Why Valuation Is More Art Than Science (AI Era)

 

Introduction

In finance, valuation is often perceived as a precise, numbers-driven process. With the rise of artificial intelligence (AI), data analytics, and automation, many believe valuation has become even more scientific and accurate. However, despite technological advancements, valuation remains deeply rooted in judgment, assumptions, and interpretation—making it more of an art than a science.

Whether it's valuing a startup, a listed company, or a merger target, no two analysts arrive at exactly the same number. Even AI models, trained on massive datasets, cannot eliminate subjectivity entirely. This blog explores why valuation continues to be an art, even in the AI-driven era.

 

1. The Illusion of Precision in Valuation

At its core, valuation uses mathematical models such as:

• Discounted Cash Flow (DCF)
• Comparable Company Analysis
• Precedent Transactions

These models give the impression of precision. A DCF model may output a valuation like ₹1,234 crore, but the reality is far less exact.

  

Why?

Because every input in the model is an estimate:

• Future cash flows
• Growth rates
• Discount rates
• Terminal value

Even a small change in assumptions can drastically alter the final valuation.

 

2. Assumptions Drive Everything

Valuation is heavily dependent on assumptions about the future—something no model can predict perfectly.

Key Assumptions Include:

• Revenue growth
• Profit margins
• Market conditions
• Competitive landscape

For example:

• One analyst may assume 10% growth
• Another may assume 15% growth

This difference alone can change valuation significantly.

 

3. Human Judgment Still Dominates

Even in the AI era, human judgment plays a central role in valuation.

Where judgment is required:

• Selecting the right model
• Choosing comparable companies
• Adjusting financial data
• Interpreting industry trends

AI can assist, but it cannot fully replace human reasoning, especially in complex or uncertain situations.

 

4. Market Sentiment Matters

Valuation is not just about intrinsic value—it is also influenced by market sentiment.

Factors affecting sentiment:

• Investor confidence
• Economic outlook
• News and events
• Market trends

For instance:

• During a bull market, companies are often overvalued
• During a recession, even strong companies may be undervalued

 

5. The Challenge of Valuing Intangibles

Modern businesses derive significant value from intangible assets such as:

• Brand value
• Intellectual property
• Customer loyalty
• Technology

These are difficult to quantify accurately.

Example:

How do you assign value to:

• A strong brand?
• User trust?
• Future innovation potential?

Even AI struggles to measure these factors effectively.

 

6. Different Methods, Different Results

There is no single “correct” valuation method.

Common approaches:

• DCF (future-focused)
• Comparable analysis (market-based)
• Asset-based valuation

Each method can give a different result.

Example:

• DCF value: ₹100 per share
• Comparable value: ₹120 per share
• Asset-based value: ₹80 per share

Which one is correct?

 

7. AI Enhances—but Doesn’t Replace—Valuation

AI has transformed the valuation process by:

• Processing large datasets quickly
• Identifying patterns
• Automating repetitive tasks

However, AI has limitations:

• It relies on historical data
• It struggles with unprecedented events
• It cannot fully understand human behavior

Example:

AI may not predict:

• Sudden regulatory changes
• Black swan events
• Shifts in consumer preferences

8. The Impact of Uncertainty and Risk

Valuation is inherently uncertain because it deals with the future.

Key uncertainties:

• Economic cycles
• Interest rates
• Political changes
• Industry disruptions

Even the best models cannot fully account for these risks.

 

9. Behavioral Biases Influence Valuation

Human biases can significantly affect valuation outcomes.

Common biases:

• Overconfidence
• Anchoring
• Confirmation bias

Example:

An analyst may:

• Favor a company they like
• Ignore negative data
• Stick to initial assumptions

Even AI models can inherit biases from training data.

10. Startups vs Mature Companies

Valuation becomes even more “art-like” when dealing with startups.

Challenges:

• No stable cash flows
• Uncertain business models
• High growth variability

Startup valuation relies on:

• Vision
• Founder credibility
• Market potential

 

11. The Role of Narrative in Valuation

Valuation is not just numbers—it’s also storytelling.

Analysts create a narrative:

• Why will the company grow?
• What is its competitive advantage?
• How will it sustain profits?

A strong narrative can justify higher valuations.

 

12. Sensitivity Analysis Shows the Uncertainty

Small changes in inputs can lead to large valuation differences.

Example:

• Discount rate: 10% → Value = ₹100
• Discount rate: 12% → Value = ₹80

This shows how fragile valuations can be.

 

13. Relative vs Absolute Valuation

Valuation can be:

• Absolute (based on fundamentals like DCF)
• Relative (based on market comparisons)

Both approaches have limitations.

Example:

A company may look:

• Cheap in DCF
• Expensive compared to peers

 

14. Real-World Deals Prove the Point

In mergers and acquisitions (M&A), valuation often depends on negotiation.

 

Factors influencing deal value:

• Strategic fit
• Synergies
• Bargaining power

Example:

A buyer may pay a premium because:

• The acquisition provides long-term benefits

 

15. AI Era: More Data, Same Uncertainty

Despite access to massive data and advanced tools:

• Predictions are still uncertain
• Markets remain unpredictable
• Human behavior is complex

AI reduces errors but does not eliminate ambiguity.

 

16. Valuation as a Range, Not a Number

Experienced analysts often present valuation as a range:

• Base case
• Bull case
• Bear case

This reflects uncertainty and different scenarios.

17. The Role of Experience

Experienced analysts develop intuition over time.

They can:

• Identify unrealistic assumptions
• Understand industry dynamics
• Make better judgment calls

AI lacks this human intuition.

 

18. Ethical Considerations in Valuation

Valuation can sometimes be influenced by incentives.

Example:

• Investment bankers may inflate valuations to win deals
• Analysts may adjust assumptions to support a target price

 

19. Continuous Evolution of Valuation

Valuation methods evolve with time:

• Traditional metrics → Digital metrics
• Tangible assets → Intangible assets
• Financial data → Alternative data

Yet, subjectivity remains constant.

 

20. Final Thoughts

Valuation will never be purely scientific, even in the AI era.

Why?

Because it involves:

• Predicting the future
• Interpreting incomplete data
• Understanding human behavior

AI enhances valuation but cannot replace:

• Judgment
• Experience
• Intuition

 

21. Models Assume Stability

Most valuation models assume:

• Stable growth rates
• Predictable cash flows

But real businesses face:

• Sudden demand shifts
• Economic downturns

Even a strong company can deviate sharply from projections within months.

22.  Input Sensitivity Makes Models Fragile

Small changes in assumptions (like discount rate or growth rate) can lead to drastically different valuations. This makes models highly sensitive and unstable, especially in uncertain environments.

 

23. Human Interpretation Still Wins

Because models break under uncertainty, analysts rely on:

• Industry experience
• Judgment
• Scenario thinking

 

24. Conclusion

Valuation sits at the intersection of numbers and judgment. While models, data, and AI provide structure and support, they cannot eliminate uncertainty or subjectivity. The assumptions we make, the narratives we believe, and the risks we perceive all shape the final outcome.

In the end, valuation is not about finding the “correct” number—it’s about understanding a range of possibilities and making informed decisions.

That’s why, even in the most advanced AI-driven world, valuation will always remain more art than science.