in writing after the scoping call. Nothing quoted on this page is a commitment until then.