The Damodaran Bot
Damodaran's AI reflection starts from a concrete data point: a colleague built an AI bot trained on his published work. The bot, by some measures, appears capable of valuing companies more accurately than his students. His honest reading of the result captures the central dilemma for every knowledge worker:
- If the bot performs well, it suggests he may be becoming obsolete
- If the bot performs poorly, it reflects on the quality of his training material
There's no comfortable answer to the paradox, which is itself the point.
Where AI Actually Threatens Jobs
Damodaran offers a useful three-dimensional framework for assessing which work is exposed:
- Mechanical vs. intuitive — formulaic work is easily automated; adaptive, judgment-heavy work is less so
- Rules-based vs. principle-based — decisions governed by clear rules are easier to encode than those requiring principled judgment
- Objective vs. bias-prone — fully objective tasks are low-hanging AI fruit; tasks where human bias currently drives outcomes may be shifted or improved by AI rather than eliminated
"Jobs that are more mechanical, rule-based, and objective are at a higher risk of being replaced by AI."
Principle-based work requiring contextual judgment remains substantially safer.
Counter-Strategies for Practitioners
Damodaran offers four constructive responses:
- Become a generalist across multiple disciplines. Cross-domain synthesis remains hard for models optimized on specific tasks.
- Blend narrative and numerical analysis. A valuation grounded in a defensible business story is more durable than a spreadsheet output.
- Keep independent reasoning skills sharp. If you stop reasoning because the bot will reason for you, you lose the only thing the bot can't easily replicate.
- Cultivate creativity through exploration. Approaching problems from unusual angles remains difficult for models trained to produce average outputs.
The Defensive Tactics (Offered Skeptically)
For the "beat your bot" framing, Damodaran lists three tactical moves — noting each has clear limits:
- Keep processes private. Protects against training data leakage but doesn't scale.
- Advocate for regulatory protections. May work in licensed professions (law, medicine, audit) but is unlikely to hold the line across the broader economy.
- Build irreplaceable expertise. The most durable moat, but also the hardest to construct — it requires genuine depth, not just credentialing.
Implications for Valuation Practice
For the valuation profession specifically:
- Mechanical valuation tasks — updating comps, maintaining data tables, running standardized DCFs — are already partially automated and will become more so
- Judgment-heavy work — selecting the right methodology, defending terminal value assumptions, integrating narrative with numbers — remains the defensible core
- The practitioner's value increasingly lives in the story, not the spreadsheet
Bottom Line
AI threats to knowledge work are real but uneven. Mechanical, rules-based, objective tasks are exposed; principle-based, narrative-rich, judgment-heavy work is more defensible. The response that actually works is the harder one — deepening expertise, broadening across domains, and preserving the reasoning capacity that makes human practitioners valuable in the first place.