Adoption
What AI adoption actually costs in India
The market is quote-gated, so buyers cannot benchmark. What actually drives the number, what each engagement type buys, and our real ladder ranges published in full.
Enterprise AI adoption in India is quote-gated (no major vendor publishes prices), so buyers cannot benchmark. The cost drivers are workflow count, integration surface, and whether governance is included. A scoped diagnostic costs a fraction of a failed pilot, and MIT found 95% of pilots return nothing.
- The Indian enterprise AI market is quote-gated; no major vendor publishes prices.
- Cost is driven by workflow count, integration surface, and whether governance is in scope.
- MIT NANDA, 2025: 95% of pilots returned no measurable P&L, against an estimated $30–40B invested.
- Chokmah's ladder: diagnostic ₹1.5–3L, sprint ₹6–12L, cohort ₹10–20L, retainer ₹2–5L/month.
- A scoped diagnostic costs a fraction of a single failed pilot.
Enterprise AI adoption in India is quote-gated (no major vendor publishes prices), so buyers cannot benchmark. The cost drivers are workflow count, integration surface, and whether governance is included. A scoped diagnostic typically costs a fraction of a single failed pilot, and MIT found 95% of pilots return nothing measurable.
Key takeaways
- The Indian enterprise AI market is quote-gated; no major vendor publishes prices.
- Cost scales with workflow count, integration surface and whether governance is included.
- MIT NANDA, 2025: 95% of pilots returned no measurable P&L, against an estimated $30–40B invested.
- Our ladder, in full: diagnostic ₹1.5–3L · sprint ₹6–12L · cohort ₹10–20L · retainer ₹2–5L/month.
Why does nobody publish AI consulting prices in India?
Because the market is quote-gated. No major India enterprise GenAI vendor publishes a rate card, and the reason is commercial: the work is priced on value and reference, not against a public list, which lets vendors price each engagement to what it is worth to that buyer. It is a rational strategy for the seller and a bad deal for the buyer, who walks into every conversation unable to tell whether a number is fair.
We think publishing honest ranges is worth more than the pricing power we give up, precisely because the market is opaque. The buyer who can see a real range before the first call arrives at that call able to reason about scope instead of guessing. So this page ends with our actual ladder, and the ranges are real, not indicative theatre.
What actually drives the number
Three things move the cost of an AI adoption engagement, and none of them is the model.
Workflow count. One workflow is a sprint. Five workflows is a programme. Cost scales roughly with the number of distinct processes you are instrumenting and changing, because each one needs its own baseline, its own evaluation harness, and its own owner.
Integration surface. A workflow that lives inside one system is cheap to change. A workflow that spans a CRM, a document store, an ERP and three approval steps is expensive, because most of the engineering is in the connections, not the intelligence. The integration surface is the single most underestimated cost driver.
Governance inclusion. Building the audit trail, the model register and the escalation path into the engagement costs more up front and far less than retrofitting them at the production gate. Whether governance is in scope changes both the price and the odds the thing ever reaches production.
Diagnostic, sprint, cohort, retainer: what each buys
The engagement types are not interchangeable price points; each buys a different thing, and they are meant to be bought in order.
| Engagement | What it buys | Chokmah range |
|---|---|---|
| Adoption diagnostic | A scored decision: which workflows to automate, which to leave alone, why | ₹1.5–3L |
| Workflow sprint | One workflow built and shipped jointly; the client owns the code | ₹6–12L |
| Capability cohort | 20–30 people trained on a capstone built from their own workflows | ₹10–20L |
| Governance & CoE retainer | Ongoing framework, harness maintenance, champion enablement | ₹2–5L/month |
These are opening ranges, quoted against your actual scope, and they are the same figures that appear on our service pages. The adoption diagnostic comes first for a reason: it is how you find out which workflow the sprint should target before you pay for the sprint.
How to compare two quotes that are not comparable
In a quote-gated market you will often hold two proposals that price different things and call them the same. Normalise them before you compare. Ask each vendor: how many workflows, named; what is the integration surface, listed; is governance in scope, yes or no; who owns the code at the end; and what are the payment milestones tied to. Two quotes that answer those five questions identically are comparable. Two that do not are not, and the cheaper one is frequently cheaper because it scoped less. The full version of this is our list of ten questions to ask any AI vendor.
The cost of the pilot you should not have run
The most expensive line item in Indian enterprise AI is not any consultant's fee. It is the failed pilot. MIT's NANDA study found 95% of enterprise GenAI pilots produced no measurable P&L return, against an estimated $30–40 billion invested across the market (MIT NANDA, July 2025). That is the base rate you are pricing against.
Seen that way, a diagnostic is cheap insurance. It costs a fraction of a single failed sprint and its entire job is to stop you funding the workflow that was never going to work. Gartner names escalating costs as one of three reasons it expects over 40% of agentic AI projects to be cancelled by the end of 2027 (Gartner, 25 June 2025). Costs escalate fastest on the projects that should have been stopped at the diagnostic and were not. The diagnostic is how you avoid being in the 40%.
What we charge and why
Diagnostic ₹1.5–3L. Sprint ₹6–12L. Cohort ₹10–20L. Retainer ₹2–5L a month. Fifty per cent upfront, milestone-billed, 45-day terms in the MSA. These are opening hypotheses that get corrected against real scope, and we publish them because a quote-gated market rewards the one vendor willing to say a number out loud.
We price as consulting, not per seat, because we are not a per-seat product: the outcome is a changed workflow, not provisioned access, and that is a workflow-absorption result rather than an adoption one. It is also why your real competitor is the unused licence, not another consultancy.
What this means for a GCC transformation owner
You will be quoted numbers with no way to benchmark them. The defence is not to find a public rate card that does not exist; it is to normalise every quote to the same five questions and to fund the diagnostic before the build. Buy the decision before you buy the thing the decision is about. In a market where 95% of pilots return nothing, the cheapest possible engagement is the one that tells you not to run the expensive one.
Sources
- MIT NANDA, The GenAI Divide: State of AI in Business 2025, July 2025. https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf
- Gartner, Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, 25 June 2025. https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027
Related reading: the ladder of engagements · the adoption diagnostic, the wedge · ten questions to ask any AI vendor
Read next
- Your real competitor is the unused licence you already pay forWe are not competing with another training vendor. We compete with the Coursera, Pluralsight and Copilot seats you already bought and no longer use, and a catalogue seat, however good, cannot change a workflow that exists only inside your organisation.
- Workflow absorptionWorkflow absorption measures whether AI has actually changed how work runs (steps redesigned, cycle time reduced, errors cut) as opposed to adoption, which only counts access such as seats and logins.
- Evaluation harnessAn agent evaluation harness is a repeatable test suite that scores an AI agent's outputs against fixed, versioned cases before and after every change, so teams can tell regression from variance.
More from the blog
- Ten questions to ask any AI consulting vendorAsk a vendor to name the workflow before the engagement starts, to say what they refuse to automate, and to show the harness they leave behind. Including the questions we would fail.
- Writing an AI business case that survives CFO reviewAI business cases fail because they cannot be falsified. Name one workflow, state a measured pre-baseline, and write the kill criteria in advance. A one-page template.
- Your Copilot licences are unused. Here is the 30-day diagnostic.Low seat utilisation is a workflow problem wearing a training problem's clothes. A 30-day audit that finds out which, without surveilling your team.
Frequently asked questions
It is quote-gated, so there is no public benchmark, but as a concrete anchor Chokmah's adoption diagnostic runs ₹1.5–3L for a two-week engagement. Prices scale with the number of workflows shadowed and the depth of the licence audit. Treat any fixed number as an opening range that gets quoted against your actual scope, not a list price.
Because the market is quote-gated and priced on value and reference rather than a public rate card, so vendors keep pricing to the sales conversation. That protects margins and lets them price-discriminate. The side effect is that buyers cannot benchmark, which is exactly why publishing honest ranges (as we do below) is itself a differentiator in this market.
Transformation. What you are buying is a workflow that runs faster or with fewer errors, which is a process outcome, not a learning outcome. Funding it from L&D frames it as training and trains you toward the 95% failure rate. GCC headcount and L&D budgets are tightening; transformation spend is not, which is also where the outcome actually lands.
It varies too widely to quote responsibly without scope, which is the honest answer. The useful frame is comparative: a scoped diagnostic and a single workflow sprint together typically cost a fraction of one failed enterprise pilot, and MIT found 95% of pilots return nothing. Budget for the diagnostic first so you are not sizing a sprint before you know which workflow it is.
For Chokmah, 50% upfront, milestone-billed thereafter, on 45-day payment terms written into the MSA. Terms vary by vendor, but the shape to look for is milestone billing tied to deliverables you can inspect, rather than a large upfront lump against a vague statement of work. Milestones tied to inspectable artefacts protect both sides.
Ready to install the workflow?
Book a free AI Reality Check and build one real thing from your own work, live.