Frictionless
AI-powered capital intelligence platform aligning startups, investors, and accelerators through structured evaluation systems.
Project Context
TIMELINE
2024–present (~18 months active design; ongoing)
STATUS
Live — first institutional partner cohort onboarding
WHAT EXCITED ME
Designing the explainability layer that makes an AI score feel legitimate to the founder whose company it's evaluating
WHAT I LEARNED
That trust in AI is an interface problem, not a model problem — the gap between a score and a decision is the most important space to design
Overview
Frictionless evaluates startup readiness against institutional investor criteria, generates a structured score, and surfaces actionable improvement pathways — functioning as an evaluation infrastructure layer, not just a matching tool.
Frictionless is an AI-powered capital intelligence platform that evaluates startups across product, traction, team, and market dimensions — aligning founders with investor thesis criteria and providing structured pathways for improving readiness beyond the funding moment.
As Chief Product Officer, I lead product strategy, system architecture, and all design execution — from AI scoring logic through dual-sided dashboard design and conversational onboarding — in a lean 0-to-1 environment.
Problem
Investors, startups, and accelerators were each making high-stakes decisions with inconsistent data and no shared framework for evaluating readiness.
Early-stage capital allocation is fragmented. Investors rely on inconsistent criteria. Startups lack structured insight into their gaps. Accelerators can't identify high-fit candidates at scale.
The challenge wasn't building a matching tool — it was designing a scalable evaluation system that could structure qualitative startup data, translate investor thesis criteria into machine-readable logic, and generate transparent readiness signals. The opportunity was an institutional-grade intelligence layer that the ecosystem didn't yet have.
Approach
I structured Frictionless as four sequential layers — data extraction, readiness evaluation, thesis alignment, and growth workflows — each designed to make the next layer more useful.
I structured Frictionless as a capital intelligence architecture composed of four core layers.
AI Data Extraction & Structuring
Automated ingestion of pitch decks, websites, and startup data into structured evaluation fields.
Readiness Evaluation Layer
Defined a proprietary Readiness Score framework translating qualitative signals into quantifiable metrics across product, traction, team, and market criteria.
Thesis Alignment Engine
Architected matching logic aligning startup profiles with investor thesis parameters, generating structured compatibility percentages.
Growth & Performance Layer
Designed task and improvement workflows enabling startups to increase readiness scores and drive operational performance beyond funding.
Guiding principles:
Make evaluation transparent and explainable
Translate subjective criteria into structured systems
Reduce friction in institutional workflows
Design scalable architecture from inception
The system was built to support iteration while maintaining institutional-grade rigor and clarity.
Key Design Decisions
Three decisions shaped the system: making scores explainable rather than just accurate, designing separate information architectures for each side of the platform, and solving the cold-start problem through AI-powered data extraction.
Making AI scoring explainable, not just accurate. Early in the process, the natural instinct was to optimize the Readiness Score for precision — surface the most accurate signal and let the output speak for itself. I pushed back on this. For institutional capital partners, a score without transparency is a score without trust. The critical design decision was to architect the scoring layer as an explainable system — surfacing not just a number, but the underlying signals and gaps driving it. This made the platform usable for investors who needed to defend their screening decisions internally, not just act on them.
Designing for both sides of an inherently asymmetric relationship. Startups and investors have fundamentally different mental models, vocabularies, and goals. A dashboard that works for a founder evaluating their own readiness looks nothing like one that works for a fund manager screening a cohort. Rather than defaulting to a single unified interface, I architected dual-sided dashboards with distinct information hierarchies — while ensuring the underlying data model remained consistent. The constraint that forced this was the realization that building for "both" without acknowledging the asymmetry produces something that works well for neither.
Sequencing onboarding to reduce the cold-start problem. A matching platform is only as useful as the quality of data it has to work with. The risk at launch was that founders would submit incomplete or surface-level information, degrading match quality and eroding investor confidence early. I designed the AI-powered onboarding to extract and structure data from existing materials — pitch decks, websites, prior documents — before asking founders to fill anything in manually. This reduced the activation burden and improved data completeness from the start.
Impact & Outcomes
The platform is live with its first institutional cohort, with the full evaluation infrastructure, scoring framework, and dual-sided dashboard architecture built from zero.
The platform is live and actively onboarding its first institutional partners, with an initial pilot cohort underway across investors and accelerators.
Established structured AI evaluation framework for early-stage capital ecosystems
Enabled transparent Readiness Score modeling across startup cohorts
Defined thesis-alignment matching logic for investors and accelerators
Structured improvement pathways driving startup performance beyond capital access
Positioned Frictionless as an institutional-grade evaluation infrastructure
The platform created a scalable foundation for aligning startups and capital through structured intelligence rather than manual review alone.
What I'd do differently
I'd have run structured feedback sessions with founders in the first onboarding cohort before finalizing the readiness scoring interface. We correctly prioritized the investor-facing experience early — they're the platform's buyers — but founders are the raw material the system depends on, and their interpretation of the scoring interface differed from how we'd designed it to communicate. We caught and corrected this during early onboarding, but earlier testing with that user group would have been cheaper and more definitive.
Structured Readiness Score framework translating qualitative startup signals into quantifiable evaluation metrics.
Thesis alignment engine matching startups to investors through structured compatibility modeling.
Conversational AI onboarding transforming unstructured pitch materials into structured evaluation data.
Growth workflows enabling startups to improve readiness scores and operational performance beyond capital introduction.
Let’s build intelligent systems together.
Open to senior product design and leadership opportunities across AI, infrastructure, and complex platforms.