Not what the technology can do. What it does to the organisation.
There is a crowded, well-capitalised market selling AI tooling, platform integration and model deployment to capability centres. There is a second, largely unoccupied market: the organisational consequence of adopting it. Successgini operates in the second.
Pilots do not fail for technical reasons.
A very large majority of capability centres are experimenting with new technology, and a substantial share are trialling agent-based systems. Only a small fraction have reached advanced maturity. The distance between experimentation and maturity is not a modelling problem. Pilots stall because nobody redesigned the work, the roles, the incentives or the accountability around them.
What the transition actually involves.
- Work Redesign
Task-level decomposition of functions to establish what is automatable, what is augmentable, and what remains irreducibly human. The output is a redesigned role architecture — not a technology recommendation.
- Human–Agent Operating Model
Where humans set intent and guardrails, where agents execute, where escalation occurs, and who is accountable when an autonomous system errs. Decision rights written down before an incident tests them.
- Capability Currency
Structured reskilling with genuine progression attached, not a training catalogue. This addresses directly the widely reported anxiety among high performers that their skills are decaying faster than they can renew them — a retention risk before it is a capability risk.
- Reward Consequence
When output per person changes materially, a reward architecture calibrated to the previous output becomes indefensible. This workstream re-architects it, and connects directly to Pillar II.
- People-Data Governance
Model governance, data boundaries and access control wherever AI systems touch employee information — the posture a global parent will require of a subsidiary, and the one most centres have not written down.
The AI Act reaches a centre in India whenever the output is used in the Union.
What moved
The Digital Omnibus on AI — Regulation (EU) 2026/1744 — entered into force on 27 July 2026, deferring high-risk obligations for standalone Annex III systems from 2 August 2026 to 2 December 2027, and embedded Annex I systems to 2 August 2028. Employment uses sit squarely in Annex III: recruitment, selection, CV filtering, candidate ranking, performance evaluation and termination decisions.
Our readingThe deferral is a resourced runway, not a reprieve. Conformity work on this scale takes twelve to eighteen months; a centre that stands down now will arrive at December 2027 in exactly the position it was in this August.
What did not move
Article 50 transparency duties applied from 2 August 2026, with legacy systems given until 2 December 2026 for machine-readable marking of AI-generated content. The Article 4 AI literacy obligation is unchanged. The prohibitions in force since February 2025 remain — and they include emotion recognition in the workplace and biometric categorisation to infer sensitive attributes, both of which have appeared in HR tooling.
What we buildAn inventory and classification of every people system in the centre, human oversight and logging designed rather than retrofitted, and a literacy position that is evidenced.
What is handed over.
- Function-level automation and augmentation map
- Redesigned role architecture with revised job families
- Human–agent operating model with decision rights and escalation paths
- Reskilling pathway design with progression and assessment criteria
- Revised reward architecture reflecting changed output
- People-data governance framework
- AI system inventory with risk classification and oversight design
Led by the CAIO mandate in concert with the CHRO mandate. The intersection is the point — neither role delivers this alone.
