Growth Stack puts AI to work for your marketing and operations inside the guardrails of your AI policy. It has four parts: AI twin videos, automated newsletters, CRM automation and AI assistant setup. The order matters. We set the guardrails first, then build the tools inside them, so the policy you sign matches the tools your team actually uses.
Pricing: from $5,000 setup, plus from $1,000/mo. Scope, timeline and final price are set in a written proposal after you send the intake.
How a Growth Stack engagement runs
- Take stock. We inventory where AI is used today (staff tools, vendors, marketing) and where borrower data touches it. This is the same AI use inventory that Fannie Mae Supplement Section 4 and Freddie Mac Guide Section 1302.8 expect you to keep.
- Write the policy. You get a mortgage-specific, sign-off-ready AI policy aligned to Fannie Mae and Freddie Mac, with a named owner and review cadence. If you already have one, we map each Growth Stack tool to it.
- Put AI to work. We build the content, newsletter, CRM and assistant pieces inside the guardrails you just set. Every automation is mapped to your policy, owner and vendor review from day one.
Everything runs by email. You send a 10-minute intake about your team, licensing, LOS and the AI tools you use today. We reply with a proposed scope and timeline, then deliver recorded video walkthroughs and your deliverables by email. A live call is optional.
The four parts, and the questions to settle before each goes live
Each part below starts with what it is, then lists the governance questions your AI policy should answer before launch. The legal rules quoted are general information, not legal advice. Have your compliance reviewer approve any borrower-facing content.
1. AI twin videos
An AI twin is a synthetic video likeness of a real person, usually a loan officer, that can present scripted content without a new recording each time. It saves filming time. It is also exactly the kind of technology Freddie Mac now asks you to train staff about: Guide Section 1302.2 says annual security training "should incorporate ... artificial intelligence (AI)-powered tactics (e.g., deepfakes, targeted phishing content)".
Questions your policy should answer before the first video goes out:
- Has the person whose likeness and voice are used given written consent, and what happens to the likeness if they leave the company?
- Who approves each script before a video is generated, and is that approval recorded?
- Will videos be labeled as AI-generated, and where?
- Where does the video vendor store the likeness and voice data, and can you have it deleted? Add the vendor to your AI vendor due diligence list.
- Does any script state a rate, payment or loan term? Regulation Z says that if an advertisement states a rate of finance charge, "it shall state the rate as an 'annual percentage rate,' using that term". It also lists triggering terms, such as the amount of any payment or the number of payments, that require additional disclosures.
2. Automated newsletters
AI can draft and schedule newsletters for your borrower and referral partner lists. The drafting is the easy part; the rules for commercial email and mortgage claims still apply to every send.
- Does every commercial email carry what the CAN-SPAM Act requires: identification that the message is an advertisement or solicitation (unless the recipient gave prior affirmative consent), notice of the opportunity to opt out, and "a valid physical postal address of the sender"?
- Are opt-out requests honored on time? The Act makes it unlawful to keep sending covered messages "more than 10 business days after the receipt of such request".
- Who reviews AI-drafted copy for mortgage claims? Regulation N prohibits material misrepresentations in any commercial communication about a mortgage credit product, including about rates, fees, whether a term is "fixed", affiliation with a government entity, and whether a consumer "has been preapproved or guaranteed".
- Which list data is sent to the AI tool that drafts the newsletter, and is that tool on your approved list?
3. CRM automation
CRM automation handles follow ups and routine tasks: reminders, status nudges, task routing and drafted replies. These automations usually run on borrower data, so they sit squarely inside your information security program.
- Which CRM fields does each automation read, and does any of it go to an AI model? Fannie Mae Section 4 says "SI outputs derived from Confidential Information should be treated as Confidential Information", so an AI summary of a borrower record is as sensitive as the record.
- Can an automated message reach a borrower without a person reviewing it first? Decide which messages need human review and write it down.
- Is the AI feature inside your CRM covered by your third-party risk review? A Freddie Mac FAQ says the third-party risk management program must address the risk "if the service provider uses artificial intelligence (AI)/machine learning (ML)".
- Who owns each automation, and how are changes approved? Section 4 requires change management for SI models, systems and use cases.
4. AI assistant setup
We set up AI assistants for your team for everyday work such as drafting, summarizing and research, configured to match your approved AI tools list.
- Which assistants are approved, for which tasks, and which data may never be entered into them?
- Has every person who uses an assistant signed your AI policy? The compliance portal tracks staff sign-offs, free for up to 10 staff.
- Is each assistant listed in your AI use inventory, with an owner and a review date?
Governance first
We start with your AI policy. If you do not have one yet, we write it with you, with a named owner and review cadence, then build the Growth Stack inside the guardrails you just set. Read what an AI policy for mortgage brokers should cover, or start with the free mortgage AI policy template.
Growth Stack does not replace your compliance review of advertising, and BrokerLumen is not a law firm. It gives you the policy, inventory and review steps so that review has something to check against.
Pricing
From $5,000 setup, plus from $1,000/mo. Final pricing depends on your team size, vendors and scope. Every purchase has a 14-day refund window; see our Terms of Service.
Need governance only, without the marketing build? See Governance Done-For-You, from $3,500.
How to start
Send a 10-minute intake. We reply by email with next steps and a recorded walkthrough, no call needed. If you would like a live conversation, you can book an optional evening or weekend video call.
Frequently asked questions
How much does Growth Stack cost?
From $5,000 setup plus from $1,000 per month. Final pricing depends on your team size, vendors and scope.
How fast is Growth Stack delivered?
Timelines depend on your team size, vendors and scope. You receive a proposed timeline by email after we review your intake.
Is an AI twin video a deepfake?
It uses the same kind of technology, which is why consent, script approval and labeling belong in your AI policy before the first video goes out. Freddie Mac Guide Section 1302.2 names deepfakes as an AI-powered tactic to cover in annual security training.
Do AI-drafted newsletters need anything special?
The same rules apply as for any commercial email and mortgage advertising: the CAN-SPAM Act requirements, including an opt-out notice and a valid physical postal address, and the Regulation Z and Regulation N rules on mortgage claims. Have a person review AI-drafted copy before it is sent.
Do I need an AI policy before Growth Stack?
Yes. Growth Stack starts with your AI policy. If you do not have one, we write it with you first, with a named owner and review cadence.
Do I need to get on a call?
No. Submit the intake form. Questions are handled by email and recorded video walkthroughs, with an optional evening or weekend video call.
What is the refund policy?
Every purchase has a 14-day refund window, including Growth Stack services. Email us within 14 days of the charge for a full refund.