From “what if?” to the AI agent
How technology has changed the way value is understood, measured and created
Rolandas Rakštys · lecturer · Vytautas Magnus University, Agriculture Academy · Faculty of Bioeconomy Development, Department of Business and Rural Development · rolandas.rakstys@vdu.lt · ORCID 0000-0003-0299-9930
Chapter 1 “Foundations of Business Value”, section 1.1.6 · English translation of v0.18.0 (2026-11-03) · set v1.0 · 14 slides, ≈ 18 min
Rakštys, R. (2026). Chapter 1. Foundations of Business Value (English translation of v0.18.0, 2026-11-03). Verslo planavimas ir vertinimas [Business planning and evaluation]. Vytautas Magnus University, Agriculture Academy. https://rakstys.eu/studijos/verslo-planavimas-ir-vertinimas/en/skyriai/ch01/ · CC BY-NC-ND 4.0
Three technological waves
“The concept of business value and the methods of measuring it did not take shape in an intellectual vacuum” – in every era they were determined also by the technological possibilities of the day.
Computerisation
from intuition to the model
The internet and data
from the model to information
AI and AI agents
from information to automated valuation
Three waves have fundamentally changed how value is understood, measured and created.
When DCF was too costly
Before the 1970s business valuation was largely a qualitative process.
Irving Fisher’s discounted cash flow (DCF) logic had been known since the beginning of the twentieth century (Fisher, 1906), but its practical application was limited: manual calculation was “far too laborious”.
“What if?” in minutes
VisiCalc was the first electronic spreadsheet. Its originator, the Harvard Business School student D. Bricklin, wanted a program in which the table could be seen as it was being built, and described his idea as “an electronic blackboard and electronic chalk in a classroom” (Power, 2004). The practical consequence: an iterative “what if” calculation took minutes rather than days. Power notes that the tale is “part myth and part fact”.
Click an assumption: revenue or the discount rate.
The “what if” logic turned valuation into “an iterative process resilient to assumptions” and extended the bounds of modelling “from a single scenario to unlimited matrices”.
Value becomes calculated
Spreadsheets democratised financial modelling: DCF valuation “ceased to be the privilege of elite investment banks and became a standard competence of the analyst”.
Markowitz’s portfolio theory (1952)
remained an academic curiosity for decades because of the complexity of the matrix computations – computers made it possible to calculate covariance matrices for hundreds of assets.
CAPM (Sharpe, 1964)
the Capital Asset Pricing Model was transformed from a theoretical model into an everyday tool for setting the discount rate.
Value became calculated rather than merely judged.
Information becomes available
The Bloomberg terminal
real-time market data; widely adopted from the 1990s
The EDGAR system
US SEC: free access to public financial statements
Damodaran’s open data
databases of risk premia and beta coefficients
XBRL and platforms
XBRL (eXtensible Business Reporting Language); Capital IQ, FactSet, Refinitiv
Valuation rests not on individual statements alone but on continuously updated data ecosystems; information asymmetry between professionals, academics and a growing proportion of practitioners was reduced substantially.
A value that DCF cannot justify
The first dot-com wave showed that traditional valuation models struggle to capture network effects, growth in the user base and the logic of the platform economy.
Amazon, 1999: very high revenue multiples in the absence of profit – a value that the classical DCF model could not justify without extreme growth assumptions.
The response: new valuation models (Parker et al., 2016):
- customer lifetime value (CLV)
- the valuation of network effects
- platform economics
- ecosystem analysis
83 % → 92 %
Capitalisation of US S&P 500 companies (Ocean Tomo, 2025)
On the methodology see section 1.1.2.
The ontology of value could no longer be confined to buildings or equipment: it had to encompass data, algorithms, software and organisational knowledge, even where the classical accounting balance sheet fails to measure them properly.
The consequence: the explanatory power of standard multiples has diminished considerably in valuing technology companies.
In Lithuania: the exchange, registers, APIs
The transformation became particularly evident in the growth of the region’s “unicorns” (Vinted, Nord Security), where the value created is largely bound up with network effects and intangible intellectual property.
Nasdaq Vilnius · since 1993
The then National Stock Exchange provided a transparent price-setting infrastructure without which the valuation of public companies would have remained speculative.
Open data and APIs from the Centre of Registers (RC)
The import of annual statements into analytical modelling can be automated. XBRL (eXtensible Business Reporting Language) in EU reporting formats: from manual data entry to algorithmic processing.
Where AI changes valuation
Large language models (LLMs) and AI agents are already changing valuation practice at several levels.
Can we answer for what we do not understand?
The “black box problem”: investors delegate complex valuations to third-party artificial intelligence platforms – the dilemma of limited transparency.
“…can we accept responsibility for a model’s value when we are unable to decompose its causal logic rationally in the accounts?”
Garbage in – garbage out
A complex model fed with poor assumptions generates results that are not better, merely more plausible in appearance – the illusion of false precision (more in Chapter 2).
Systemic fragility
where many market participants use the same algorithms, correlated valuation distortions can be amplified
Dependence
on third-party data infrastructure, model licences and platform standards – an error can be replicated across the whole market
Responsibility
who answers for an AI-generated valuation? The question remains legally and ethically unresolved
Three waves, three breaks
| Era | Years | Technology | Break |
|---|---|---|---|
| Computerisation | 1970–1995 | the electronic spreadsheet | from intuition to the model |
| The internet and data | 1995–2020 | data infrastructure | from the model to information |
| AI and AI agents | 2020→ | LLMs and AI agents | from information to automated valuation |
A deeper analysis of the relation between technology and value, synthesising all the themes of the book and looking to future prospects, is given in Chapter 19.
References
Fisher, I. (1906). The Nature of Capital and Income. Macmillan.
Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223–2273. https://doi.org/10.1093/rfs/hhaa009
Markowitz, H. (1952). Portfolio Selection. The Journal of Finance, 7(1), 77–91. https://doi.org/10.2307/2975974
Ocean Tomo. (2025). Intangible Asset Market Value Study. Ocean Tomo, a part of J.S. Held. https://www.oceantomo.com/intangible-asset-market-value-study
Parker, G. G., Van Alstyne, M. W., & Choudary, S. P. (2016). Platform Revolution: How Networked Markets Are Transforming the Economy – and How to Make Them Work for You. W. W. Norton.
Power, D. J. (2004, September 30). A brief history of spreadsheets. DSSResources.COM. https://dssresources.com/history/sshistory.html
Sharpe, W. F. (1964). Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of Risk. The Journal of Finance, 19(3), 425–442. https://doi.org/10.2307/2977928
Rakštys, R. (2026). Chapter 1. Foundations of Business Value (English translation of v0.18.0, 2026-11-03). Verslo planavimas ir vertinimas [Business planning and evaluation]. Vytautas Magnus University, Agriculture Academy. https://rakstys.eu/studijos/verslo-planavimas-ir-vertinimas/en/skyriai/ch01/ · CC BY-NC-ND 4.0
These slides are based on the textbook Verslo planavimas ir vertinimas [Business Planning and Evaluation]. The text was translated, structured and visually laid out with the help of artificial intelligence tools (Claude), following the protocol for working with AI and the verification procedures set by the author.