AI Beyond 2026

Exploring Business Opportunities
for AI Wrapper Software

Hyeongkeol Nam, Mingle LabsNov 5, 2026

AI Beyond 2026

About me

Hyeongkeol Nam
Mingle Labs

Before

Hyperconnect · Scatter Lab

Built AI and social products at both

2 years

All in on AI wrappers

Two years of shipping, failing and learning at Mingle Labs

Now

HashChat

Formerly Mingle. Nonstop, real-time interpreting across 3+ languages

AI Beyond 2026

Contents

1. AI Wrapper Case Studies

AI Beyond 2026

Case Studies

Definition

What is a wrapper?

“Software built by putting
a UI on top of a foundation
model like GPT.”

AI Beyond 2026

Case Studies

OVERVIEW

15 AI wrapper categories
with real revenue

01Character chatCharacter.AI · Zeta

02AI aggregatorsOpenRouter

03Note-takingGranola · tiro

04AI diet trackingCal AI

05Vibe codingCursor · Lovable

06AI study helpersStudyFetch

07AI browsersAside · Ego(lite)

08Language learningSpeak · heyring

09Photo editingSNOW · Fotor

10Professional servicesHarvey · Abridge

11Fortune tellingForceteller · Jeomsin

12YouTube summariesLilys AI · Eightify

13Shorts & highlightsOpusClip · AlphaCut

14Photo managementSwipewipe · Google Photos

15Workplace searchGlean · Notion AI

We’ll take a closer look at the four highlighted fields

AI Beyond 2026

Case Studies

CHARACTER

Field 01Character Chat

C.AI

Zeta

Crack

Ditto

Whif

SpicyChat.AI

JanitorAI

Talkie

Candy AI

Chai

Caveduck

LoveyDovey

DotDotDot

BabeChat

Lumo

AI Beyond 2026

Case Studies

CHARACTER

One of the hottest fields
Character.AI started it

Founded by ex-Google stars
Rejoined Google in a $2.7B deal

AI Beyond 2026

Case Studies

CHARACTER

How big is
character chat?

$30M+Zeta + Crack (OOC), combined 2026E worldwide revenue, growing 10x every year2026 estimate (provided figures)

4 hrsa day per Zeta user in JapanAI Times, Jul 2025

> YouTubeZeta’s daily time spent in KoreaApp usage data, 2025

10M+Zeta sign-ups worldwideScatter Lab, 2025

17%operating margin at Zeta (Q2 ’25)Scatter Lab, Jul 2025

$120M+consumer spend on AI companion apps in 2025Appfigures, Aug 2025

AI Beyond 2026

Case Studies

CHARACTER

At its core, character chat is
spicy stories & dopamine

AI Beyond 2026

Case Studies

CHARACTER

In Korea, Zeta beat
YouTube on time spent

Zeta ahead of YouTube in daily time spent in Korea

AI Beyond 2026

Case Studies

CHARACTER

Global apps only sell subscriptions: thin revenue
Korean apps squeeze the dopamine
and charge about 7¢ a turn

AI Beyond 2026

Case Studies

CHARACTER

History of
Character Chat

Phase 1Sep 2022 – Aug 2024

Key: model quality

From the C.AI launch to Google’s deal in Aug ’24

Phase 2Aug 2024 – Oct 2026

Key: content supply

A race for killer content, ended by C.AI’s bankruptcy

AI Beyond 2026

Case Studies

CHARACTER

What matters in Phase 3
The creator ecosystem

AI Beyond 2026

Case Studies

CHARACTER

Like webtoons,
it all comes down
to the business model!

Revenue structure of the webtoon industry

AI Beyond 2026

Case Studies

CHARACTER

Where is this field headed?
Merged with webtoons and shorts, it could grow 10x or more
Expect a crowded field, and keep an eye out for a hectocorn

AI Beyond 2026

Case Studies

CHARACTER

My take:
the hidden opportunity
Local-model businesses

1

Inference cost is the margin

Per-turn pricing brings revenue, but API bills eat the profit

2

General models limit content

The product is spicy storytelling, yet Big Tech APIs play it safe

3

Serve character-tuned models

Tune open models for role-play, then serve them yourself or on-device

AI Beyond 2026

Case Studies

AGGREGATOR

Field 02AI Aggregators

OpenRouter

Poe

Higgsfield

AI Beyond 2026

Case Studies

AGGREGATOR

One API and one bill
for countless AI models
and infrastructure

AI Beyond 2026

Case Studies

AGGREGATOR

How big is
aggregators?

$50Mannualized fee revenue, OpenRouterOpenRouter

23M+Combined accounts: OpenRouter 8M + Higgsfield 15MCompany figures, as shown

100Ttokens per month through OpenRouterOpenRouter

3M/daygenerations per day on HiggsfieldHiggsfield

$500Mannualized transaction volume, HiggsfieldHiggsfield

5%OpenRouter fee on every call, on top of provider pricesOpenRouter fee structure

AI Beyond 2026

Case Studies

AGGREGATOR

Aggregators take a 5% cut
much like Stripe or Visa

Customer Apps

OpenRouter 5% fee

Inference Provider 95% payout

AI Beyond 2026

Case Studies

AGGREGATOR

The core job: serving
countless open models
reliably

AI Beyond 2026

Case Studies

AGGREGATOR

In practice,
the two blur together

Pure Aggregators

OpenRouter

Poe

Higgsfield

vs

Inference Provider

FriendliAI

Replicate

WaveSpeed

Together

Baseten

Fireworks

AI Beyond 2026

Case Studies

AGGREGATOR

Aggregators route
traffic across
inference providers

Inference providers

FriendliAI

Replicate

WaveSpeed

Together

Baseten

Fireworks

Aggr.

OpenRouter API

AI Beyond 2026

Case Studies

AGGREGATOR

My take:
the hidden opportunity
Character chat & contents

1

Open source keeps prices stable

With per-turn pricing, cheaper inference is the edge

2

Reliable open-model serving

Hop across providers without downtime

3

Free choice of approved models

Pick models without being bound by general API policies

AI Beyond 2026

Case Studies

NOTES

Field 03Note-taking

Countless note-taking apps

AI Beyond 2026

Case Studies

NOTES

How big is
note-taking?

$2.9MTiro’s annualized revenue, Korea’s best-known AI note-taker₩4B annualized (≈$2.9M at ₩1,400/$)

40M+Otter.ai usersOtter.ai, 2025

$1.5BGranola valuation, 6x in 10 monthsSeries C, Mar 2026

$1B+Fireflies valuation, profitable since 2023Fireflies, Jun 2025

$250Mexpected 2025 revenue at Plaud, 1M+ devices soldPlaud co-founder, 2025

$3.5BAI meeting assistant market in 2025Grand View Research

AI Beyond 2026

Case Studies

NOTES

Everyone is making money
and growing fast

Growth chart with 12.7% average weekly growth

About 10% growth every week

AI Beyond 2026

Case Studies

NOTES

The endgame:
owning the
context layer

Char expanding into a single memory layer

AI Beyond 2026

Case Studies

NOTES

My take:
the hidden opportunity
Messaging

1

Notes are only part of it

Most work context lives in chats, not meetings

2

Own the context layer

Owning where conversations happen creates workflow lock-in

3

AI-native from day one

Grow into a messenger with summaries, to-dos and search built in

AI Beyond 2026

Case Studies

DIET

Field 04AI Diet Tracking

Foodvisor

Cal AI

SnapCalorie

NutriScan

CalZen

BiteSnap

MyFitnessPal

Lose It!

Cronometer

Nutrola

Calory App

Rex.fit

AI Beyond 2026

Case Studies

CORE FEATURE

Snap a photo,
get calories and macros

Food is recognized from the photo, and calories and macros are logged automatically.

Scanning a pancake photo and showing calories and macros

AI Beyond 2026

Case Studies

DIET

How big is
AI diet tracking?

$30M+Cal AI’s annual revenue, reached in under two yearsTechCrunch, Mar 2026 (reports go up to $40M)

15M+Cal AI downloads in under two yearsTechCrunch, Mar 2026

Mar ’26MyFitnessPal acquires Cal AI, a teen-built appTechCrunch, Mar 2026

~$6BDiet and nutrition app market in 2025 (estimates range $2.4B to $14B)Market Research Future, Roots Analysis

$1.2BCalorie-tracking apps segmentMarket Research Future

~2xMarket forecast to roughly double by 2030 ($5.8B to $10.2B)Mordor Intelligence

AI Beyond 2026

Case Studies

DIET

Another red-hot field
Cal AI started it

Cal AI founder — Zach Yadegari · founded it in high school

AI Beyond 2026

Case Studies

DIET

Fully paid from day one
and still hugely popular

Cal AI 3-day trial and subscription pricing — Charging upfront beats converting free users
Inference costs make it paid-only anyway

AI Beyond 2026

Case Studies

DIET

Where is this field headed?
Likely absorbed as one feature of the vast healthcare industry
Linking into other healthcare workflows will be key

AI Beyond 2026

Case Studies

DIET

My take:
the hidden opportunity
Medical consultations

1

Your data is the asset

Daily meal and weight logs are where consultations begin

2

From prevention to care

Like Pomellow (eating disorders), Noom and Counsel Health

3

In Korea, check the rules

Telemedicine rules apply, so start with non-medical coaching or abroad

AI Beyond 2026

Case Studies

VTUBER

One more:
AI VTubers

1

Becoming a VTuber is still too hard

2D rigging, 3D modeling, mocap studios: costly and slow

2

Generative models are collapsing the pipeline

Bringing models like Wan Animate and DreamActor to live streams

3

Go live from one photo

Virtually starts a stream from a single image in 10 seconds

AI Beyond 2026

Case Studies

VTUBER

The key:
simpler character creation

How

Live AI animation · vibe rigging

Animate one image with generative models, or let AI handle rigging

Proof

IRIAM proved the demand

Simple avatar streaming alone: 6.21M downloads, ¥8.5B+ in annual revenue

AI Beyond 2026

Case Studies

VTUBER

1/200 the cost,
live in 10 seconds

$7/hrAI live operation, vs up to $1,500/hr for mocapVirtually estimates

$0Setup and production, vs $1,000–$25,000 for 3DVirtually estimates

10 secFrom a single still image to a live streamVirtually

AI Beyond 2026

Case Studies

VTUBER

How big is
AI VTubers?

$2.2–3.0BGlobal VTuber market size, 2026E

$45M+Japan market size

$36M+China market size

$150M+Global production and operations

$22M+US production and operations

$7M+Korea production and operations

AI Beyond 2026

Case Studies

MY VIEW

Five ideas
worth pursuing

1

Character chat → local-model businesses

2

Aggregators → character chat & content

3

Note-taking → messaging

4

AI diet tracking → medical consultations

5

AI VTubers

AI Beyond 2026

Case Studies

STUDY

More fieldsAI Study Helpers

QANDA

Photomath

Gauth

Brainly

StudyFetch

AI Beyond 2026

Case Studies

BROWSER

More fieldsBrowsers

Perplexity Comet

Dia

Arc Search

Aside

Ego Lite

AI Beyond 2026

Case Studies

LANGUAGE

More fieldsLanguage Learning

Duolingo

Speak

ELSA Speak

heyring

AI Beyond 2026

Case Studies

PHOTO

More fieldsPhoto Editing

Remini

FaceApp

Photoroom

SNOW

Carat

Concept

AI Beyond 2026

Case Studies

EXPERT

More fieldsProfessional
Services

Harvey

SuperLawyer

Norm AI

Legora

LBOX

NearDoc

Abridge

Pilot

Counsel Health

Rogo

Hebbia

AlphaSense

Inkle

Maxed

Cranston

Black Ore

AI Beyond 2026

Case Studies

FORTUNE

More fieldsFortune Telling

Personalized LLM readings and pay-per-consult
are a natural fit

Forceteller

Jeomsin

Hellobot

Co–Star

AI Beyond 2026

Case Studies

SUMMARY

More fieldsYouTube
Summaries

Video summaries are becoming a Big Tech default
Survival means tying into notes and study

Lilys AI

Eightify

NoteGPT

AI Beyond 2026

Case Studies

CLIPS

More fieldsShorts &
Highlights

Auto-cutting long-form into short-form
growing with the creator economy

OpusClip

Vizard

Vrew

AlphaCut

AI Beyond 2026

Case Studies

PHOTOS

More fieldsPhoto
Management

AI apps to search, clean up and organize photos
Big Tech has the OS and cloud advantage

Swipewipe

Cleanup

Clever Cleaner

Slidebox

Google Photos

Apple Photos

AI Beyond 2026

Case Studies

KNOWLEDGE

More fieldsWorkplace
Search

Search scattered docs and chats, permissions intact
The front line of the context-layer race

Glean

Notion AI

Guru

DeepWiki

2. What Comes Next

AI Beyond 2026

What Comes Next

Limits

AI wrappers
create real value
but have no moat

“If your startup's core value proposition can be cloned in a weekend, you don’t have a business, you have a feature.”

“당신 스타트업의 핵심 가치가 주말 사이에 똑같이 복제될 수 있다면, 당신은 비즈니스를 가진 게 아니라 단지 하나의 '기능'을 가졌을 뿐이다.”

- Andrew Ng

AI Beyond 2026

What Comes Next

Limits

Elad Gil argues
moats can
come later

“Serving a customer need well is often more important (and harder) to think about than defensibility.”

“해자에 대한 생각보다는 고객 문제를 잘 푸는 게 중요하다.”

- Elad Gil

AI Beyond 2026

What Comes Next

MOAT

Wrapper first,
moat next

“The mistake isn't building an AI wrapper — the mistake is staying one.”

“AI 래퍼(Wrapper)로 시작하는 것이 실수가 아니라, 끝까지 AI 래퍼 상태로 남아있는 것이 실수다.”

- Paul Graham

AI Beyond 2026

What Comes Next

VISION

Where I think we should aim
"Ambient AI"

Ambient means “all around you.”

AI Beyond 2026

What Comes Next

VISION

Design Principles
of Ambient AI

Workflow Integration

Weave AI naturally into the customer’s workflow

Always-On Intelligence

Offer AI help automatically, as naturally as breathing

GUI to CUI

Replace complex GUIs with conversation

AI Beyond 2026

What Comes Next

WORKFLOW INTEGRATION

First,
workflow integration

1

Workflow Integration

2

Always-On Intelligence

3

GUI to CUI

AI Beyond 2026

What Comes Next

Workflow Integration

Let’s recall how software was built
until summer 2024, before Cursor took off

AI Beyond 2026

What Comes Next

Workflow Integration

AI was already
helping a lot with coding

ChatGPT

AI Beyond 2026

What Comes Next

Workflow Integration

“Developers are done”
was already a familiar line.

ChatGPT

AI Beyond 2026

What Comes Next

Workflow Integration

But to get that help, you had to
copy and paste code by hand.

VS Code

Copy & paste back and forth

ChatGPT

AI Beyond 2026

What Comes Next

Workflow Integration

Cursor’s team spotted this
and built AI into the editor

VS Code

ChatGPT

Cursor

AI Beyond 2026

What Comes Next

Workflow Integration

Then vibe coding took off,
just from the right integration.

Cursor

AI Beyond 2026

What Comes Next

Always-On Intelligence

Second,

AOI

1

Workflow Integration

2

Always-On Intelligence

3

GUI to CUI

AI Beyond 2026

What Comes Next

Always-On Intelligence

In April, X’s auto-translate
took the world by storm

AI Beyond 2026

What Comes Next

Always-On Intelligence

It just flipped the old
“Show translation” button on

AI Beyond 2026

What Comes Next

Always-On Intelligence

Yet people loved it
Why?

AI Beyond 2026

What Comes Next

Always-On Intelligence

Because it saved them
a single click

The bigger reason may be X’s recommender becoming i18n-aware.
That worked because X rebuilt its recommender to be AI-native a year earlier.

AI Beyond 2026

What Comes Next

Always-On Intelligence

Users found even
one click too much

AI Beyond 2026

What Comes Next

Always-On Intelligence

A feature that needs a click
was practically no feature at all

AI Beyond 2026

What Comes Next

Always-On Intelligence

Still, auto-applied AI
isn’t always a win

AI Beyond 2026

Case Studies

Always-On Intelligence

iPhone autocorrect fails

AI Beyond 2026

What Comes Next

Always-On Intelligence

Smarter AI will get it right on its own
Until then, designers have to design it right

AI Beyond 2026

What Comes Next

GUI to CUI

Third,

GUI to CUI

1

Workflow Integration

2

Always-On Intelligence

3

GUI to CUI

AI Beyond 2026

What Comes Next

GUI to CUI

One reason vibe coding feels easy:
no more fighting GUI complexity

AI Beyond 2026

What Comes Next

GUI to CUI

App Store releases,
AWS setup: same story

Even inviting internal testers in App Store Connect, it’s hard to find where the invite code comes from

AI Beyond 2026

What Comes Next

GUI to CUI

AI solves exactly this
Once things get complex, CUI wins by far

AI Beyond 2026

What Comes Next

GUI to CUI

Same with Google Ads:
editing an ad is a maze

AI Beyond 2026

What Comes Next

GUI to CUI

Same when a banking app frustrates you
you end up calling support
Why?

AI Beyond 2026

What Comes Next

GUI to CUI

Same when a banking app frustrates you
you end up calling support
to solve it in a conversational UI.

AI Beyond 2026

What Comes Next

GUI to CUI

GUI to CUI alone means
browsers and OSes
must go AI-native!

Comet

Dia

Arc Search

Aside

Ego Lite

iOS

macOS

Android

Windows

Linux

AI Beyond 2026

What Comes Next

Ambient AI

In short,
whoever nails Ambient AI
in their domain survives

1

Workflow Integration

2

Always-On Intelligence

3

GUI to CUI

AI Beyond 2026

What Comes Next

Ambient AI

Sometimes an AI wrapper
eats the existing
workflow

AI Beyond 2026

What Comes Next

Ambient AI

But incumbents may
reach Ambient AI
first

AI Beyond 2026

What Comes Next

Opportunity

So what should
AI wrapper startups do?

AI Beyond 2026

What Comes Next

Opportunity

1. Solve a real problem with a sharp wrapper

AI Beyond 2026

What Comes Next

Opportunity

1. Solve a real problem with a sharp wrapper
2. Grow into that domain’s Ambient AI

AI Beyond 2026

What Comes Next

Opportunity

My biggest bet:
AI-native messaging

AI Beyond 2026

What Comes Next

Opportunity

AI plus messaging and social
will transform most industries

AI+Messaging
Social

Sales & work

Shopping

Education

Travel

Content

Character chat

Companions

Dating

Medical, legal, finance and more

AI Beyond 2026

What Comes Next

Opportunity

Clayton Christensen
The Innovator’s Dilemma

The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail

"The reason is that good management itself was the root cause."

"실패의 근본 원인은 바로 훌륭한 경영 그 자체였다."

- Clayton M. Christensen

AI Beyond 2026

What Comes Next

Opportunity

The chance to overtake the giants
is right now

AI Beyond 2026

What Comes Next

Thank you

Thank you
Q & A

Hyeongkeol Nam, Mingle Labs

Exploring Business Opportunitiesfor AI Wrapper Software

Hyeongkeol NamMingle Labs

Hyperconnect · Scatter Lab

All in on AI wrappers

HashChat

1. AI Wrapper Case Studies

What is a wrapper?

15 AI wrapper categorieswith real revenue

Field 01Character Chat

One of the hottest fieldsCharacter.AI started it

How big ischaracter chat?

At its core, character chat isspicy stories & dopamine

In Korea, Zeta beatYouTube on time spent

Global apps only sell subscriptions: thin revenueKorean apps squeeze the dopamineand charge about 7¢ a turn

History ofCharacter Chat

Key: model quality

Key: content supply

What matters in Phase 3The creator ecosystem

Like webtoons,it all comes downto the business model!

Where is this field headed?Merged with webtoons and shorts, it could grow 10x or moreExpect a crowded field, and keep an eye out for a hectocorn

My take:the hidden opportunityLocal-model businesses

Inference cost is the margin

General models limit content

Serve character-tuned models

Field 02AI Aggregators

One API and one billfor countless AI modelsand infrastructure

How big isaggregators?

Aggregators take a 5% cutmuch like Stripe or Visa

The core job: servingcountless open modelsreliably

In practice,the two blur together

Aggregators routetraffic acrossinference providers

My take:the hidden opportunityCharacter chat & contents

Open source keeps prices stable

Reliable open-model serving

Free choice of approved models

Field 03Note-taking

How big isnote-taking?

Everyone is making moneyand growing fast

The endgame:owning thecontext layer

My take:the hidden opportunityMessaging

Notes are only part of it

Own the context layer

AI-native from day one

Field 04AI Diet Tracking

Snap a photo,get calories and macros

How big isAI diet tracking?

Another red-hot fieldCal AI started it

Fully paid from day oneand still hugely popular

Where is this field headed?Likely absorbed as one feature of the vast healthcare industryLinking into other healthcare workflows will be key

My take:the hidden opportunityMedical consultations

Your data is the asset

From prevention to care

In Korea, check the rules

One more:AI VTubers

Becoming a VTuber is still too hard

Generative models are collapsing the pipeline

Go live from one photo

The key:simpler character creation

Live AI animation · vibe rigging

IRIAM proved the demand

1/200 the cost,live in 10 seconds

How big isAI VTubers?

Five ideasworth pursuing

Character chat → local-model businesses

Aggregators → character chat & content

Note-taking → messaging

AI diet tracking → medical consultations

AI VTubers

More fieldsAI Study Helpers

More fieldsBrowsers

More fieldsLanguage Learning

More fieldsPhoto Editing

More fieldsProfessionalServices

More fieldsFortune Telling

More fieldsYouTubeSummaries

More fieldsShorts &Highlights

More fieldsPhotoManagement

More fieldsWorkplaceSearch

2. What Comes Next

AI wrapperscreate real valuebut have no moat

Exploring Business Opportunities
for AI Wrapper Software

Hyeongkeol Nam
Mingle Labs

15 AI wrapper categories
with real revenue

One of the hottest fields
Character.AI started it

How big is
character chat?

At its core, character chat is
spicy stories & dopamine

In Korea, Zeta beat
YouTube on time spent

Global apps only sell subscriptions: thin revenue
Korean apps squeeze the dopamine
and charge about 7¢ a turn

History of
Character Chat

What matters in Phase 3
The creator ecosystem

Like webtoons,
it all comes down
to the business model!

Where is this field headed?
Merged with webtoons and shorts, it could grow 10x or more
Expect a crowded field, and keep an eye out for a hectocorn

My take:
the hidden opportunity
Local-model businesses

One API and one bill
for countless AI models
and infrastructure

How big is
aggregators?

Aggregators take a 5% cut
much like Stripe or Visa

The core job: serving
countless open models
reliably

In practice,
the two blur together

Aggregators route
traffic across
inference providers

My take:
the hidden opportunity
Character chat & contents

How big is
note-taking?

Everyone is making money
and growing fast

The endgame:
owning the
context layer

My take:
the hidden opportunity
Messaging

Snap a photo,
get calories and macros

How big is
AI diet tracking?

Another red-hot field
Cal AI started it

Fully paid from day one
and still hugely popular

Where is this field headed?
Likely absorbed as one feature of the vast healthcare industry
Linking into other healthcare workflows will be key

My take:
the hidden opportunity
Medical consultations

One more:
AI VTubers

The key:
simpler character creation

1/200 the cost,
live in 10 seconds

How big is
AI VTubers?

Five ideas
worth pursuing

More fieldsProfessional
Services

More fieldsYouTube
Summaries

More fieldsShorts &
Highlights

More fieldsPhoto
Management

More fieldsWorkplace
Search

AI wrappers
create real value
but have no moat

Elad Gil argues
moats can
come later

Wrapper first,
moat next

Where I think we should aim
"Ambient AI"

Design Principles
of Ambient AI

First,
workflow integration

Let’s recall how software was built
until summer 2024, before Cursor took off

AI was already
helping a lot with coding

“Developers are done”
was already a familiar line.

But to get that help, you had to
copy and paste code by hand.

Cursor’s team spotted this
and built AI into the editor

Then vibe coding took off,
just from the right integration.

In April, X’s auto-translate
took the world by storm

It just flipped the old
“Show translation” button on

Yet people loved it
Why?

Because it saved them
a single click

Users found even
one click too much

A feature that needs a click
was practically no feature at all

Still, auto-applied AI
isn’t always a win

Smarter AI will get it right on its own
Until then, designers have to design it right

One reason vibe coding feels easy:
no more fighting GUI complexity

App Store releases,
AWS setup: same story

AI solves exactly this
Once things get complex, CUI wins by far

Same with Google Ads:
editing an ad is a maze

Same when a banking app frustrates you
you end up calling support
Why?

Same when a banking app frustrates you
you end up calling support
to solve it in a conversational UI.

GUI to CUI alone means
browsers and OSes
must go AI-native!

In short,
whoever nails Ambient AI
in their domain survives

Sometimes an AI wrapper
eats the existing
workflow

But incumbents may
reach Ambient AI
first

So what should
AI wrapper startups do?

1. Solve a real problem with a sharp wrapper
2. Grow into that domain’s Ambient AI

My biggest bet:
AI-native messaging

AI plus messaging and social
will transform most industries

Clayton Christensen
The Innovator’s Dilemma

The chance to overtake the giants
is right now

Thank you
Q & A