Last post was the agent that reads the tax code. This one covers its neighbor: “small business capital,” the agent I asked to find money for a business of one.
My business is a sole proprietorship. It does software development, consulting, Amazon publishing, and crypto and stock trading. Revenue swings from about $10,000 to $100,000 a year, and lately it’s been on the low end. That’s a hard profile to fund, and I wanted to know exactly how hard.
The rule it set for itself #
The first report came back on Sept 30. Before any numbers, it explained how to read them:
- Verified (official): it read the figure on the bank’s or agency’s own page that day.
- Third-party: it came from a deal or review site because the official page wouldn’t load.
- UNVERIFIED: it couldn’t confirm it.
- Estimate: its own math, labeled as such.
Then one line I wish every research report had: “I did not make up any figures.” Each report ends with a list of everything it couldn’t verify.
I didn’t ask for that system. It’s the most useful part of the work, because it tells me which numbers I can act on and which ones I have to check first.
Most of the answers were “not you” #
A lot of the research was ruling things out, with a reason for each:
- Online lenders: one line of credit requires an LLC or corporation, so a sole proprietor is out. Another wants $100,000 a year in revenue. A nonprofit lender wants $300,000 a year and two years in business.
- The expensive ones that would say yes: the agent found one lender’s average line-of-credit APR was 59.8%, and another’s calculator assumed 3.3% interest a month. Its advice was to avoid those unless the spending clearly pays for itself.
- Local programs: a nearby city’s grant had closed and required a physical location. That city’s microloan requiresone full-time job created for every $35,000 borrowed , which is a tough rule for a business with no employees.
- State programs: the big California grant had no open round, and its tax credit requires job creation.
None of that is bad news, exactly. It’s the list of doors I’d otherwise have spent an afternoon knocking on.
The red flag in my own business #
The most useful finding was about trading.
The agent quoted the SBA rule that makes “speculative businesses” ineligible for SBA loans (13 CFR 120.110). One nonprofit lender it checked excludes crypto businesses outright. One online lender lists cryptocurrency as an ineligible industry. And the business credit card terms it read say crypto purchases don’t count toward sign-up bonuses.
It also pointed out that trading gains don’t show up the way lenders want. Lenders look at business revenue and business bank deposits, and for most people trading gains are reported separately from business income.
Its advice: apply as the software, consulting and publishing business, keep trading out of the business description, and keep it out of the business bank account. A separate business checking account isn’t just tidy. One lender it found asks for three months of business bank statements.
What actually fits #
After the “not you” list, a short list was left. These are the figures as of Sept 30; offers change, so the agent said to recheck each one on the day I apply:
- A $10,000 grant from Verizon’s Small Business Digital Ready program. You unlock the application by finishing two free courses, and there’s no storefront, revenue or EIN minimum stated.
- Kiva U.S. loans of $1,000 to $15,000 at 0% interest , with no fees. Crypto businesses are excluded, so again, apply as the software or consulting business.
- An SBA microloan through a Los Angeles nonprofit lender: $5,000 to $50,000 at 7.5% to 9%.
- A 0% intro APR business credit card. The agent’s bottom line, labeled as its own estimate: at my revenue, a 0% card is the cheapest unsecured credit I can get. Chase says sole proprietors can apply with an SSN and that revenue can be $0. It also flagged that the 0% ends after 12 months, then rates run roughly 17% to 29%.
- A grant of up to $4,000 from the National Association for the Self-Employed, if the membership pays for itself.
It also found business checking sign-up bonuses, but I covered bank bonuses in post 4, so I’ll skip them here.
Building business credit from zero #
The agent laid out a plan in order, with the cost and wait time for each step:
- Get a free EIN from the IRS. It takes minutes online. It said to never pay a third-party site for one.
- Get a city business license , and a county fictitious business name filing if I use a trade name.
- Open business checking and run only business income through it.
- Request a free D-U-N-S number , which can take up to 30 business days.
- Open two or three net-30 supplier accounts that report to business credit bureaus , buy only what I need, and pay early.
Then it gave a realistic timeline, labeled as its own estimate: about 6 to 12 months before there’s enough history to support a microloan application. And it was blunt that personal credit and cash flow would still drive most approvals.
The license I didn’t know I needed #
I work from home, and I’d never thought about a city license. The agent read my city’s municipal code and quoted the line that makes it unlawful to carry on any business in the city without one. Home-based businesses count.
The good news was that it’s a flat tax, not one based on revenue: a $40 base for a solo operator with no employees, so my revenue swings don’t change it. City fees come on top, and it marked the current amounts as unverified because it couldn’t open the newest fee schedule.
The federal R&D money, and why I’m not describing my idea #
A week later, on Oct 7, I asked about one more option: America’s Seed Fund, the federal SBIR program for small-business R&D. I have an early-stage logistics idea, and I wanted to know if it could qualify.
The numbers it pulled from the NSF’s official solicitation:
- Phase I is up to $305,000 for 6 to 18 months. NSF takes no equity.
- A sole proprietorship is eligible. No LLC required.
- You can’t apply directly. You first send a short “Project Pitch,” and NSF typically takes 1 to 2 months to respond.
- The pitch limits are strict: two per company in any 12 months, and three ever for the same project.
Then it told me what I didn’t want to hear. The next deadline, Nov 4, wasn’t realistic: the pitch response alone takes longer than the four weeks left. It pointed me to Mar 4, 2027 instead. It quoted NSF’s own list of what it won’t fund, including “software implementation” and “application of existing technologies,” and said a plain app version of my idea would be declined. It searched the full SBIR award dataset for prior art and found work close enough that reviewers would know it. And it said a one-person, idea-stage team is a weakness, so I should line up technical advisors before pitching.
One more piece of advice is why this section is vague: it suggested filing a provisional patent before talking about the idea publicly. So I’m not.
What I learned #
- Tagging every claim beats a confident answer. “Verified,” “third-party,” “unverified” and “estimate” told me which numbers to trust and which to check, without rereading every source.
- The “no” list is worth as much as the “yes” list. Most of the value was ruling out lenders and programs before I spent time on them.
- Look at your own business the way a lender would. Trading is part of how I make money, but for funding it’s a red flag. The agent spotted that before any application did.
- A good research agent tells you to slow down. Wait for the right deadline, don’t waste a pitch, protect the idea first.
This isn’t financial or legal advice, and every figure here was as of the date the agent checked it.
— Blake (written through my AI agents)