Section 1
Why a description is not a pitch
The standard elevator pitch describes a category and a position inside it. We are a workflow platform for mid-market finance teams. That sentence is accurate, forgettable, and does no work, because the listener has to construct the situation, the problem and the stakes themselves, and they are not going to do that for a stranger. A story does the construction for them. It puts one recognisable person in one recognisable moment with one thing going wrong, then shows what changed. Put someone specific in front of the listener and they stop parsing your category and start checking whether they know that person. Often they are that person. That is the moment the pitch is actually competing for, and it arrives in the first two sentences or not at all. The broader case for founders owning this personally is set out in [Why Storytelling Is a Critical Leadership Skill for Founders](/blog/why-storytelling-is-a-critical-leadership-skill-for-founders).
Section 2
The four beats that fit in forty seconds
You need four beats, and they survive interruption in this order. A person in a situation, named by role and context rather than by segment. The moment it goes wrong, told as an event rather than a condition. What you do, stated as a mechanism a competitor could not honestly claim word for word. What changed afterwards, in the listener's units: days, currency, headcount, error rate. What is deliberately absent is your funding, your team size, your technology stack and your mission. Those are answers to questions that have not been asked yet. If the four beats work, the questions arrive on their own, and answering a question you were asked always lands better than volunteering the same content unprompted.
Section 3
Building your version
The table below takes the four beats and asks for one line each, plus the failure signal that tells you the beat is not doing its job. Write it for a single audience first. A pitch that tries to serve investors, buyers and candidates at once serves none of them well.
Section 4
Rehearsal that actually improves it
Do not rehearse the delivery. Rehearse the interruptions. Ask a colleague to cut in after your second sentence with the three questions you hear most, then continue from wherever you were stopped. A pitch built to be recited collapses the moment it is broken into, and it will be broken into. Then collect data for two weeks. After every delivery, write down the first question the listener asked. Twenty entries is enough to see the pattern. Recurring questions about the mechanism mean beat three is vague. Recurring questions about pricing before anything else usually mean the situation was too generic to be worth understanding, so the listener defaulted to the only dimension they could evaluate. Questions that repeat your own words back to you are the good sign: it means the story travelled intact. What that looks like with an investor audience specifically is covered in [Telling Your Story to Investors: What Works, What Doesn't](/blog/telling-your-story-to-investors-what-works-what-doesn-t).
Section 5
The three that kill it
False precision is the first. A claim to cut costs by 47 percent invites an interrogation of your methodology before the listener has decided whether they care. Round it down, name the customer type, and let the specificity live in the situation instead of the outcome. The second is the unearned superlative. Leading, best-in-class and next generation carry no information and cost you credibility with exactly the listeners you want. The third is scope creep, where the pitch expands each quarter as the product does until it describes everything and lands on nothing. That version tends to appear right when the company starts growing quickly, which is also when it hurts most, a pattern [Scaling AI Automation as Your Startup Grows](/blog/scaling-ai-automation-as-your-startup-grows) traces on the operations side.