Luma Fintech — AI Strategy for Structured Products Marketplace
Full boardroom debate on Luma Fintech's comprehensive AI strategy covering competitive positioning, 12-month execution plan, resourcing, and monetization strategy for the structured products marketplace.
Participants
Chief Strategy Officer
Chief Financial Officer
M&A Advisor
Chief Marketing Officer
Chief Technology Officer
Venture Capital Investor
General Counsel
Questions (2)
Luma Fintech is a structured products marketplace platform. We need the boardroom to debate and produce a comprehensive AI strategy. Specifically: (1) Current State Assessment — what is the state of AI in fintech today, particularly for structured products? What are competitors doing? (2) Market Positioning — where should Luma position itself with AI relative to Halo Investing, Simon Markets, and traditional bank platforms? What is the defensible AI moat? (3) Comprehensive AI Strategy — what should Luma's AI strategy look like across the value chain (product creation, pricing, distribution, advisor tools, analytics, compliance)? (4) 12-Month Execution Plan with quarterly milestones, headcount, budget, outcomes, and risks.
Data-First, Compliance-Embedded, Build-Heavy AI with One Selective Acqui-Hire
How should Luma best monetize AI? Three competing models: (1) Stickiness play — AI bundled free for retention. (2) AUM/volume uplift — AI drives more transactions at existing pricing. (3) Premium tier — AI features priced separately. Which model (or combination) maximizes enterprise value? What does advisor willingness-to-pay look like? How should monetization be sequenced?
Prove, Then Price — Three-Tier Model with 'Gate Outputs, Not Inputs'
Luma Fintech is a structured products marketplace platform. We need the boardroom to debate and produce a comprehensive AI strategy. Specifically: (1) Current State Assessment — what is the state of AI in fintech today, particularly for structured products? What are competitors doing? (2) Market Positioning — where should Luma position itself with AI relative to Halo Investing, Simon Markets, and traditional bank platforms? What is the defensible AI moat? (3) Comprehensive AI Strategy — what should Luma's AI strategy look like across the value chain (product creation, pricing, distribution, advisor tools, analytics, compliance)? (4) 12-Month Execution Plan with quarterly milestones, headcount, budget, outcomes, and risks.
Conclusion
Data-First, Compliance-Embedded, Build-Heavy AI with One Selective Acqui-Hire
The board recommends a phased, build-heavy AI strategy investing $10-15M over 12 months with 12-14 new hires and one targeted acqui-hire ($3-5M). The strategy is anchored on Luma's $300B+ proprietary structured product dataset as the defensible moat, with compliance positioned as a competitive differentiator rather than a blocker. The flagship product is an AI Structuring Copilot for advisors, launched first to the RIA channel (lighter regulatory burden) in Q2, expanding to the BD channel in Q3 after compliance frameworks are in place, and to Europe in Q4 ahead of EU AI Act deadlines.
Consensus
- +AI is a genuine strategic lever for Luma — the $300B+ structured product lifecycle dataset is a rare, proprietary asset no competitor can replicate
- +The AI Advisor Copilot is the highest-value, highest-urgency AI use case — every member independently identified this as the flagship product
- +Compliance is not a blocker — it is a feature. 'FINRA-ready, EU AI Act compliant' is a competitive differentiator neither Halo nor iCapital can currently claim
Dissenting Views
Maintains that one strategic acquisition ($15-25M range) of a derivatives pricing AI company would accelerate Luma's roadmap by 12+ months. Argues the board is underweighting competitive risk — if iCapital deploys AI across the combined SIMON platform, Luma's window closes. Strength of reasoning: moderate. The competitive risk is real, but the CFO and CTO's counterarguments on integration cost and data moat are stronger.
Worried that a phased, conservative approach produces a 'boring' Series D narrative. Wants at least one 'hero feature' shipped and generating metrics within 6 months to anchor the next fundraise. This is a legitimate concern — addressed by the RIA-channel copilot launch in Q2.
Key Tensions
Budget Discipline vs Acquisition Ambition
Conservative organic investment ($8-12M)
Phased organic build preserves financial flexibility, targets 12-18 month payback, avoids integration risk and forced fundraise scenario
Aggressive acquisition strategy ($35-50M)
Speed to capability through acquisition, arbitrage window on AI-fintech multiples, competitive urgency before iCapital/Halo move
Board leans toward CFO's position. M&A's plan too capital-intensive for Luma's stage. One targeted acqui-hire ($3-5M) justified for structured products AI talent.
Build vs Buy
Build in-house on the data moat
$300B data moat is non-transferable — any acquired model needs full retraining on Luma's data anyway. Building at $2.5-3.5M/year is cheaper and produces purpose-fit AI.
Acquire AI capabilities externally
Organic hiring takes 6-9 months. Acquisition buys 12-18 months of velocity. Competitive window is closing.
CTO's position wins. Data moat argument is decisive. Build-first with selective acqui-hire as talent accelerant, not product acquisition.
Speed to Market vs Regulatory Compliance
Launch fast, compliance catches up
Speed-to-market critical for Series D narrative. AI features as competitive weapons in sales cycles immediately.
Compliance-first architecture
FINRA Reg BI and EU AI Act create hard constraints. BD compliance teams will block non-compliant AI tools regardless of quality.
CMO co-opted GC's argument — compliance IS the competitive moat. Three-wave launch: education tools first (low risk), assist-mode second (human-in-loop), autonomous third (post-compliance cert). GC and CMO fully aligned.
Strategic Sequencing of $63M
AI-first allocation
AI compounds — invest first in what has non-linear returns. Geography scales linearly.
Balanced allocation across AI, insurance, geography
Don't over-concentrate. Maintain optionality across all three vectors.
CSO's framework adopted: $15-20M AI (compounds), $15-20M life insurance (extends, already committed), $15-20M geography (seed $5-8M, hold rest). $8-13M reserve.
Individual Positions
How should Luma best monetize AI? Three competing models: (1) Stickiness play — AI bundled free for retention. (2) AUM/volume uplift — AI drives more transactions at existing pricing. (3) Premium tier — AI features priced separately. Which model (or combination) maximizes enterprise value? What does advisor willingness-to-pay look like? How should monetization be sequenced?
Conclusion
Prove, Then Price — Three-Tier Model with 'Gate Outputs, Not Inputs'
The board achieved its first unanimous position: a phased hybrid model starting with Model B (volume/AUM uplift) at existing pricing, transitioning to a three-tier premium structure after compliance certification. The breakthrough insight was the VC's 'gate outputs, not inputs' reframe — all users feed the data flywheel (free AI inputs), but premium analytics outputs are monetized. This preserves the network effect while creating a separable revenue line for Series D.
Consensus
- +Model B (volume/AUM growth at existing pricing) is the right starting position — lowest risk, fastest to market, measurable
- +Basic AI features must be bundled free — they are table stakes within 18-24 months
- +Premium tier is the right long-term destination but premature launch triggers AI-washing enforcement and fiduciary liability
- +The commodity vs proprietary distinction drives the free/paid split — commodity LLM features free, proprietary $300B dataset analytics premium
Dissenting Views
Remains concerned that 9 months of free AI with no direct revenue creates a 'show me the money' problem for Series D. Proposes compromise: publish AI usage metrics and declared pricing roadmap by Month 6 so investors can model revenue potential. 'The intent to monetize is almost as powerful as actual revenue at Series C/D stage.'
Argues free-first approach makes Luma a less attractive acquisition target in months 6-9. Bloomberg or SS&C would prefer to see revenue-generating AI. Concedes regulatory incident would be more damaging than delayed monetization.
Key Tensions
Free AI Strengthens Moat vs Premium AI Essential for Valuation
Free AI builds unassailable network effect
Premium fragments the flywheel, trains on only 30% of users, creates competitive vulnerability. Free AI drives 100% adoption and compounding data advantage.
Premium AI revenue essential for exit multiples
Investors and acquirers want AI-attributed ARR. Bundling free makes AI an invisible cost center. NRR >120% with premium tier commands 2-3x higher multiples.
VC's 'gate outputs not inputs' reframe resolves both: universal free AI inputs (preserves flywheel) + premium AI output monetization (creates revenue line). Hybrid approach satisfies both sides.
Free-Tier Economics — CFO vs CMO
Free AI is financial negligence
Requires halving churn to justify $10-15M — unprecedented. $2-4M/yr perpetual COGS. Maximum 90-day trial.
Free AI drives land-and-expand at 13:1 CAC ratio
Free tier runs budget models at $300-500K/yr, not $2-4M. 10K users at 8% conversion = $4M revenue against $300K compute.
CMO wins on economics (cost was overestimated 5-10x). Compromise: permanent bounded free tier at $300-500K/yr cap, 90-day Pro trial, premium Intelligence tier recoups costs with margin.
Regulatory Risk of Premium AI — GC vs VC/CFO
Premium AI creates fiduciary liability and AI-washing risk
Two-tier suitability problem, SEC enforcement trend, Investment Advisers Act obligations when charging for advisory-adjacent AI.
Premium pricing is standard SaaS tiering, not advisory
Bloomberg model — tiered analytics without fiduciary liability. SEC cases targeted fraud, not legitimate pricing.
GC wins the timeline — premium tier gated behind compliance certification (Month 9-12). VC/CFO win the destination — premium tier IS the right model once compliance infrastructure exists. Both satisfied.