Rung 3 · ₹10–20L
Capability Cohort
Ten to twelve weeks, 20–30 people, a capstone built on your own workflows. Never a first engagement.
The Capability Cohort is a 10–12 week programme for 20–30 people that builds the ability to design, build and evaluate agentic workflows, with a capstone on your own workflows. It teaches orchestration, not prompt basics, and is never sold as a first engagement. It follows a Diagnostic and a Sprint.
- 10–12 weeks, 20–30 people, capstone built on your own workflows.
- Teaches agent orchestration, tool use and evaluation, not Prompt Engineering 101.
- Never a first engagement; it follows a Diagnostic and at least one Sprint.
- Capability is measured by what people build, not by completion certificates.
- Priced ₹10–20L per cohort, indicative; the next rung is the Governance & CoE retainer.
- Rung
- 3
- Duration
- 10–12 weeks
- Price
- ₹10–20L
- Shape
- 10–12 weeks · 20–30 people · capstone built on your own workflows
Indicative. Priced per engagement.
Who this is for
A transformation owner at a GCC who has already run at least one Workflow Sprint, has seen agentic delivery work in their own environment, and now wants to build the capability across 20–30 of their engineers so the organisation can run workflows itself. This rung scales capability after delivery is proven, not before.
What you get
- 20–30 of your people able to design, build and evaluate agentic workflows, not just prompt.
- A capstone built on your own real workflows during the programme: working artefacts, not exercises.
- Internal capability to run the next workflow without an outside vendor in the room.
- A shared evaluation and governance discipline carried across the cohort, not held by one person.
- A pipeline of candidate workflows surfaced by the cohort as they learn to see them.
Deliverables
The cohort capstones
Working agentic artefacts built on your own workflows during the programme, owned by your teams and reviewed against the same evaluation discipline we use.
A capability baseline and exit assessment
What the cohort could do at the start and at the end, measured against building and evaluating workflows, not attendance or completion certificates.
Reusable internal materials
The patterns, harness templates and review practices the cohort used, left with you so the next group can be run internally.
A workflow-candidate backlog
The workflows the cohort identified as they learned to spot them, ranked for future sprints.
What this rung will not do
We name what we refuse to sell before we start. The honesty constraint is the product.
- Not a first engagement. We do not sell a cohort as an opening move. It follows a Diagnostic and at least one Sprint.
- No completion certificates as the measure. Capability is measured by what people can build, not by attendance.
- No Prompt Engineering 101 syllabus. The free tier covers prompt basics; this cohort teaches orchestration, tool use and evaluation.
- No catalogue course. If it could be delivered asynchronously to a million people, it is already free and we do not resell it.
- No logo-for-discount arrangement. The price is not traded for a case-study credit or a brand mention.
Frequently asked questions
A 10–12 week programme in which 20–30 of your people learn to design, build and evaluate agentic workflows, with a capstone built on your own real workflows rather than toy exercises. It teaches orchestration, tool use and evaluation (the parts the free training tier does not cover), and leaves your organisation able to run the next workflow without an outside vendor in the room.
Because training a large group to build agentic workflows before anyone has seen one delivered in your environment produces confident people building the wrong things. The ladder is sold in order: a Diagnostic to decide what to build, at least one Sprint to prove delivery works here, then a cohort to scale the capability. Selling the cohort first is the mistake that makes training the thing that fails, and that refusal is the point of the rung.
A catalogue course is asynchronous, generic and, for prompt basics, already free from Microsoft, AWS, IBM and Nasscom FutureSkills. The cohort is the opposite: live, built on your own workflows, focused on orchestration and evaluation, and measured by what people can build at the end. If it could be delivered to a million people asynchronously, we would not sell it, because it is already free.
By capability, not attendance. We baseline what the group can build and evaluate at the start and assess it again at the end against the same standard, and the capstones are working artefacts reviewed with the same evaluation discipline we apply to a Sprint. Completion certificates are not the measure: the measure is whether your people can now build and score a workflow themselves.
Indicatively ₹10–20L per cohort, quoted per engagement against group size and the workflows in scope. It is billed 50% upfront and milestone-billed, like every engagement. The price reflects the programme, and it is not discounted in exchange for using your name as a reference.
Not sure which rung fits?
Start with a free AI Reality Check. Ninety minutes, on-site, one real workflow built live.