Jay Han
Project DefaultSeptember 2026

AI Distribution
Infrastructure

We are building the distribution layer between
companies and consumers through AI interfaces.

Raising $750k pre-seed

02 / Why Now

AI assistants are becoming
the next consumer interface.

Every computing shift creates a new way for companies to reach consumers.

Web

Companies Websites Consumers

Search

Companies Search / Ads Consumers

Mobile

Companies App Store Consumers

AI

Companies AI Assistants Consumers

Users will discover, evaluate, and act on products through AI interfaces.

Project Default owns the distribution layer for this transition.

03 / ProblemThe proposal is distribution

AI is becoming the new vendor discovery layer

A user request can become a product or stack recommendation before the user visits a vendor.

ChatGPT / GPT-5.6 Sol High / Temporary Chat
User request, summarized:
“How should I build a shopping website?”
FrameworkNext.js + TypeScript
DatabasePostgreSQL / Supabase
PaymentsStripe Checkout
DeploymentVercel

One AI answer.
One distribution moment.

AI responses are becoming the place where companies are discovered.

We sell a disclosed place in that proposal.

04 / Default momentConcept conversation

The default proposal

One matched proposal first. Alternatives open through conversation.

01 / Default

“The recommended fit is [matched provider].”

Sponsored placement.

“Expected cost: [range], billed by [matched provider]. Proceed?”

02 / User: “Anything else?”

Other matched sponsors

Show the remaining sponsors in fit-score order. Exclude the proposed provider.

03 / User: “Other than those?”

Neutral alternatives

Show non-sponsored companies in fit order.

An explicit vendor choice always wins.

No matched sponsor: a neutral multi-option answer. No persistent alternatives button.

05 / Search vs. DefaultModel comparison

Keyword vs. situation

Conversation supplies context for matching a sponsor to a user’s task.

DimensionConventional search-ad flowProject Default
Starting signal“Web hosting”“Deploy my SaaS this week”
ContextA queryTask, constraints and qualification answers
PlacementSponsored search resultSponsored proposal inside the answer
Next stepUser visits a landing pageUser consents in the conversation
06 / ProductEnterprise sponsorship contracts

The AI distribution layer

Free or ultra-low-cost AI enables user access.

01 / Core

Slot fee

Fixed per period.
Paid upfront.
Scope follows the advertiser’s target.

02 / Optional

Performance fee

Additional fees on agreed, verified outcomes. Negotiated per contract.

03 / Optional

Exclusivity

Solo placement for the period. Sum of the slot fees plus a premium.

Enterprise sponsorship contracts power the distribution layer.

Intended scale: million-dollar enterprise contracts, with scope and pricing negotiated directly. Buyer access and amounts are validated after funding, together with investors. No click-based pricing.

07 / MatchingTarget-fit matching

Bought by sponsors.
Chosen by fit.

Money never enters the score.

Sponsor input

Target customer
in natural language

No forced targeting menus.

+
User input

Conversation
and qualification

Only observable attributes.

=
Sole ranking criterion

Target-fit
score

Highest-fit matched sponsor.

Payment buys eligibility for placement

Slot fees and conversion rates never determine the matching score.

Training follows the same rule

Target-fit labels and advertiser feedback. Never conversion labels.

08 / EconomicsThe first bar to clear

The serving-margin test

Sponsor revenue > AI serving cost

Recognized revenue / same period

Slot fees

+

Performance fees ×
recognized events

>
Serving cost / same period

AI inference + tool execution
+ infrastructure + support

Across all users,
including non-converters.

Then, company profit

Subtract development, sales and operating costs from serving margin.

Capacity follows collected funds

Sponsor cash plus investment sets seats and allowances. Investment is not revenue.

09 / PilotDemo first. Measurement plan follows.

Seven-day proof

Initial product: an AI development assistant with sponsored deployment choices.

PROJECT DEFAULT / CONCEPTFree AI models / limited usage

Build and deploy my project this week

01 Select a stack
02 Build the application
03 Deploy with consent
Thin client + our gateway
All model calls route through our server.
Enrollment
Public, first come. “What will you build and deploy this week?”
Scope
Stack selection, build and deploy.
Capacity
Models are free within a limited allowance. Users choose the model; expensive models use allowance faster. Seats and allowances follow secured funding.
Evidence
Ad delivery, early response and serving cost. Measures agreed after the demo.
01

Working demo

Prove the product works.

02

7-day pilot

Validate ad value and cost.

03

Large sponsorship sales

Sell using pilot evidence.

04

Full launch

Launch the service.

Seven days of operation, not development. Limited pilot slots. Measures follow the demo, with sponsors. Delivery and early response only.

10 / CompetitionMarket position

The competitive position

Incumbent AI assistants

Sponsored defaults require a new trust model and business constraints that incumbents face.

Search advertising

Starts with a query. The landing page and sales process follow the click.

Affiliate / comparison sites

Close to conversion, but users still need to visit them.

Our placement sits inside a conversation the user already has.
11 / RisksEach needs evidence

Four ways this fails

RiskTest or response
Sponsors will not payValidate buyers and pricing after funding, with investor introductions.
Serving costs exceed revenueMeasure all-user cost in the pilot. Set allowances from secured funds.
Users reject sponsored defaultsDisclose every proposal and respect explicit choice. Agree response measures with pilot sponsors.
Compliance or settlement failsCounsel review before the pilot. Match sponsor records to our logs for settlement.
12 / FounderSolo founder / Pre-seed

Jaeha (Jay) Han

Builder of AI systems
and developer tools.

Selected proof points
Threemmary ↗

Built a Gemini-based summarization application.

TooManyOres Modrinth / CurseForge

Built and distributed a Minecraft mod on two platforms.

Spectrum-State ↗

Built experimental AI systems and developer tools.

Building with investor participation

Jay builds product and AI systems. Investor introductions and advice support commercial validation after funding.

13 / AskProject Default
Pre-seed

$750k

Capital and strategic partners
to build the AI distribution layer.

Jaeha (Jay) Han
jay@nitg3n.xyz

01

Fund the test

Working demo, then a seven-day pilot. Sponsor payments plus investment fund operating capacity.

02

Validate the business together

After funding, use investor introductions and access to validate who pays, how much, and how we reach them.

03

Make operating decisions together

First hires and operating decisions follow investor advice after funding. Pilot evidence supports larger sponsorship sales.

Detailed allocation and runway remain open.