Claude Enterprise for the seven-person GFG Securities team. Nothing more, nothing less. Live in one to two weeks.
Not an estimate. A ceiling we set ourselves.
$20 × 12 months × the tier's 20-seat floor = $4,800 a year, fixed. The seat is the license and the governance around it.
$45,000 ÷ 7 people = about $536 a month each. Sized for daily use as a working tool, not an occasional one.
All-in at the full cap. For context, one Bloomberg terminal runs about four times that per person.
A credit is simply pre-purchased work. Every time someone asks Claude to read, draft, review, or analyze, it draws down credits, the way a phone plan draws down data. We buy them upfront so Compliance can cap the total, and so the tool is ready the moment the team needs it. The seat unlocks the door. The credits are the actual work done inside.
Read a data room of hundreds of pages and surface the risks, the comparables, and the questions to ask, in an afternoon instead of a week.
Heavy draw. Large documents in, detailed analysis out. The single biggest use of credits.
Compare a Credit, HMA, or franchise agreement against our standard terms and flag every deviation before it reaches counsel.
Medium draw. Runs many times a week across the team, so it adds up steadily.
Turn a template and deal inputs into a first draft, then proofread the subscription booklet, so people edit instead of starting from a blank page.
Medium draw. Concentrated around live deals, lighter between them.
Build the Puntamar, Cacique, and Venado investor updates, and pull sector and tourism trends into a clean brief.
Lighter draw. Predictable, scheduled work that fills the quieter weeks.
This is why usage, not seats, is the larger number. Seven people using Claude as a daily working tool consume far more in work performed than in licenses. The $45,000 cap sizes that work for a busy year and guarantees it never exceeds it.
Usage is prepaid in credits and capped centrally, so the ceiling cannot be exceeded. If the team's workload ever justifies more, raising the cap is a deliberate leadership decision, not an overage.
The team already spends on individual Claude accounts today. That spend is unmanaged, expensed person by person, and invisible to Compliance. Enterprise replaces those accounts, so part of this budget is reallocation of existing spend, brought under supervision. The current figure is being pulled from expense records.
The Securities team, mapped. Nothing more, nothing less.
Pricing note: this tier bills at a twenty-seat minimum, so the seat fee is $4,800 a year fixed. We activate our seven. More seats would be a future decision.
Everyday work and live deal data never travel the same path.
The chat window the team types into for everyday work: research, comparables, drafting, contract review. Every session is governed by activity records, retention settings, and Compliance review.
Live deal data and MNPI never touch a chat window. They run system to system through the API, a direct pipeline, under Zero Data Retention: the vendor keeps nothing after the session ends.
Chat sessions never see live KYC, client account numbers, SSNs, or live MNPI. That work goes through door two.
The seven seats are the tool people use. The API is how we wire that intelligence into our own systems later.
A single proof-of-concept. One narrow task automated end to end, for example first-pass document review, to prove the value before we build wider.
One GFG system connected to Claude, so intelligence flows directly into an existing workflow rather than a person copying between windows.
Several systems working together. Sized only after a starter proves out, so we commit to the larger build with evidence, not a guess.
These are build estimates, not committed spend. Each would carry its own metered usage on top, and each is a deliberate future decision. Nothing here is needed to launch the seven seats.
We are not paying more. We are choosing the version built for books and records.
| What it gives us | Team | Enterprise |
|---|---|---|
| A FINRA exam asks for these four | ||
| A record of who did what, and whenAudit logs | ✕ | ✓ |
| Compliance can pull and review every conversationCompliance API | ✕ | ✓ |
| We decide how long conversations are keptCustom data retention | ✕ | ✓ |
| Seats switch on and off from our own directorySCIM provisioning | ✕ | ✓ |
| Additional Enterprise controls | ||
| We hold the encryption keys, not the vendorCustomer-managed encryption keys | ✕ | ✓ |
| Processing never leaves the United StatesUS-only inference | ✕ | ✓ |
Both tiers: the vendor never trains on GFG data, passes independent security audits, and uses our single company login. Only Enterprise carries the books-and-records controls.
Green goes first. Amber waits for Compliance. One gets reframed.
Set up. Compliance drafts the policy while the purchase completes. Both run in parallel.
Green work begins: research, comparables, contract review, closed-deal models.
Amber workflows pilot under Compliance review.
Market studies, hotel comparables, sector trends, due diligence on prospects, legal contract review, financial models on closed deals, internal knowledge base.
OM and Credit Memo drafting, subscription booklet automation, investor reports for Puntamar, Cacique, and Venado, shared contract repository, financial models on pipeline deals.
Reframed to AI-assisted form structure, with PII handled by humans. Full automation of live KYC data is not viable at our risk profile.
The Compliance Officer is involved at the design stage for amber items, not after the fact.
Four prerequisites, consistent with FINRA Notice 24-09. Examiners will ask for each.
Two pages. Defines who may use AI tools, what data classes are permitted as input, and how outputs are reviewed and retained.
No live KYC, no client account numbers, no SSNs, no live MNPI in chat sessions. Live MNPI work runs through the API with ZDR.
FINRA Rule 2210 principal pre-approval extended to AI-drafted communications with the public. Existing review workflow updated, not duplicated.
Workspace retention set per data policy. AI-assisted records flow into our existing 17a-4 archive: three or six years by record type, first two years easily accessible.
Seven seats, all Securities, via self-serve checkout. Seats fixed at $4,800 a year. ZDR terms pursued with Anthropic sales for the data pipeline.
Two-page document, two weeks. Runs in parallel with the purchase.
From decision to people working: one to two weeks.
Cookies on GFGSecurities.com
At GFG Securities, LLC, we use cookies and other technologies to enhance your experience and provide you with our services. These cookies are used for functional, analytical, and advertising purposes. By using our website, you consent to the use of cookies in accordance with our Privacy Policy.
Please take a moment to review our Privacy Policy for more detailed information about how we use cookies and other tracking technologies.
If you choose to decline the use of cookies, please be aware that some features of our website may not function properly, and your information won’t be tracked for analytical and advertising purposes. However, a single cookie will be used in your browser to remember your preference not to be tracked.
By continuing to use our website, you acknowledge that you have read and understood our use of cookies and agree to our Privacy Policy.