An AI consultant,
available right now.

Early-stage decisions carry the most leverage and arrive when there is least money to spend on them. Entity type, pricing, first hire, the first contract - each compounds for years, and each costs several hundred dollars to ask an expert about.

So they get made on intuition, and the cost surfaces three years later as a restructuring, a repricing, or a clause nobody read.

A 10% price increase drops almost entirely to the bottom line. A 10% cost reduction does not, and it is far harder to find. Yet pricing is typically set once, early, by feel, and never revisited.

The Business Copilot will build the model - unit economics, margin, the sensitivity of the whole thing to a 10% move - and it will ask the question most people avoid, which is what the outcome is worth to the buyer rather than what it costs you to produce.

LLC, S-corp election, C-corp: the right answer depends on net income, whether you intend to raise, your state, and how much administrative overhead you will actually tolerate.

The honest version includes the costs. An S-corp election means payroll, a separate return, and a reasonable-compensation standard that the IRS does enforce. It is often worth it above roughly $80,000 of net income and it is not free, which is the part the confident articles leave out.

The expensive way to test a business idea is to build it. The cheap way is to try to sell it, and the gap between those two is where most of the wasted years in early-stage work live.

The reason people build first is that building is comfortable and selling is not. A codebase does not tell you it is not interested. Ten conversations with people who fit the profile will, and they will do it in a fortnight rather than in eight months.

The conversations have to be structured correctly to be worth anything. Asking whether someone would use a thing produces politeness. Asking what they currently do about the problem, what it costs them, what they have already tried and what they paid for it produces evidence. The single most useful signal is whether they have already spent money attempting to solve it - people who have are people with a budget line.

A pre-sale is the strongest validation available short of revenue. Someone who will pay before the thing exists is telling you something no survey can.

What this does not mean is that you should not build. It means that the version you build after twenty conversations is a substantially different and better-aimed product than the one you would have built after none, and the conversations cost a fortnight.

A business can grow and lose money on every transaction, and it is remarkably easy not to notice for a year.

The arithmetic is not complicated. What does it cost to acquire one customer, what does it cost to serve them, what do they pay, and how long do they stay. Four numbers, and most small businesses can only confidently state one of them.

Cost to serve is where the surprises are. It includes support, payment processing, hosting, the onboarding that somebody does manually, and the share of a founder week that goes into keeping existing customers rather than finding new ones. Leaving out the manual work is the standard error, because it feels like effort rather than cost right up until it needs hiring for.

Lifetime value is where the optimism is. Assuming a long retention on three months of data produces a number that justifies almost any acquisition spend, and it is the most common way a plausible business plan turns out to be wrong.

Run the four numbers pessimistically. If the business works on conservative assumptions it will work; if it only works on optimistic ones, you have learned something important for the cost of an afternoon rather than a year.

Every small company has them. The invoicing routine that only one person knows. The deployment step that is not written anywhere. The client relationship that exists in one inbox.

Each is a single point of failure, and the cost is invisible until somebody takes a holiday or leaves. Then it becomes two weeks of archaeology, performed under pressure, by people reconstructing decisions nobody recorded.

Documentation gets skipped because it is never the most urgent thing on any given day, and because writing it feels like describing something everybody already knows. Both are true and neither survives the first departure.

The version that actually gets done is small. One process, written as steps, by the person who does it, in half an hour. Not a wiki project, not a documentation initiative - one document, and then another one next month. Ten of those is most of a functioning operation.

The second-order benefit is more valuable than the insurance. Writing a process down almost always exposes that three of the steps exist for reasons nobody remembers, which is the cheapest process improvement available to any company.

The distinction between profit and cash is the single most consequential thing an owner can understand, and it is routinely learned the hard way.

A business can be profitable on paper and unable to pay wages, because profit records a sale when it is invoiced and cash records it when the money arrives. If customers pay in sixty days and suppliers require thirty, growth consumes cash - and the faster it grows, the more it consumes. That is why fast-growing companies fail, which sounds paradoxical and is arithmetic.

Inventory does the same thing more visibly. Stock is cash converted into something that cannot pay a bill, and an optimistic purchase decision is a cash decision wearing an operations label.

The practical instrument is a rolling cash forecast rather than a profit and loss statement. Thirteen weeks is the conventional horizon: what is coming in, what is going out, and what the balance is at the end of each week. It takes a spreadsheet and an hour, and it converts a vague anxiety into a dated number.

The levers, once visible, are unglamorous. Invoice immediately rather than monthly. Ask for deposits. Negotiate supplier terms, which is a conversation most owners never have. Chase receivables before they are overdue rather than after.

A credit facility arranged while things are comfortable is available; one sought while they are not, generally is not. That timing point is worth more than most financial advice.