EBITDA multiples are the most commonly requested and most commonly hallucinated metric in AI business tools. Ask almost any AI tool what your company might be worth — or what EBITDA multiple a company in your sector typically commands — and you will receive a confident, specific answer. That answer is often wrong in ways that are difficult to detect and expensive to act on.

This is not because the AI is malfunctioning. It is because EBITDA multiples are fundamentally situational. They depend on a constellation of factors that are highly specific to your company, your industry niche, your market position, and the current deal environment. Any tool that produces a multiple without accounting for all of these variables is producing an inference, not a valuation. And the problem is that the inference often looks identical to the real thing.

Why Multiples Are So Variable

EBITDA multiples are not stable, broadly applicable numbers. They shift dramatically based on a set of well-understood value drivers. Understanding each of these is the first step in evaluating whether any multiple estimate — AI-generated or otherwise — is applicable to your situation.

Industry Sector

Software companies (particularly those with recurring revenue) have historically commanded meaningfully higher EBITDA multiples than manufacturing, distribution, or professional services businesses. Healthcare has its own range, which varies further by sub-sector. Media and content businesses trade differently than technology-enabled services businesses that look similar on a revenue line. These are not small differences — the range of EBITDA multiples across industries is wide, and a multiple that is accurate for one sector is potentially misleading in another.

Company Size

Smaller companies in the private market consistently trade at lower multiples than larger companies in the same industry, for two structural reasons: they carry higher execution risk (the loss of a key customer or team member has larger proportional impact) and they are less liquid (the universe of qualified buyers for a $5M EBITDA business is smaller than for a $50M EBITDA business). This size-related discount is real and consistent, and AI tools that produce multiples without adjusting for company size systematically overestimate value for smaller businesses.

Growth Rate

A company growing revenue meaningfully faster than its peer group commands a premium multiple over a company growing at a slower rate, assuming equivalent profitability and quality. This growth premium is particularly pronounced in sectors where market share capture is competitively significant. Conversely, a declining-revenue business may trade at a discount to EBITDA even if the near-term cash flows look stable.

Recurring vs. Project Revenue

Contracted, recurring revenue is worth more than equivalent project-based revenue. Buyers value predictability: they pay for the assurance that revenue will be there in the periods after close. A business with 80% of revenue under multi-year contracts commands a different multiple than a business with equivalent EBITDA but mostly project revenue that must be re-won each cycle. The mix matters, and AI tools frequently treat "revenue" as a single undifferentiated category.

Customer Concentration

If a meaningful share of revenue comes from a small number of customers, buyers price that concentration risk directly into the multiple. The concern is straightforward: post-close, if one of those customers churns or renegotiates terms, the acquirer's return on their investment is materially impaired. The degree of concentration discount varies by industry and specific circumstances, but it is a consistent and often material factor.

Management Dependency

In founder-led businesses, a portion of the company's value — customer relationships, strategic direction, sales capability, operational knowledge — may reside in the founder rather than in the institutional business. Buyers assess what the business looks like if the founder exits at or shortly after close. If that picture is uncertain, the deal structure will reflect it: lower upfront consideration, earnouts tied to performance, consulting arrangements, or escrow holdbacks.

Market Conditions

Private company M&A multiples are sensitive to the broader market environment: interest rates (which affect the cost of leveraged acquisition financing), public market valuations (which set anchor points for private market expectations), deal activity (which affects competition among buyers and sellers), and sector-specific dynamics. A multiple that was accurate in a low-rate, high-liquidity environment may not be accurate in a different one.

The core problem: A valid EBITDA multiple for your company requires simultaneous adjustment for all of these variables. Any multiple estimate that does not account for your specific industry classification, size range, growth rate, revenue quality, customer concentration, and current market vintage is not estimating your multiple — it is producing an industry average that may not apply to your situation at all.

Why AI Tools Produce Unreliable EBITDA Multiple Estimates

AI tools synthesize their multiple estimates from publicly available information. This corpus includes public company trading multiples (which are for larger, more liquid businesses), disclosed private equity transactions (which are selectively disclosed and skew toward larger deals), published industry guides (which aggregate across wide ranges and vintage years), and general financial content in the training data.

Private company transaction data — specifically, what was actually paid for a company in your exact industry niche, your exact size range, with comparable growth and revenue quality characteristics, in the current market environment — is not public. It lives in proprietary databases that require professional access and cost to query. The comparables that matter most for your valuation are almost certainly underrepresented in any AI tool's training data.

The result is that AI-generated multiples are effectively an average over a wide, poorly-defined peer set, stated with the precision and confidence of a specific, sourced estimate. The stated range of "5x–7x EBITDA" may reflect averaging across many deals in many industries and many years. The correct answer for your specific company in the current market could be anywhere from 3x to 12x, depending on the factors described above.

The False Precision Problem

The particular danger of AI-generated multiples is not that they are wrong — it is that they are stated with a precision that implies the analysis was done. A range of "5x–7x EBITDA" feels specific. It is the kind of range a banker might give you. It provides a false floor and ceiling that anchors subsequent thinking.

When you present this range to your board, your co-founders, or your family as your expected exit valuation, it takes on a life of its own. People plan around it. Career decisions get made around it. Lifestyle adjustments get made around it. And when the actual market feedback is different — either higher or lower — the recalibration is painful in proportion to how concretely the number was relied upon.

The most expensive version of this problem is when a seller anchors their go-to-market price expectation to an AI-generated multiple, communicates that expectation to buyers early in a process, and then must either walk away from interest or retrade when the market feedback differs. Both outcomes are worse than starting with an accurate valuation expectation.

What a Proper Multiple Estimation Requires

Generating a credible EBITDA multiple range for your specific company is not a simple query task. It requires:

This is not a process that an AI tool with a general question-and-answer interface can replicate. It requires professional access to private transaction data, sector knowledge, and analytical judgment about comparability.

How to Use AI-Sourced Multiples Appropriately

None of this means you should ignore AI-generated multiple estimates entirely. They have appropriate uses in the analytical process:

What AI-sourced multiples should not be used for is pricing a transaction, communicating expectations to buyers or investors, planning liquidity around, or anchoring any decision with multi-million dollar consequences.

The Cost of Overestimating Your Multiple

Entering a sale process with a price expectation that the market won't support is one of the most common and most preventable M&A mistakes for mid-market companies. The costs are concrete:

An accurate valuation expectation, established before you go to market, eliminates all of these costs. It allows you to make an informed decision about timing, prepare the business to address the specific factors that are suppressing your multiple, and enter conversations with expectations aligned to what the market will support.

The right sequence: Before anchoring on any multiple, validate it against: (1) a professional advisor with access to comparable private transaction data in your specific industry; (2) a quality of earnings analysis that establishes your actual normalized EBITDA; and (3) an honest assessment of the value-driver factors described above and how your business scores on each of them.

A Note on Revenue Multiples

Everything said above applies with equal force to revenue multiples, which are commonly used in software and high-growth businesses where EBITDA may be minimal or negative. Revenue multiples are even more sensitive to the specific factors described — growth rate, net revenue retention, gross margin profile, and contract quality — and are even more likely to be distorted by AI tools that synthesize from public software company data (which skews toward larger, higher-growth, more liquid businesses than the typical private mid-market software company).

KCENAV valuation assessments use structured scoring across the specific value drivers that affect multiples in your industry segment — not AI-generated estimates from an undisclosed data source. The goal is to give you a grounded, methodology-transparent view of where your business stands and what would move your valuation range in either direction.