Plain-language definitions of the deal terms a founder selling an AI company, or a buyer looking at one, actually meets — from LOI and earnout to ARR, tail period and covered parties.
A contract in which each side promises to keep what it learns confidential and to use it only to assess the deal. It is normally the first document signed, because nothing useful about a company can be shared before it exists.
A one- or two-page anonymous summary of a company for sale: sector, size, rough financials, why it is attractive — but not the name. It is sent to possible buyers to find out who is interested before any identity is revealed.
CIMConfidential Information MemorandumDeal process
The full written case for buying a company — what it does, how it makes money, its customers, its financials and its risks. It goes to buyers who have signed an NDA, and it is the document most first offers are based on.
An early, non-binding note from a buyer saying roughly what they would pay and on what terms, usually as a range. It is a way to sort serious buyers from curious ones before anyone spends money on diligence.
A written outline of the deal a buyer intends to do: price, structure, timetable and conditions. Most of it is non-binding, but a few clauses usually are — exclusivity and confidentiality in particular — so it is signed with legal advice, not as a formality.
A promise by the seller to stop talking to other buyers for an agreed period, so the one buyer can spend money on diligence without being outbid mid-way. Also called a no-shop. It is the point where a seller loses competitive leverage, which is why its length is negotiated carefully.
The buyer checking that the company is what it was said to be — financial, legal, technical, commercial. Findings do not only kill deals; more often they change the price, the structure, or what the seller has to promise in the contract.
The controlled place where a seller puts the documents a buyer needs — contracts, accounts, cap table, IP assignments. Access is granted per person and usually tracked, so the seller can see what has been read and by whom.
The binding contract that actually transfers the business. It sets the final price and its adjustments, what each side promises is true, who carries which risk afterwards, and what must happen before completion.
The seller’s list of exceptions to the promises made in the contract. If the contract says "there is no litigation", this is where an existing case is named. Disclosing something properly generally removes it as a basis for a later claim.
The moment ownership actually changes hands and the money moves. It often happens some weeks after signing, once the conditions in the contract — approvals, consents, financing — have been met.
What the business itself is worth, independent of how it is financed: the value of the operations before counting cash in the bank or debt owed. Most multiples are quoted against it because it lets two companies be compared without their balance sheets getting in the way.
What the owners actually receive: enterprise value, plus the cash in the business, minus the debt it owes. This is the number a founder should look at, because a headline price can shrink considerably by the time debt is settled.
A valuation expressed as a multiple of annual revenue. Common for software companies that are growing rather than maximising profit, because their profit figure understates what the business is worth.
A valuation expressed as a multiple of annual EBITDA. The usual measure for profitable businesses, since it prices what the company earns rather than what it sells.
Valuing a company by looking at what similar public companies trade at. Quick and widely understood; its weakness is that a private company of a different size, growth rate and customer base is rarely as comparable as the label suggests.
Valuing a company by what buyers actually paid for similar companies. Closer to reality than public trading multiples because those prices include the premium a buyer paid for control — but the data is thinner and often older.
Valuing a company by forecasting the cash it will produce and reducing each future year to what it is worth today. Rigorous in form, but the answer moves a long way on small changes to the assumptions, so it is usually a sanity check rather than the number.
The last twelve months of actual results, whatever the financial year happens to be. Buyers price on TTM rather than on a calendar year because it is the most recent full year of evidence.
A recent short period projected out to a full year — most often the latest month multiplied by twelve. Useful for a fast-growing company, and easy to flatter with one good month, so buyers usually check it against TTM.
The annualised value of subscription revenue that repeats — contracted and expected to continue. One-off fees, services and usage that is not committed do not belong in it, though they frequently get counted anyway.
EBITDAEarnings Before Interest, Taxes, Depreciation and AmortisationMetrics
A measure of operating profit that strips out financing, tax and accounting charges for past spending, so two businesses can be compared on how well the operations themselves earn.
What is left of revenue after the direct cost of delivering the product. For AI companies this is where inference and model-serving costs land, which is why their margins can look unlike traditional software.
What last year’s customers are worth this year, counting upgrades, downgrades and cancellations but no new customers. Above one hundred per cent means the existing base grows on its own — the single number buyers most often ask for first.
The rate at which customers or their revenue leave. Worth asking which is being quoted: losing many small accounts and losing one large one produce very different pictures from the same headline figure.
A rough test for software companies: growth rate plus profit margin should reach forty. It is a shorthand for the trade-off between growing and earning, not a law, and it means little for a company early enough that both numbers are volatile.
How much cash the company consumes each month. Net burn — spending minus income — is the figure that matters, because it is what actually shortens the runway.
How long the company can operate before the money runs out: cash divided by net burn, usually expressed in months. It sets the deadline for a raise or a sale, and it is the first thing a buyer works out for themselves.
What it costs, on average, to win one customer — sales and marketing spend divided by customers gained. Read next to how much that customer is worth over time, and how long it takes to earn the cost back.
The total gross profit a customer is expected to produce before they leave. Highly sensitive to the churn assumption behind it, so a confident LTV built on thin retention data is worth little.
The buyer acquires the company itself by buying its shares, so contracts, employees and liabilities move with it. Usually simpler for the seller and the outcome most founders expect.
The buyer acquires chosen assets — code, customers, brand, sometimes staff — rather than the company. It lets a buyer leave unwanted liabilities behind, which is exactly why sellers usually prefer a share purchase.
Part of the price paid later, only if the business hits agreed targets. It bridges a disagreement about what the company is worth, and it is the most frequent source of post-deal argument — which is why how the target is measured matters as much as its size.
Part of the price left as a loan from the seller to the buyer, repaid over time with interest. The seller is effectively financing part of their own exit and ranks behind the banks if things go wrong.
The seller keeps a stake in the business under its new owner instead of cashing out fully. Common where the buyer wants the founder to stay invested in what happens next.
Part of the price held by a neutral third party for an agreed period, available to the buyer if the seller’s promises turn out to be wrong. Released to the seller if nothing surfaces.
A correction to the price at completion so the business is handed over with a normal level of day-to-day operating cash. It routinely moves the final number, and it surprises sellers who assumed the headline price was the price.
The register of who owns what: shares, options, convertibles and the terms attached to each. A messy or contradictory cap table is one of the most common reasons a deal slows down.
The right of certain investors to be paid before ordinary shareholders when the company is sold. In a modest exit it can absorb most of the proceeds, so founders should model their own outcome after preferences, not before.
Shares or options earned over time rather than granted at once, usually with a cliff before any of it is earned. Buyers look at it closely, because it determines who is tied to the company after the deal.
A short summary of the terms an investor proposes — valuation, amount, control rights. Mostly non-binding, but it sets the shape of everything that follows, so what is conceded here is rarely won back later.
An investment that converts into shares at a later priced round rather than buying shares now. It avoids agreeing a valuation early; the caps and discounts on several stacked SAFEs decide how much a founder is actually giving away.
A loan that is expected to convert into shares at a future round instead of being repaid. Unlike a SAFE it is debt, so it carries interest and a maturity date — and if the round does not arrive, that date still does.
Pre-money is what the company is agreed to be worth before new investment; post-money is that plus the money invested. Which one a quoted valuation refers to changes how much of the company the founder keeps, so it is always worth asking.
The reduction in an existing shareholder’s percentage when new shares are issued. Not automatically bad — a smaller share of a larger company can be worth more — but it compounds quietly across rounds and option pools.
A short raise meant to carry the company to a larger round or to an exit. Usually raised from existing investors and on faster terms, because it exists to buy time.
The investor who sets the terms, does the deepest diligence and usually takes the largest share of a round. Other investors often wait for a lead before committing, which is why a round without one can stall.
An existing investor’s right to put more money into future rounds to keep their percentage. Valuable to investors and worth counting for founders, since it reduces the room available to new ones.
The engagement under which an adviser is appointed to sell a company: what they will do, for how long, and how they are paid. Everything else on this list — retainer, success fee, tail — is a term inside it.
A mandate under which only one adviser may run the sale for its term. Advisers ask for it because the work is front-loaded and easy to lose to a rival who did none of it.
A fee paid during the engagement regardless of whether the company sells, typically monthly. It funds the work up front and is often credited against the success fee at completion.
The adviser’s payment on a completed sale, normally a percentage of the price. What counts as "the price" is worth pinning down: whether deferred amounts, earnouts and assumed debt are included changes the fee materially.
A window after the mandate ends during which the adviser is still owed a fee if the company is sold to a buyer they introduced. It stops a seller from waiting out the contract to avoid paying for the introduction.
The agreed list of buyers an adviser introduced, which the tail period applies to. Keeping it explicit and approved as the process runs is what prevents an argument about who found whom after the deal is done.