A company has Rs 100 crore of deployable capital. The options: plant expansion (Rs 50 crore, projected 22% IRR), acquisition (Rs 80 crore, projected 18% IRR), new product development (Rs 15 crore, projected 28% IRR but high uncertainty), debt repayment (Rs 30 crore, guaranteed 9% cost savings), geographic expansion (Rs 25 crore, projected 20% IRR, unproven market), technology investment (Rs 10 crore, projected efficiency gains of Rs 3 crore per year) and cash reserve (Rs 0 deployment, optionality preserved).
The capital is finite. The options are not mutually exclusive but are constrained by the Rs 100 crore budget. Some combinations work (technology + debt repayment + new product = Rs 55 crore). Others do not (plant expansion + acquisition = Rs 130 crore, exceeding available capital).
The traditional approach: the CFO builds a spreadsheet, ranks the options by IRR, presents the top three to the board and the board approves.
The AI-augmented approach: a capital allocation decision engine that evaluates every option and every feasible combination across multiple dimensions simultaneously, producing a decision space rather than a ranked list.
What the Copilot Does
Constructs the decision space
The copilot evaluates every feasible combination of investments within the capital constraint. With seven options and a Rs 100 crore budget, the number of feasible combinations runs into dozens. Each combination has a different aggregate NPV, aggregate IRR, risk profile, liquidity impact, covenant consequence and strategic alignment.
The copilot presents the five highest-value combinations, ranked by probability-weighted NPV, with each combination’s risk profile, liquidity trajectory and covenant impact displayed alongside the return.
Stress-tests every option
For each investment, the copilot runs the scenario analysis automatically: base, downside, severe downside. It identifies the assumptions that each investment depends on, tests them against historical evidence and flags assumptions that are above the 75th percentile of historical outcomes (aggressive) or below the 25th percentile (conservative).
Models interaction effects
Plant expansion and geographic expansion together may generate synergies (shared logistics) or conflicts (management bandwidth constraint). The copilot models these interactions, adjusting the combined NPV for estimated synergies and dis-synergies.
Evaluates optionality
Debt repayment reduces leverage and creates capacity for future borrowing. Cash reserve preserves flexibility for an unforeseeable opportunity. New product development creates an option on a new revenue stream. The copilot values the optionality created by each investment alongside its NPV.
Produces the Capital Allocation Dashboard
For each recommended combination:
Combination A: Technology + Debt Repayment + New Product + Geographic Expansion (Rs 80 crore) NPV: Rs 32 crore. Probability-weighted NPV: Rs 24 crore. Maximum downside: Rs 18 crore. Cash reserve: Rs 20 crore. Covenant impact: leverage improves from 2.8x to 2.4x.
Combination B: Plant Expansion + Technology (Rs 60 crore) NPV: Rs 28 crore. Probability-weighted NPV: Rs 22 crore. Maximum downside: Rs 35 crore. Cash reserve: Rs 40 crore. Covenant impact: leverage stable at 2.8x.
Combination C: Acquisition (Rs 80 crore) NPV: Rs 26 crore. Probability-weighted NPV: Rs 15 crore. Maximum downside: Rs 55 crore. Cash reserve: Rs 20 crore. Covenant impact: leverage increases to 3.4x.
The dashboard presents the decision space. The board selects the combination that best balances return, risk, liquidity and strategic alignment. The AI remains the analytical engine. The board remains the decision-maker.
What the Copilot Cannot Do
The copilot cannot assess strategic intuition: whether the market will shift in a direction that makes one investment brilliant and another obsolete. It cannot evaluate management chemistry: whether the team has the capacity to execute three initiatives simultaneously. It cannot judge competitive response: whether the plant expansion will provoke a competitor to build a larger plant.
These judgments require human experience, industry knowledge and competitive instinct that AI does not possess. The copilot provides the analytical foundation. The board provides the judgment.
In Northrop Management Private Limited’s capital allocation advisory, the copilot framework is deployed to ensure that the board sees the full decision space, not just the CFO’s preferred option. The framework does not replace board judgment. It ensures that judgment is applied to the best available analysis rather than to a single spreadsheet.
Ashish Chaudhary, frames the governance enhancement directly: “AI does not decide where the next rupee should go. It constructs the decision space: every feasible combination of investments, stress-tested, interaction-modelled and displayed with return, risk, liquidity and covenant impact side by side. The board that sees the decision space makes a choice. The board that sees a single recommendation makes an approval. The difference is governance.”
Questions for the Boardroom
- When we make capital allocation decisions, do we evaluate individual projects in isolation or combinations of projects within the capital constraint?
- Has each investment option been stress-tested under downside scenarios, and does the board see the probability-weighted NPV alongside the base-case NPV?
- Have we modelled the interaction effects between investments (synergies, conflicts, management bandwidth constraints)?
- Does the board see the covenant impact and liquidity trajectory of each capital allocation combination?
- Are we preserving optionality (cash reserve, reversible investments, real options) or committing the full capital budget to irreversible deployments?
Closing Implication
Capital allocation is the highest-leverage decision a board makes. Every rupee deployed on one investment is a rupee not available for another. The copilot does not make the decision. It ensures the board sees every feasible combination, stress-tested and displayed across multiple dimensions, before the decision is made. The board that decides from a decision space governs with full information. The board that decides from a single recommendation governs with the CFO’s preference, which may or may not be the combination that creates the most value at acceptable risk.