Axios Finish Line: Go start a business
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Brendan Lynch/Axios
I'm offering you four specific ways to get more out of AI: better prompting, improving AI memory, starting a business using AI (tonight) and running a business using AI (Monday).
The barrier to starting a business isn't capital anymore. It isn't a team. It's just you.
- The old rule: Before you could launch, you'd need to assemble an army: lawyer, accountant, developer, designer, copywriter, researcher. This cost — and complexity — left a million ideas unborn.
- The new rule: Anyone with a strong idea and solid AI prompting skills can model and prep a new business in a weekend.
Why it matters: This is THE under-appreciated upside of the same AI boom that many fear will eviscerate existing jobs. It has never been this easy to start something with so little capital.
- The startup boom appears to be underway. 580,612 new businesses formed in March 2026, according to Registered Agents Inc.'s monthly Business Formation Report — a 14% year-over-year jump.
- The share of solo-founded startups climbed from 23.7% in 2019 to 36.3% by mid-2025, per Carta.
When we started Politico in 2007 and Axios in 2017, it took months to sketch it out, mock up designs and scrub legal obstacles to our names and business.
- We could do in hours now what would have taken weeks. You can too.
The new math, function by function:
- Legal scaffolding. Describe your setup to Claude or ChatGPT and get a plain-English LLC vs. S-Corp breakdown, a filing checklist and a draft operating agreement. You fact-check and fine-tune instead of paying a lawyer.
- Market research. Paste in the concept. Ask for the steelman case against it, the existing players, the pricing, the complaints. Build a customer survey in an afternoon. You talk it through with the AI instead of hiring a generalist researcher or firm.
- Financial model. Outline your business case out loud: how you will make money and spend it. AI builds the spreadsheets and forecasts. Stress test and iterate live with the LLM: "What assumptions are weakest? What threats or opportunities am I missing?" That's two weeks of junior analyst work done before lunch.
- Brand and copy. Tell the LLM about your customer — not the demographic, the person — and generate a logo, a design, a homepage, email sequence and pitch. All that used to cost a lot of money and take six weeks. Now, it requires your judgment over an hour.
- Product. Sketch out features and get working prototypes without writing code. Iterate even more simply with voice or text commands. You can easily change how it looks or works in minutes. If your business is more physical than digital, AI drafts your supplier outreach and negotiation scripts. Service? Positioning, packaging and pricing — all built by you and your AI.
- Support, onboarding, docs. Done before launch with a near-zero marginal cost.
The "what's left?" test: We've reached the decisions where the value of being human is everything. What remains are the parts that were always hard — the parts no tool can replicate.
- Judgment: Knowing what's worth building.
- Taste: The ability to discern "good enough" from "market-ready."
- Trust: The human-to-human connection that survives the bot-generated marketing.
- Resilience: The grit to keep going when v1 doesn't land. Startups are fun as hell, but still hard.
The bottom line: For 30 years, the excuse for not starting was the cost of starting. Consider that excuse expired.
📺 Worth watching ... theMITmonk: How I'd Build a 1-Person AI Business (0 to $1M+) ... Dan Martell: How to Build a $10M Solo AI Business ... Anik Singal: How To Build a One-Person Solo Business Using AI
If you start a business based on this column (and we hope you do!), shoot Jim a note: [email protected].
📈 If you're a CEO or on a CEO's team: Ask to join Jim's new weekly Axios C-Suite newsletter.
