Two-day MVP
LineRunner
A live MVP built to let people try the product for free, register demand for a paid version, and convert market feedback into a sharper build decision.
Visit LineRunnerJason Thomas Ferrell
AI strategy and execution.
For leaders deciding what to build, buy, fund, or greenlight, Jason turns complex AI opportunities into working software, agent systems, diligence artifacts, and film production workflows.
Two-day MVPs
LineRunner shows how quickly a product can get into users' hands, gather demand, and shape the paid build.
Full products in under a month
Elenqa, Screenvox, and the enterprise contract intelligence system show complete workflows, billing, and production-shaped architecture at unusual speed.
Enterprise systems inside one month
The contract intelligence build reached roughly 250K lines across bulk clause analysis, private MCP tools, and AWS deployment architecture.
Early OpenAI product lineage
GPT-3 legal reasoning work before enterprise LLM product patterns were obvious.
Selected clients and collaborators
Work across AI filmmaking, early OpenAI productization, leading private equity value creation, enterprise AI, legal evidence, and high-stakes finance.
For CEOs, investors, and product teams
AI strategy and systems.
Diligence, product strategy, agents, document intelligence, secure builds, and full production systems for teams deciding what to build, buy, fund, or ship.
See AI systems workFor producers, directors, studios, and brands
AI filmmaking.
Commercials, proof-of-concept videos, shorts, remakes, animatics, and feature workflows directed with a dedicated AI film production system.
Go to the cinema pageStrategy backed by products that actually ship.
The consulting value is not just advice. It is the ability to test an AI product thesis in days, turn user feedback into an MVP, and scale the right idea into production software.
Under 100 hours
Elenqa
A full AI debate learning product with teacher and student workflows, realtime voice, transcripts, grading, pricing, and customer-facing launch surface.
Visit ElenqaOne-month enterprise build
Enterprise Contract Intelligence Infrastructure
A roughly 250K-line enterprise contract intelligence system with ingestion, bulk clause and term extraction, redlines, private MCP tools, governance surfaces, and AWS-native deployment architecture.
One-month product
Screenvox
A screenplay-to-table-read product with character voices, synced animatic stills, billing, and customer workflows. The product ultimately evolved into Slate.
AI Systems
AI strategy that can survive diligence and become software.
For CEOs, PE/VC managing directors, and product teams: assess where AI creates or threatens enterprise value, then build the prototypes, agents, document intelligence, and secure deployments needed to test and capture it.
01
2026
Enterprise legal AI infrastructure
Enterprise Contract Intelligence Infrastructure
A secure contract intelligence system for bulk clause analysis, grounded legal questions, private MCP access, and AWS-native deployment inside customer-controlled environments.
Role
Fractional Chief AI Officer, product architect, solo implementation lead, AWS deployment architect.
Stakes
PraxisIQ needed a product that could satisfy enterprise security expectations, deploy inside customer AWS accounts, and answer complex legal questions across a company's agreements.
Built
Built in roughly one month as a production-class system: ingestion, indexing, bulk clause and fact extraction, redline workflows, RBAC, governance surfaces, authenticated MCP tools, and an AWS marketplace deployment model.
Why it matters
The system shows what enterprise AI buyers need: secure workflows, grounded answers, clear controls, agent-accessible tools, and a deployment model that respects customer data.
- Private MCP tools for contract questions, redlines, audits, renewals, costs, queues, summaries, facts, and evaluations.
- Lower-cost OSS models handle baseline analysis; frontier models handle deeper reasoning when the workflow justifies the cost.
- Designed for bring-your-own-cloud AWS deployment with enterprise controls and customer-specific isolation.
02
2025-2026
Private equity value creation
AI Diligence & Portfolio Strategy
AI strategy analysis for a leading private equity firm and portfolio-company contexts: where AI threatens a company, where it can expand the product line, and where existing distribution becomes an advantage.
Role
AI strategist, technical diligence partner, product architecture advisor.
Stakes
The useful question is concrete: what could an AI-native competitor build against this company, and what should this company build first because it already owns the customer channel?
Built
Delivered analysis across potential investments and existing portfolio companies, mapping disruption risk, product opportunity, data and workflow advantages, defensibility, execution paths, and the operating assumptions behind AI value creation.
Why it matters
This work gives investors and operators a clearer view of AI risk and AI upside before capital, management attention, and engineering teams get committed.
- Assessed AI-native competitive threats against incumbent software categories.
- Identified product opportunities that fit existing distribution channels and customer relationships.
- Translated market and technical analysis into concrete roadmap and value-creation recommendations.
03
2025-2026
AI filmmaking system
Slate
An AI director's assistant for turning scripts and source films into shot language, references, continuity, reels, and generated scenes.
Role
Founder, product architect, full-stack builder, AI director, workflow designer.
Stakes
AI video becomes production technology when the director can control sequence logic: characters, locations, coverage, continuity, camera grammar, cost, sound, and iteration.
Built
Slate combines screenplay parsing, scene automation, shot planning, provider orchestration, visual runs, continuity checks, score generation, source-motion windows, media QA, and ongoing OpenUSD scene-engine work.
Why it matters
For producers and directors, Slate turns AI video from prompt experiments into a controlled development and production workflow.
- Designed for collaborative creative workflows with durable production structure.
- Used with professional filmmakers and investor-backed development.
- Supports the film proof-of-concept work shown in the cinema section.
04
One-month build
Screenplay-to-table-read product
Screenvox
A full product for uploading a screenplay, casting character voices, generating a performed table read, and syncing it with an animatic made from generated stills.
Role
Founder, product architect, full-stack builder, AI workflow designer.
Stakes
Writers, producers, and directors need to hear a script and feel its pacing long before a full production exists. Screenvox turned a screenplay into something closer to a pitch asset: performed, visualized, and easy to review.
Built
Built in about one month with screenplay parsing, character extraction, voice casting, dialogue and action rendering, generated looks for characters, locations, and props, still-image animatic sequencing, audio synchronization, billing, and a customer product surface.
Why it matters
Screenvox ultimately evolved into Slate. It proved the screenplay-to-production workflow, then exposed the business constraint: voice-generation economics made price competition hard against better-capitalized competitors, so the opportunity moved toward deeper filmmaking control.
- Turned uploaded screenplays into multi-character table reads with differentiated voices.
- Generated character, location, and prop looks, then synced still-image animatic beats to the read.
- Included billing and customer workflows; the product was retired as the learning moved into Slate.
05
2020-2022
Early GPT-3 legal AI lineage
CS DISCO + OpenAI
Early GPT-3 work on legal evidence, factual support, contradiction, entailment, and document-grounded reasoning.
Role
Principal data scientist, OpenAI implementation partner, legal-AI product strategist.
Stakes
Legal AI has to connect claims to evidence spans, handle contradiction, and create lawyer-grade traceability across large document sets.
Built
Agreements with both CS DISCO and OpenAI included one of OpenAI's earliest partner agreements with a revenue-share structure. The work explored GPT-3, legal entailment, Enron annotations, contradiction datasets, zero-shot experiments, DeBERTa fine-tuning, and proof-table workflows.
Why it matters
This is where the later agent work started: evidence-first reasoning at scale, before LLM product patterns were obvious.
- Direct work inside the early OpenAI partner ecosystem before GPT-3 became a public market category.
- Research artifacts across contract NLI, evidence evaluation, contradiction detection, grounded summaries, and proof tables.
- The pattern carried forward into Narrative Equity, Slate, and enterprise contract intelligence infrastructure.
06
2024-2025
Omnichannel voice agents
Echobach
A hardware-backed voice-agent platform for realtime audio, telephony, messaging, calendar workflows, memory, and long-running orchestration.
Role
Founder, AI architect, hardware/software/product builder.
Stakes
Useful voice agents have to survive interruptions, route tools, preserve context, resume later, use phone and calendar systems, and behave like operational assistants.
Built
Echobach evolved from a Raspberry Pi push-to-talk device into a configurable agent platform spanning OpenAI Realtime, Twilio voice, SMS, email, contacts, RAG, workflow logs, scheduler resumes, and user-defined tools.
Why it matters
It pushed toward the kind of assistant people actually want: one that can speak, call, schedule, research, remember, and follow through.
- Physical device work across audio client, Wi-Fi setup, battery telemetry, GPIO controls, and enclosure prototypes.
- Realtime voice workflows with phone-number management, interruption handling, and tool routing.
- Explorations included multi-party scheduling, research, outreach, and persistent follow-through.
07
2023-2024
Agentic investment assistant
Narrative Equity
An early agentic market assistant connecting tool use with portfolio state, filings, news, fundamentals, analytics, forecasting, and paid workflows.
Role
Founder, architect, product builder, quantitative workflow designer.
Stakes
Investment workflows demand tool use, traceability, market data, portfolio context, risk awareness, and judgment about when language output is useful.
Built
Narrative Equity integrated GPT-style function calling with AlphaVantage, Finnhub, SEC data, company fundamentals, news sentiment, technical indicators, portfolio analytics, optimization routines, and Stripe-based paid flows.
Why it matters
It translated evidence-grounded reasoning into a market workflow before agents became a default product category.
- Function-calling architecture across market APIs, filings, portfolio state, and analytics.
- Classical finance concepts sat beside LLM reasoning: volatility, covariance, efficient frontier, risk metrics, and forecasts.
- Established a product pattern for later enterprise and creative agent systems.
08
2023-2024
High-stakes finance AI
Millennium Partners
AI architecture and strategy work for language data, internal workflows, and finance contexts where model output touches real operational risk.
Role
AI strategy advisor, architecture consultant, evaluation and workflow partner.
Stakes
In high-stakes finance, retrieval quality, latency, privacy, evaluation, data rights, and domain judgment matter as much as model capability.
Built
NDA-limited advisory work covered language data workflows involving newswires, filings, research reports, portfolio workflows, internal AI systems, and strategy selection.
Why it matters
The work sits at the level where AI architecture has to withstand capital, confidentiality, speed, and institutional scrutiny.
- Advised on language-alpha and internal AI workflow architecture.
- Connected hands-on system design with executive build-buy-evaluate decisions.
- Operated in a setting where vague AI claims have little value.
09
2025
Post-essay education AI
Elenqa
A realtime AI debate product that lets students demonstrate understanding through argument, rebuttal, evidence, and recorded performance.
Role
Solo MVP architect and builder across product, backend, frontend, realtime voice, grading, and payments.
Stakes
When students can outsource take-home essays, assessment has to move toward live reasoning, nuance, and defensible argument.
Built
Built in under 100 hours across teacher assignment flows, student debate sessions, OpenAI Realtime conversations, recordings, transcripts, structured grading, rubric review, exports, Supabase, FastAPI, Next.js, and Stripe.
Why it matters
Elenqa shows how quickly a market shift can become a tested product when product judgment and implementation stay together.
- Teacher and student workflows from assignment to recorded debate and grading.
- Realtime AI interlocutor designed to challenge the student's position.
- A fast, complete MVP with payments and production-shaped architecture.
AI Filmmaking
Commercials, proof films, and feature workflows built with a real production system.
For producers, directors, studios, agencies, and brands: development footage, commercials, shorts, remakes, animatics, and feature workflows directed with Slate. The cinema page has the broader reel set and film-specific offer.
Open the cinema pagePublic-domain remake proof / Feature study
Santa Claus Conquers the Martians
A Roger Avary collaboration taking a low-budget 1964 public-domain film and reimagining it as a Pixar-style children's adventure with modern production value.
Historical feature proof / Feature study
Wings
A Roger Avary collaboration updating a 1927 silent aviation film with action, speed, music, and spectacle calibrated for modern audiences.
Prestige remake proof / Feature study
The King of Kings
A Roger Avary collaboration using Cecil B. DeMille's silent biblical epic as a test for updating archival spectacle, reverence, and coverage for modern audiences.
Background
Business judgment, product execution, and creative direction in one operator.
Jason Thomas Ferrell has an MBA from the McCombs School of Business at The University of Texas at Austin, has served as an executive at two venture-backed businesses, and has spent more than a decade advising CEOs and operators on AI, machine learning, data science, and product strategy.
That consulting work spans legal technology, investment technology, fine-art authentication, cinema, parking infrastructure, healthcare, and medical workflows. The breadth matters: product judgment gets sharper when it has been tested across industries with different data, customers, regulations, economics, and operating constraints.
Executive operator
Leadership experience inside venture-backed companies, with strategy tied to shipped work.
AI and data science consultant
More than a decade of AI, machine learning, data science, and product strategy work for CEOs and operators.
Cross-industry judgment
Experience across legal, investment, fine-art, cinema, parking, healthcare, and medical contexts.
Board perspective
A decade of nonprofit board service and the governance discipline that comes with it.
Technical builder
Hands-on architecture and implementation across agents, document intelligence, legal AI, finance, education, and film workflows.
Creative technologist
AI filmmaking practice backed by Slate, a production system for controlling script-to-screen workflows.
Practice
Creative judgment, technical strategy, and implementation.
This is for teams that need more than a memo or a one-off demo: a film concept, investment thesis, product roadmap, MVP, or production AI system built quickly enough to shape a real decision.
AI filmmaking
Direct AI films, commercials, and remake studies with control over script interpretation, shot language, continuity, edit, tone, and production workflow.
PE AI diligence
Assess disruption risk, product expansion, defensibility, distribution leverage, and realistic value-creation paths for a leading private equity firm, targets, and portfolio companies.
Fractional CAIO
Turn board-level AI ambition into product strategy, model architecture, team direction, and systems that can be shipped.
Agentic products
Design tools, memory, state, orchestration, evaluation, recovery, cost controls, and long-running workflows.
Scientific judgment
Bring classical ML, data science, experimentation, and measurement discipline into modern LLM product development.
For film and media
For producers, directors, studios, agencies, and brands.
Proof-of-concept videos and commercials
under a week
Short films
one to two weeks
Feature films
six to eight weeks
For companies and investors
For CEOs, PE/VC firms, product teams, and portfolio companies.
Market-test MVPs
one to three days
Full MVPs and launchable products
under a month
Enterprise production systems
one to three months
Film, product, and strategy inquiries
Bring the project that needs both judgment and execution.
Producers and directors can move from script, remake, or campaign idea to table reads, animatics, proof-of-concept videos, and commercials in under a week, short films in one to two weeks, and feature film builds in six to eight weeks. CEOs, investors, and product teams can move from AI thesis to market-test MVP in days, launchable MVP or full product in under a month, and enterprise production system in one to three months depending on scope, data, security, and integrations.