Gains:
- Ability to configure and execute steps of discounted cash flow (DCF) and multiplier methods with AI
- Ability to test the effect of critical assumptions such as discount rate, growth and terminal value on the result with AI
- Ability to question the valuation result produced by AI with sensitivity and plausibility control
Valuation is “how much is this company or asset worth?” It is a systematic search for an answer to the question. This question lies at the heart of an acquisition, partnership, IPO or investment decision. There are two big families of valuation: discounted cash flow (DCF), that is, taking the cash the company will generate in the future to its current value; and multiples, that is, making comparisons based on the market value of similar companies. AI dramatically speeds up valuation: it configures DCF steps, discounts cash flows, calculates terminal value, and produces sensitivity tables. But the harsh reality of valuation is that the outcome is extremely sensitive to a few critical assumptions. Changing the discount rate by one point can move the value by tens of percentages. So it is not a single number of values that the AI produces, but a range of values based on assumptions, that is the actual output.
In this unit, we will learn to carry out the steps of the DCF and multiplier methods with AI, test the impact of critical assumptions on the result, and check the valuation result for reasonableness.
Concepts: DCF: Method that discounts future cash flows to the present using a discount rate. Discount rate: Rate used to find the present value of future money; reflects the risk. Terminal value: Summary of all cash flows after the forecast period. Multiplier: Benchmark ratio that divides the market value by a financial quantity (profit, EBITDA, turnover). EBITDA: Earnings before interest, taxes and depreciation.
Steps of DCF
DCF looks complex, but its logic is chained. Having the AI build this chain step by step provides both transparency and makes it easier to isolate the error.
step
what to do
critical assumption
1. Cash flow forecast
Free cash flow for the next 5 years
growth, margin
2. Discount rate
The rate reflecting the risk is determined
cost of capital
3. Reduction
Each year's cash is withdrawn to today
discount rate
4. Terminal value
Summarized after the 5th year
terminal growth
5. Total
Reduced values + terminal
combination of all
Critical fact: Usually more than half of the DCF value comes from the terminal value. Terminal value is extremely sensitive to the small difference between the discount rate and the terminal growth rate. These two assumptions are the most fragile point of DCF and must be subjected to sensitivity analysis.
Caution: Telling the AI to "value this company" and getting a single number goes against the spirit of valuation. If the discount rate is 13% instead of 11%, the value drops perhaps 25%. An odd number gives misleading precision; The correct output is a band of values within a reasonable assumption range.
Building DCF with AI
DCF framework prompt: "Set up a 5-year DCF for a company. Inputs: 1st year free cash flow $20 million, 8% annual growth, discount rate 14%, terminal growth 4%. Show steps separately: (1) 5-year cash flow, (2) discounted value of each year, (3) terminal value plus discounted value, (4) total company value. Write formulas, constant embedding."
Sensitivity matrix prompt: "Construct a two-variable sensitivity table for the same DCF: discount rate in the rows (12%-16%, 1 point step), terminal growth in the columns (2%-5%, 1 point step). Show the total company value in each cell and comment on how sensitive the value is to these two assumptions."
Multiplier valuation prompt: "Value a company with the multiplier method. Its EBITDA is 30 million TL. The EV/EBITDA multiplier of similar companies is in the range of 6x-8x (consider this range as 'resources required'). Calculate the firm value band of the company with this range and give it in a format comparable to the DCF result."
Reasonableness check prompt: "Check the DCF value you found as follows: (1) how many times the company's annual EBITDA does the resulting value correspond to and is this multiplier reasonable according to the sector? (2) what percentage of the total value is the terminal value? If it is over 80%, explain why it is risky. (3) how much higher is the discount rate than the terminal growth; if the difference is very small, warn."
Weak Prompt / Strong Prompt
WEAK: "How much is this company worth?" (Result: a single unsupported figure; which assumption, which method is unclear.) STRONG: "Set up DCF with 5-year free cash flow, 14% discount, and 4% terminal growth assumptions, show the steps; then add a sensitivity matrix for the discount and terminal growth, and report the terminal value as a share of the total, so I get a value band, not a single number."
Mini Cases
Case 1 — Weight of terminal value. An analyst values a company at 240 million TL with DCF. When it runs the plausibility prompt, the terminal sees that the value is 86% of the total. So almost all of the value is based on a growth assumption 5 years ahead. When the analyst reduces terminal growth from 4% to 3%, the value drops from 240 to 205 million. The decision is made by clearly reporting this vulnerability.
Case 2 — Two methods in one. In a purchase valuation, DCF gives 180 million TL, and the multiplier method (6x-8x EBITDA) gives a band of 180-240 million TL. The overlap of two independent methods increases confidence; The fact that DCF is at the lower end of the band indicates that DCF assumptions are relatively conservative. By comparing the two methods, the analyst provides a realistic range for negotiation.
Case 3 — Impossible difference. In a DCF, AI took the discount rate as 5% and the terminal growth as 4.5%; Because the difference is so small, terminal value explodes and company value becomes absurdly high. In his plausibility check, the analyst finds that this difference is very narrow: if the discount rate is only slightly above growth, the terminal value is mathematically overinflated. Once the assumptions are made realistic, the value returns to normal.
Tip: Be sure to use DCF in conjunction with a multiplier method, not alone. If two independent methods converge, your confidence increases; If they are very far apart, there is something questionable in the assumptions of at least one of them.
The Golden Rule of Valuation
Valuation is not a matter of precision, but a matter of judgment and range. Maintain three principles when using AI:
- Band, not single number: Convert each valuation into a band within a reasonable assumption range.
- Own the assumption: Don't let the AI choose the discount rate and growth; You determine and justify it.
- Cross-validation: compare DCF and multiplier; The big divergence is a warning.
Common mistakes
- Valuing with a single number. The value given without sensitivity is misleadingly accurate.
- Ignoring the numerator of terminal value. If most of the value comes from the terminal, the risk is there.
- Leaving the discount-growth gap narrow. The small difference mathematically inflates the terminal value.
- Leaving the guesswork to AI. The discount rate is a judgment; There must be a source and justification.
- Being satisfied with one method. Not comparing DCF and multiplier makes the error invisible.
In summary
Valuation is a matter of judgment that draws future cash into the present (DCF) or compares it with peers (multipliers). AI quickly establishes the steps, but the result is extremely sensitive to several critical assumptions, especially the discount rate and terminal growth. Produce a band of values with a sensitivity matrix, not a single number; check the share of the terminal value in the total; Be sure to cross-validate the DCF with a multiplier method. Avoid the illusion of certainty in valuation.
Application task
Have the AI install a 5-year DCF; You give the cash flow, discount rate and terminal growth assumptions and have the steps shown separately. Then ask for a bivariate sensitivity matrix for the discount rate and terminal growth. Run the plausibility check prompt and evaluate the share of the terminal value in the total and the discount-growth difference. Finally, evaluate the same company with an EBITDA multiplier band and compare it with the DCF result.
checklist
- [ ] I showed the DCF steps (cash flow, discount, terminal, total) separately.
- [ ] I determined and justified the discount rate and terminal growth myself.
- [ ] I constructed a two-variable sensitivity matrix.
- [ ] I checked the share of the terminal value in the total.
- [ ] I verified that the discount-growth difference is reasonable.
- [ ] I cross-validated the DCF with a multiplier method.
- [ ] I presented the result as a value band, not a single number.