07 · Engagement & Next Steps

How we would work together, and what happens in the first 90 days

A concrete plan for turning this proposal into a launched product: engagement options, a 30/60/90-day plan with deliverables, the first week day by day, who to hire and when, how we operate, and the questions I need you to answer.

Commercial terms to be scoped togetherCode and accounts owned by the company from day 1Weekly live demos

1 · Engagement options

How I would engage

You are hiring a CTO for a company that does not exist yet as a product. The right shape of engagement depends on your runway, timeline and how much you want to de-risk the relationship before committing. Here are four options. Commercial terms for each are to be scoped together.

Option	Scope	Time commitment	Cadence	Best when
A · Advisory CTO	Architecture, roadmap, vendor and partner selection, hiring support, technical diligence. You hire or contract the builders.	About 1 day a week	Weekly 60-minute session plus async review of ADRs and pull requests	You already have an engineering team or agency and need senior direction
B · Fractional build lead	Everything in A, plus I write a large share of the MVP and lead the first engineer. Owns delivery of the 90-day plan.	3–4 days a week	Weekly demo, weekly metrics, daily async update	You want speed and a working product before committing to a full-time hire
C · Full-time founding CTO	Full technical ownership: product engineering, security, compliance infrastructure, hiring and the technical side of fundraising.	Full time	Embedded in the founding team	Funding is in place and you want a long-term technical co-owner
D · Build lead, then convert (recommended)	Start as B for 90 days against the plan below, then convert to C if the milestones are met and we both want to continue.	3–4 days a week, then full time	As B, with a formal day-90 review	You want evidence of fit and delivery before a long-term commitment, on both sides

Pricing, equity and term are deliberately not shown. They depend on your funding position and are to be scoped together.

2 · First 90 days

30 / 60 / 90 days, with concrete deliverables

Each phase ends with something you can see, use and show to others, not just a status report. Dates move only by explicit decision, recorded in the decisions log.

Days 1–30

Decide and de-risk

Every big decision written down, the riskiest assumptions tested, the foundation in place.

Product brief and success metrics agreed with the founding team (one page, signed off).
15–20 customer interviews across Maya, Dev, Priya and Ben, recorded and synthesised.
First 10 Architecture Decision Records: stack, auth, data model, AI provider and routing, broker rail, content CMS, archive, analytics.
Securities counsel engaged; written memo on the education-only, publisher's-exclusion posture in progress.
Production repo, CI, preview environments, error tracking and product analytics live. Prototype code triaged into keep / rewrite / discard.
AI coach guardrail pipeline v1 (input classifier, retrieval over vetted lessons with citations, output checks) and the first 150 eval prompts running in CI.
Append-only archive of every coach turn from the first real conversation.

Days 31–60

Build the core loop

A private alpha in which real people learn, practise and set a goal end to end.

Private alpha with 50–100 hand-picked users from the interview pool and the founders' networks.
First 10–12 lessons in the CMS with mastery checks, reviewed by the content lead and counsel.
Simulator v1 on paper-trading rails with drawdown scenarios, diversification score and behavioural nudges. No trading rewards.
Goal planner with ranges rather than point forecasts, plus disclosure templates for hypothetical projections.
Eval suite grows to 300+ prompts: advice-bait, prompt injection, vulnerable-user scripts.
SOC 2 readiness started on a compliance-automation platform; written information-security programme drafted.
Weekly metrics dashboard: activation, D7, lessons mastered, coach refusal accuracy, cost per conversation.

Days 61–90

Prove it and plan the raise

Evidence a founder can take to investors and partners.

Closed beta with paid plans live: Free and Plus ($8/mo or $72/yr, annual by default).
Eval pass rate of 98% or better on advice-bait prompts, the gate for a broader launch.
One B2B2C pilot (employer or credit union) signed or in final negotiation, with per-seat pricing as an explicit learning goal.
Broker-partner due diligence complete (Alpaca as default; one alternative assessed) and a recommendation for Phase 2.
Counsel memo finalised; marketing claims reviewed; AI disclosures and point-of-delivery disclaimers in place.
Phase 2 plan, budget and hiring plan presented as a board-ready pack, with D30 retention and conversion data from the beta.

Why the compliance work starts in month one

Impersonal education is protected by the publisher's exclusion,^[1] but tailored coaching risks being treated as personalised advice.^[2] Counsel needs to see the coach design before it reaches real users.
FINRA treats AI chatbot sessions as firm communications that must be supervised and retained.^[3] Retrofitting an archive is far harder than starting with one.^[4]
EU AI Act Article 50 disclosure applies from 2026-08-02,^[13] and the FCA now expects disclaimers at the point AI financial information is delivered.^[12] Building it in from the start makes later markets cheaper.

What it costs to run in the first year

The AI coach runs at about $0.035–0.04 per conversation with model routing on list prices, about $150 a month at 1k monthly active users.^[15]
A SOC 2 compliance platform runs about $7.5–15k a year^[7], and a Type II audit about $12–20k for a company our size.^[6] Broker partners will ask for it in Phase 2.
The full reference stack is about $2–3k a month at 1k MAU (see the Architecture document for 10k and 100k).

3 · Week one

The first five days

Momentum in week one sets the tone for the engagement. This is what you would see.

Mon
Founder working session
Half-day session on vision, non-negotiables, target launch date and constraints. Walk the prototype together and capture every reaction.
Tue
Access and foundations
Company-owned accounts for code, cloud, domains and analytics. Repo, CI and preview deploys stood up. Nothing lives on my personal accounts.
Wed
Research kick-off
Interview guide and screener live; first 5 interviews booked. Counsel shortlist (2–3 securities firms) sent to the founders.
Thu
Decisions on paper
First 5 ADRs drafted and reviewed. Draft 90-day plan with owners and dates. Founding-engineer job description drafted.
Fri
First weekly demo
30-minute live demo of what changed this week, followed by the decisions log and next week's plan. This cadence starts on day 5 and does not stop.

4 · Team plan

Who to hire, and when

A small, senior team beats a large, junior one at this stage. The sequence below assumes option D and a U.S.-first, education-only Phase 1. I would run every technical hire's process; you would meet every finalist.

Role	When	Why this role, why then	Profile
Securities counsel (outside, fractional)	Week 1–2	The education-only posture, coach design, disclaimers and marketing claims need sign-off before real users. Counsel is also needed for broker-partner contracts in Phase 2.	Securities firm with Investment Advisers Act and broker-dealer experience; fintech clients
Founding engineer	Month 1–2	Doubles build capacity and removes single-person dependency on me. Owns the simulator, data pipelines and integrations.	Senior full-stack TypeScript, Postgres, product-minded; has shipped consumer apps; comfortable with AI tooling
Content lead / Head of Learning	Month 1–3	Owns the curriculum the coach retrieves from, the brand voice and the review queue. Without this role the AI has nothing trustworthy to stand on.	Credentialed (CFP or CFA preferred) educator who writes plainly and has run content production
Product designer	Month 2 (contract), full time by month 4–6	Activation and retention decide this business. Onboarding, lesson UX and the simulator need a dedicated design owner.	Consumer mobile and web, strong in behavioural design, research-literate, anti-dark-pattern
Growth / community lead	Month 4–6	Once activation is proven: search, newsletter, video and the creator programme, all within compliance review.	Content-led growth background in a regulated category
Second engineer (AI quality and data)	Month 6+	Eval harness, red-teaming, analytics and personalisation as usage grows.	LLM evaluation, data engineering, security mindset
Chief Compliance Officer	Phase 3 only	Required only if WealthCreate registers as an adviser. Until then, counsel plus the partner's compliance team cover it.	Registered-adviser compliance experience

5 · Operating cadence

How we would run the work

Light process, high visibility. You should never have to ask what is happening.

Weekly live demo (30 min)

Working software, not slides. What shipped, what we learned, what we decided, what is next. Recorded for anyone who misses it.

Architecture Decision Records

Every significant technical decision written down in the repo: context, options, decision, consequences. New hires and investors can read why the system is the way it is.

Weekly metrics review

Activation, D7 and D30 retention, concepts mastered, first-action rate, coach refusal accuracy, guardrail incidents, cost per conversation. Finance-app D30 is 2–7% at the median, so that is the bar to beat.

Monthly risk and compliance review

Eval results, flagged conversations, content review backlog, security findings and any regulatory changes. Quarterly external red-team of the coach.

Daily async update

Three lines in a shared channel: done, next, blocked. No standing daily meeting with the founders.

Monthly customer day

The whole team watches real users try the product. Nothing changes a roadmap faster.

Retention benchmark: Business of Apps finance-app data.^[10] Red-teaming approach: Promptfoo-style adversarial suites.^[11]

6 · Open questions

Ten questions for the founding team

These are the answers that change the plan. I would like to work through them in our first working session; several have a recommended default so we are never blocked.

1
Where do we launch first, and where do you need to be within two years?
U.S.-first is my recommendation. UK and EU each add their own promotion, AI-disclosure and licensing regimes, and that changes counsel, data residency and partner choice.
2
What are your licensing ambitions: stay education-only, embed investing via a partner, or become a registered adviser?
This decides how much regulatory infrastructure we build in the first year and whether we need a CCO on the roadmap.
3
What funding and runway do we have, and when is the next raise?
The 90-day plan should end in the evidence your next round needs. Runway also sets how fast I hire.
4
Who owns the curriculum and the credentialed voice?
The AI coach is only as trustworthy as what it retrieves. A CFP or CFA on the team, a licensing partner, or commissioned content are very different cost and IP positions.
5
Is WealthCreate the brand, or a working name?
Domain, trademark search and app-store names take weeks. I would start the search in week 1 if the name is not settled.
6
B2C first or B2B2C first?
My recommendation is direct-to-consumer with one employer or credit-union pilot. If you have a warm channel partner, the order may flip, and so do the first hires.
7
Do you want to hold assets (AUM) eventually, or stay subscription-led?
AUM changes the partner choice (self-directed rail vs managed portfolios), the regulatory path and the valuation story.
8
What is the target launch date, and what is fixed about it?
A date tied to a raise, a partner or a seasonal moment such as tax season or a Trump Account milestone changes scope decisions.
9
Is a teen or family product in scope, and when?
Adults-only avoids amended COPPA obligations at launch. A teen product needs a custodial-account partner and separate compliance work.
10
How do you want to make decisions with your CTO?
Who has the final call on product, brand and spend, and what level of detail do you want each week? Clear decision rights matter more than any framework.

Context: Trump Accounts went live on July 4, 2026^[14]; amended COPPA compliance has been required since 2026-04-22^[9]; Alpaca's Broker API supports both paper trading and real accounts^[5]; non-bank fintechs are covered by the FTC Safeguards Rule.^[8]

7 · Risks to the engagement

What could go wrong between us, and how we prevent it

Most CTO engagements that fail do so for predictable reasons. Naming them up front is the cheapest fix.

Risk	What it looks like	Mitigation
Scope drift	Every idea becomes a Phase 1 feature and the launch date slides.	One-page product brief, a written 90-day plan, and a decisions log. New scope replaces existing scope; it does not add to it.
Slow founder decisions	Engineering waits on brand, content or partner choices.	Every open question has a recommended default and a decide-by date. If a date passes, the default ships and can be revisited.
Regulatory surprise	Counsel reads the coach design late and forces rework.	Counsel engaged in week 1–2; coach design and disclaimers reviewed before any real user sees them.
Content bottleneck	The platform is ready but there are not enough reviewed lessons.	Content lead hired early; AI-assisted drafting with human and compliance review; launch with a tight 10–12 lesson spine, not a library.
Key-person dependency	Knowledge sits in my head and my accounts.	Company-owned accounts from day 1, ADRs for every major decision, documented runbooks, and a founding engineer by month 2.
Prototype mistaken for product	Demo-grade shortcuts end up in production.	Explicit keep / rewrite / discard triage in month 1; production code goes through review, tests and security checks.
Misaligned expectations	Different views on hours, pace, or what “done” means.	Written engagement scope, a formal day-90 review, and either side free to change the arrangement at that point.

8 · Next step

Let's talk

The fastest way to judge whether I am the right CTO is to spend 45 minutes together walking through the prototype and the questions above. Bring your hardest objections. If it is a fit, the week-one plan starts the following Monday.

Ready when you are.

Chris Therriault · ServiceVision. Full-stack and AI platform builder; I write the code, ship the product, and own the outcome.

Email chris@servicevision.net Open the prototype

Back to the Executive Summary Try the AI coach