Home / Case studies
Proof of work
Build blueprints and client outcomes
Three detailed blueprints showing exactly how a Kompass build runs — the sequence, the ramp curve, the cost shape and what the centre owned at the end of year one.
Sectors we build for
and SaaS
and BFSI
and life sciences
and mobility
and industrial
portfolios
GCC mandates carry confidentiality terms, so the blueprints below are written as programme patterns rather than named accounts — the sequence, the ramp, the cost shape and the governance are exactly as they run. Named client references are shared during scoping, matched to the shape of build you are considering.
Blueprint one
Platform engineering centre for a US software company
Build-Operate-Transfer · Hyderabad · 12 people at month three, 85 at month eighteen
The situation
A mid-market SaaS company with roughly USD 240 million in revenue could not hire platform and infrastructure engineers fast enough in its home market. Two prior attempts with outsourcing vendors had produced capacity but no ownership: every time a contract ended, the context left with it. The board wanted a centre it would eventually own, but the CTO had no bandwidth to run an incorporation.
What the build looked like
- Weeks 1–3: capability mapping across platform, SRE, data and QA; Hyderabad selected over Bengaluru on cost per productive FTE and attrition
- Weeks 2–8: entity incorporated and registrations completed in parallel with a site-leader search
- Week 9: site leader joined; first engineering pod of six offers issued the same week
- Week 14: floor live in managed seats, payroll running, first sprint connected to the US release train
- Months 6–18: three further pods, an engineering manager layer, and a dedicated floor at month eleven
- Month 22: entity and employees transferred to the client at the pre-agreed consideration
Where it landed
Hiring the site leader in week nine rather than week twenty was worth a full quarter. Everything downstream — the hiring bar, the interview loop, the credibility with senior candidates — ran through that one appointment.The decision that mattered
Blueprint two
Finance and global business services hub for a UK insurance group
Wholly owned subsidiary · Delhi NCR · 60 people across finance, actuarial support and claims operations
The situation
A listed insurance group had board approval for India but no operating experience there. The constraint was not cost — it was the regulator’s view on outsourcing and data residency. The structure had to be a wholly owned entity from day one, with an audit trail their compliance function could defend.
What the build looked like
- Weeks 1–5: regulatory and data-residency design agreed with the group’s compliance and legal functions before any hiring began
- Weeks 3–11: incorporation, registrations, transfer pricing structure and the intercompany service agreement
- Weeks 6–16: finance lead and actuarial lead hired; Gurugram chosen over Noida on candidate commute geography
- Week 18: dedicated floor with segregated access control, CCTV and clean-desk enforcement to satisfy the group audit
- Months 5–12: claims and finance operations pods, monthly close moved to the India team over three cycles
Where it landed
Designing the data-residency and access model before the first job description went out. Bringing compliance in at week one cost five weeks; bringing them in at week twenty would have cost the go-live date.The decision that mattered
Blueprint three
AI and data unit for a private-equity-backed logistics platform
Managed capability centre · Bengaluru · 28 people, senior-weighted
The situation
A portfolio company with a four-year hold period needed applied AI capability it could not hire at home at any price. An entity made no sense: the payback ran past the likely exit. The sponsor wanted capability on the balance sheet story without a legal structure to unwind at sale.
What the build looked like
- Weeks 1–2: charter written around ownership of two model families rather than a headcount target
- Weeks 3–7: senior-weighted hiring at a 1:2 senior-to-junior ratio, paid at the 75th percentile and held there
- Week 8: team live under a managed structure with present-assignment IP clauses in every individual contract
- Months 3–6: evaluation infrastructure, data lineage and model-risk governance established before the first production model
- Month 14: conversion path to a wholly owned entity documented and priced, ready for the buyer’s diligence
Where it landed
Refusing to size the team by headcount. Twenty-eight people who owned two model families outperformed the fifty-five the original plan called for, at well under the budgeted run-rate.The decision that mattered
What these have in common
Three different models, one pattern
The charter came before the headcount
Every one of these builds started by naming what the centre would own end to end. None started with a number of people to move.
Leadership was hired first
The site or functional lead was in place before volume hiring began. That single sequencing choice is worth roughly a quarter on every programme.
The exit was designed at the start
Transfer terms, IP assignment and documentation standards were fixed at signature — not negotiated later, when the centre had become valuable.
Next step
Ask us for references on your shape of build.
Tell us the capability, the headcount and the horizon, and we will point you at the closest work we have done and, where the client permits it, put you on a call with them.