It's 9 PM and You're Not a Lawyer: The Small Business Owner's Playbook

It's 9:14 PM. The kids are finally asleep. The dishwasher is running. You have a twenty-page vendor contract open on your laptop because the account exec wants it signed "by end of week" and this is the first hour you've had to look at it.

You are, in the legal-and-strategic-counsel sense, completely alone. Your CFO is you. Your general counsel is you, approximately, if you squint. Your contract review team is also you, plus whatever you can remember from the last one of these you signed.

There's a clause on page 7 about auto-renewal. There's something on page 12 about limitation of liability that might matter. There's a "Term and Termination" section on page 14 that you should probably read twice. There's an indemnity section on page 17 that says "indemnify" nineteen times. There's probably something important on page 19 too but honestly at this point you are skimming.

Every small business owner has this Tuesday. Vendor contracts. Commercial leases. NDAs from potential customers who want you to sign before they'll even get on a call. SaaS terms of service that are a hundred pages because they include the TOS of seven sub-processors. You have a company to run. You are not a lawyer. You also can't afford to miss the one clause that costs you thirty grand in eighteen months.

Here is what actually works.

The move: stop reading front to back

The mistake every owner makes at 9 PM on a contract is starting on page one. Contracts aren't written to be read linearly. They're written to be reviewed by someone who already knows what to look for.

You don't. That's fine. The tool knows what to look for — AI contract review points out the risky clauses it matches, though a clause it doesn't list is not proof the contract doesn't have one. You just want the list, in plain English — and a way to check anything that surprises you: Chat names the source document, and the page, slide or section when CorpGPT can place it.

The playbook

Step 1. Drop it in

Open CorpGPT. Drag the contract in. If there are exhibits, schedules, SLAs, or an Order Form — drop them in too. If your own Master Services Agreement template is the one being "accepted" here, drop that in as well so the tool can compare.

While it processes, refill the water glass.

Step 2. Get the plain-English summary

Open the Make pane ("Make from these sources"). Ask for:

• Plain-English summary (Summary) — one page on the terms it finds: what you're buying, what they're providing, how long, what it costs, how it renews, how you get out, who owns what, and who's liable for what. Written for a human, not a lawyer.
• Red flag list (Contract review) — the clauses it finds that a reasonable SMB shouldn't accept without negotiation or at least knowing about: auto-renewal, uncapped liability, unilateral price increases, unilateral termination, indemnities that flow only one direction, jurisdiction locks, IP or data ownership issues.

Two outputs. Each reads the passages that best match it, so treat them as a map of where to look — then read those clauses.

Step 3. Interrogate Chat

Open Chat on the same screen. Ask the questions a careful business owner wants answered before signing:

• "What happens if we need to cancel six months in? What notice do we have to give, and are there any penalties?"
• "Can the vendor raise prices mid-term? Is there a cap?"
• "What's our total exposure if something goes wrong? Is it capped?"
• "Who owns the data we put into this tool? What happens to it if we leave?"
• "If they breach the contract, can we terminate? What's the cure period?"
• "If we breach, what can they do? Same cure period?"
• "Is there an auto-renewal? What's the notice window and when does it trigger?"
• "Is there anything in here about exclusivity or non-compete? How broad?"
• "Where do disputes get resolved? Is there a mandatory arbitration clause?"

Chat names the source document, and the page, slide or section when CorpGPT can place it. You click through to verify anything that matters.

You now know where the clauses that matter are.

Step 4. Decide the three things

Your decision list:

1. Sign as-is — once you've read the clauses it pointed to and found nothing you'd want a lawyer to see.
2. Sign with edits — there are one to three clauses you want to push back on. Ask CorpGPT: "Draft a one-paragraph reply to the account exec asking for these three changes, in a friendly professional tone." Send it.
3. Stop and call a lawyer — the deal is bigger than you thought, the exposure is real, or there's something in here (non-compete, exclusivity, personal guarantee) that is above your pay grade to decide alone. Put it down. Email your attorney in the morning.

For a standard vendor contract, you'll often land on (1) or (2). For the rest — acquisitions, major leases, partnership deals — you've now saved your lawyer's time and money by going in with specific questions instead of "please read this."

Step 5. Sign or don't

Either way: you've read the clauses that matter. You didn't gamble at 9 PM on a Tuesday.

Beyond the one contract

Build a contract library

Drop every vendor contract, SaaS agreement, NDA, commercial lease, and employment agreement into a collection in My Documents — "Active Contracts — 2026." Every renewal, every new vendor, every "hey, what did we sign with that supplier?" moment is one question to Chat.

Ask Chat to find anything

Describe what you're looking for — "the SaaS tool with the 60-day cancellation window", "the lease with the CPI escalator" — and Chat searches for the passages that best match and names the document they're in. If it finds nothing, look again yourself before you conclude it isn't there.

Run a renewal audit quarterly

Once a quarter, go through the collection one contract at a time and ask: "Does this contract auto-renew? What's the notice window to get out? Is there a price escalator, and when does it trigger?" Chat reads the passages that best match each question, so asking about one contract at a time keeps each answer focused.

You find out about the auto-renewing $800/month SaaS you forgot about before it renews. That one audit pays for a year of CorpGPT by itself.

The features doing the work

Make pane — plain-English summaries, red-flag reviews (Contract review), renewal summaries. They land in Things I've made.

Chat across everything — ask any question about any contract you've uploaded, get an answer that names the source document, and the page, slide or section when CorpGPT can place it; find contracts and clauses across your whole library by intent, not exact keyword; draft the reply to the account exec. No Ctrl-F. No re-reading clauses you already read.

My Courses — when a vendor drops a new product agreement that's actually complex (think distribution agreements, revenue-share deals, licensing), turn it into a short course — eight steps per module, from a summary to a quiz — and read it a module at a time before the call.

Record audio — record an in-person meeting (with consent); it uses your device's microphone, so for a phone or video call, upload the audio file your call tool saved instead (MP3, M4A…, up to 50 MB; audio, not video files). The Transcript tab shows who spoke. Cross-reference what they promised on the call against what's in the signed contract.

Why this actually works

Three forces are doing the real work.

First, red flags are pattern-matching. The AI recognizes the common shapes: what a standard auto-renewal looks like, what an unusual one looks like, and what a predatory one looks like. You don't have to know the difference in advance. The tool tells you where to look, and Chat names the source document, and the page, slide or section when CorpGPT can place it.

Second, citations make it checkable. Chat names the source document, and the page, slide or section when CorpGPT can place it. You push back on the vendor with "clause 8.2 says X" instead of "the contract says somewhere that…". Vendors take the first one seriously. They ignore the second one.

Third, your contract library finally compounds. Every new vendor contract makes the next review faster because you now have your precedent searchable. "Last time we signed a cloud-hosting agreement, we negotiated the SLA to 99.95% — let's do that again here" is one quick question, not a "where did we put that thing?"

What this can't do

Be honest about this.

CorpGPT is not your lawyer. It doesn't give legal advice. It doesn't represent you in a negotiation. It doesn't know your risk tolerance, your business's specific regulatory environment, or the fact that this vendor sued their last customer. For anything material — an acquisition, a major lease, a partnership, a personal guarantee, anything with the word "exclusive" in it — get a lawyer.

What it does is the reading and the finding: the plain-English summary of what the contract says, the clauses worth a second look, the "what does it say about raising prices mid-term?" question, the renewal dates. That's the grunt work. Whether a clause is acceptable is a legal judgment — yours, or your lawyer's. The judgment calls, the strategic trade-offs, the high-stakes negotiation — that's you (or your lawyer, when the stakes justify it).

The bottom line

Vendor contract. Twenty pages of legal. It's 9 PM. You are definitely not a lawyer.

Now you know what's in it. Now you sign with your eyes open, or you push back with specifics, or you stop and call a lawyer before you sign something you shouldn't have.

Sign smart. Sleep soundly. Run the business — not the filing cabinet.

Open CorpGPT. Drop the contract in. Go to bed.

---

Keep reading

• 200-Page Contract at 4:45 on a Friday — how lawyers find the clauses that matter in a weekend ambush. The professional-grade version of this same playbook.
• Tax Code Changed. Again. — how accountants translate new tax rules into client-ready answers they can check against the source documents.
• Two Minutes Before the Client Call — how financial advisors prepare for client calls from the fund documents they uploaded.
• Eighty Pages of Policy. Find the Clause. — how insurance agents find a clause in a long policy form and read what it says.
• Ten Minutes Between Sessions — how therapists study their training material and CE and draft psychoeducation handouts, with client information kept out.
• Beat the Binder: The Nurse's 12-Hour Shift Playbook — how nurses study unit policies and protocols with cited answers.
• Doctors: Study Your Guidelines, Draft the Admin — how physicians study the guidelines they uploaded and draft the non-patient admin, away from patient care; answers name the source document, and the page, slide or section when CorpGPT can place it.
• Find the Guideline Behind 'Will I Qualify?' — how mortgage brokers find the guideline section behind a client's question and read it; the lender makes the call.
• Turn One Chapter Into a Week of Lessons — how teachers build a full week of materials from a single source chapter.
• Your Client Wants 30 Listings by Saturday — how real estate agents review 30 disclosure PDFs.
• Cram 400 Pages Before Morning — the student side of the same playbook.
• Client Deck Tomorrow. It's 8 PM. — how consultants synthesize four reports into a draft deck and pressure-test the story.
• Another Meeting to Recap the Last Meeting — how project managers turn meeting recordings into the summary, decisions, and action items.
• Answer Once. Help Forever. — how HR managers answer handbook questions with cited answers.
• Prep the Hiring-Manager Intake — how recruiters turn the JD and scorecard into a reading checklist and prep the intake; the resumes and the decisions stay with the recruiter.
• How Are You Different From [Competitor]? — how sales reps prepare cited battlecard answers before the call and draft the follow-up after it.
• Brand Says Bold. Legal Says Careful. — how marketing managers draft campaigns and prepare them for brand and legal review.
• Seventeen Transcripts. One Deadline. — how writers turn a pile of interview transcripts into themes, pulled quotes, and a working outline.
• Ninety-Minute Interview. Show Notes by Tuesday. — how indie podcasters draft the post-production package from one episode's transcript.
• Sixty Videos In. Did I Already Post This? — how content creators turn their back catalog into a searchable content brain.
• Sixty Papers. One Lit Review. One Weekend. — how researchers turn a stack of PDFs into summaries, themes, citations, and gaps.
• Three Investor Meetings. Three Different Decks. — how founders tailor a pitch per investor.
• 47-Page Brief. Proposal Due Friday. — how freelancers turn dense briefs into a first proposal draft.
• Your Kid Just Asked About Algebra. — how parents help with homework using the actual textbook.
• Medicare Sent a Letter. 12 Pages. Tiny Font. — how older adults turn any letter into a plain-English one-pager.
• See all use cases — small business owners, therapists, nurses, doctors, insurance agents, advisors, accountants, lawyers, brokers, agents, teachers, students, and more.
• Plans & pricing — a free Trial to start, no credit card required.