Reskilling budgets in Indian banking rarely pay off on their own — pairing training investment with active change management to reduce employee resistance is essential, the authors argue, and they propose an Invest–Engage–Institutionalize framework to help firms convert reskilling into measurable ROI.

Main Finding

Reskilling in Indian BFSI is economically justified only when training produces real, sustained on‑the‑job adoption; unmanaged employee resistance driven by job‑insecurity substantially reduces or nullifies measurable ROI. The paper proposes an integrated Invest–Engage–Institutionalize (IEI) sequence that treats reskilling as a single, sequential decision problem linking financial planning (ROI) with resistance‑focused change management and capability embedding.

Key Points

• Economic framing

  • Human capital theory: distinguish firm‑specific vs general (portable) skills. Employers have stronger incentives to fund firm‑specific training; general skills risk turnover and underinvestment.
  • To capture full value of portable skills, organizations may need retention mechanisms (career pathways, bonded learning, promotion-linked commitments).

• Measurement and evaluation

  • Use Kirkpatrick’s four levels plus Phillips’ ROI level: Reaction, Learning, Application, Business Results, and Monetized ROI.
  • Most Indian BFSI reskilling reporting is concentrated at Levels 1–2 (satisfaction, learning); few programs measure Levels 3–5 (behavior change, business impact, financial return).

• Behavioral barriers

  • Job insecurity and pre‑adoptive appraisals (perceived threat and usefulness) drive resistance more than an unwillingness to learn.
  • Routine/clerical banking roles are especially exposed to automation; fear of displacement is empirically common in South Asian banking samples.

• Change management

  • Classical change models (Lewin; Kotter) emphasize unfreezing, active change, and refreezing/institutionalization; absent reinforcement, new behaviors revert.
  • Structured post‑training supports (mentoring, peer shadowing, manager training) are strongly associated with positive ROI in industry evidence.

• Integrated gap

  • Literature typically treats finance/ROI and HR/change‑management separately. This separation is a practical barrier: investment decisions that ignore adoption dynamics will overstate expected returns.

Data & Methods

• Design: conceptual thematic review synthesizing economic and organizational‑behavior literatures, plus recent empirical studies on AI adoption in banking (focus on South Asia/India).
• Sources: four types — foundational human capital and training‑evaluation texts; organizational‑behavior literature on change and job insecurity; recent empirical AI‑adoption studies in banking; industry/institutional reports on reskilling investment. Fifteen sources retained for synthesis, with empirical AI adoption studies prioritized from the last five years.
• Analysis: thematic synthesis producing the IEI integrated framework.
• Limitations: no primary data collected; broad treatment of “BFSI” without disaggregation by bank type (public/private/NBFC/insurance); quantitative parameter estimates and causal verification left for future empirical work.

Implications for AI Economics

• Valuation of reskilling must include adoption probability and behavioral frictions

  • Standard ROI calculations should not only monetize training costs and potential productivity gains but adjust expected benefits by the likelihood of on‑the‑job adoption (i.e., the transition from learning to sustained behavior).
  • Behavioral costs (reduced discretionary effort, turnover, litigation risk, reputational/regulatory costs) need to be treated as negative components or risk adjustments in cost‑benefit models.

• Policy and firm strategy

  • Firms should prioritize firm‑specific reskilling investments (higher capture of benefits) and design retention/contractual mechanisms when investing in general AI/data skills.
  • Regulators or industry consortia could subsidize general AI upskilling or support portable‑skill certification to correct market underinvestment and reduce employee mobility frictions.
  • Implement structured post‑training supports (mentors, manager coaching, shadowing, role redesign) as budget line items; these materially affect ROI and should be costed in economic models.

• Measurement and empirical research needs

  • Move evaluation beyond completion and satisfaction metrics: collect data on tool use, workflow changes, processing times, error rates, complaint volumes, turnover, and monetize these outcomes.
  • Empirical work should estimate: adoption elasticities to change‑management effort; the marginal ROI of different support components (mentoring, managerial incentives); heterogeneity across bank types and roles; and optimal split between firm‑specific and general training investments.

• Modeling reskilling as a sequential decision problem

  • Treat investment choice (how much to spend, what content) and implementation choice (how to manage resistance and institutionalize changes) as linked stages. Economic models of firm-level human capital should incorporate a behavioral adoption stage with associated probabilities and costs.
  • Comparative statics: higher perceived job insecurity lowers adoption probability—so firms facing greater insecurity should allocate more to engagement and institutionalization per rupee of training expenditure to achieve the same ROI.

• Practical evaluation guidance for practitioners/economists

  • When building ROI projections, include:
    • Full loaded training costs + post‑training support costs.
    • An empirically justified adoption discount factor (probability that training leads to sustained behavior).
    • Monetized benefits only from Levels 3–5 outcomes, with sensitivity analysis over adoption rates and turnover.
  • Use randomized or phased rollouts where feasible to estimate causal impact of training+engagement interventions on productivity and retention.

Future research priorities highlighted by the paper: collect primary panel data on training, adoption, and outcomes in Indian BFSI; quantify the cost‑effectiveness of specific engagement interventions; and explore regulatory instruments to mitigate market underinvestment in general AI skills.