How quickly will large enterprises actually adopt AI?

Harry Stebbings · May 17, 2023

Enterprises currently have no idea what to do with AI

Scope: says he doesn't blame them for it

39:12 20VC: Why the AI Bubble Will Be Bigger Than The Dot Com Bubble, Why AI Will Have a Bigger Impact Than COVID, Why No Models Used Today Will Be Used in a Year, Why All Models are Biased and How AI Kills Traditional Media with Emad Mostaque, Founder & CEO @

Howie Liu · Aug 25, 2023 · hedged

Enterprises are nowhere near a tornado phase where every enterprise knows it wants AI and is ready to buy.

His conversations with enterprise leaders suggest they are not close to that yet.

Scope: based on his own conversations; allows that the 'train may arrive' later

0:00 20Product: Enterprises are not Adopting AI Yet, When Will AI Break Into Enterprise, What are the Blockers, What Do Enterprises Need from AI & Why Services Companies Will Win in the Next 10 Years of AI Implementation with Howie Liu, Founder & CEO @ Airtabl

Harry Stebbings · Aug 25, 2023

Large enterprises lag drastically behind on basic software adoption — many 50,000–500,000 person companies don't even use or know Slack.

He knows CEOs of such companies personally who have never heard of Slack.

Scope: European context

9:37 20Product: Enterprises are not Adopting AI Yet, When Will AI Break Into Enterprise, What are the Blockers, What Do Enterprises Need from AI & Why Services Companies Will Win in the Next 10 Years of AI Implementation with Howie Liu, Founder & CEO @ Airtabl

Howie Liu · Aug 25, 2023 · hedged

We are nowhere near the tornado phase where every enterprise knows it wants AI and is ready to throw its own resources at deploying it everywhere.

From his own sales conversations, enterprises are still trying to figure out what AI is and what they can use it for.

Scope: based on his own conversations; it may still happen — 'maybe the train will arrive'

23:01 20Product: Enterprises are not Adopting AI Yet, When Will AI Break Into Enterprise, What are the Blockers, What Do Enterprises Need from AI & Why Services Companies Will Win in the Next 10 Years of AI Implementation with Howie Liu, Founder & CEO @ Airtabl

Harry Stebbings · Sep 27, 2023

We are overestimating the speed at which large corporates will adopt AI

Most corporates in London and Europe are so far behind they don't even know what Slack is

Scope: Europe/London corporates as the reference case

54:57 20VC: "How Being a Founder Almost Killed Me"; We Have Lied to a Generation of Founders | The Hardest Truths About Being a Founder Revealed | Why AI Co-Pilot is BS, Seat Pricing is Over & User Interfaces are Stupid with Christian Lanng

Harry Stebbings · Dec 20, 2023 · hedged

AI technology progress is excellent while enterprise adoption is badly lagging, which is concerning

Many people rate AI development 10/10 and enterprise adoption dismally; a third of European corporates reportedly don't even know what Slack is

Scope: based on what many people he speaks to say

37:24 20VC Roundtable: Spotify, Adobe & Linkedin CPOs on How AI Changes The Future of Product, Why AI is Now the Product, How TikTok Changed Product, Why Cost is the Biggest Barrier to LLM Usage & Why Incumbents Can Adopt AI Faster Than Any Prior Innovation Cyc

Harry Stebbings · Oct 28, 2024

European large enterprises are so far behind on software adoption that an agent-based SaaS ecosystem is ten years out for them

62% of European large enterprises don't know what Slack is and 91% don't know Notion; getting them onto cloud at all is a challenge

Scope: specific to European large enterprises

12:38 20VC: Why SaaS is Dead | Why AI First Companies Will Win | We are in the Middle of a Cold War for AI Talent | Why Europe is F******* and We Need to Stop Whining with Daniel Khachab, Co-Founder @ Choco

Kim Graves · Jun 27, 2025

AI innovation has been impressively fast but enterprise AI adoption has been very slow, with the average enterprise still in wait-and-see mode

Enterprises have heavy existing process and many competing stakeholder priorities, so aligning them on what AI means for them takes a long time; their top-of-mind question is how to integrate AI into existing workflows

Scope: based on a survey Notion recently ran

54:02 20Sales: How to Layer Enterprise Sales on PLG | How to Sell AI Tools To Enterprises That Are Scared | Should Reps Own Their Own Pipeline | Mistakes All Founders Make When Moving From Founder-Led to Rep-Led Sales with Kim Graves

Jesse Zhang · Sep 19, 2025

The narrative that AI will transform every use case is the most overhyped thing in AI; most use cases, especially in enterprise, have not been transformed

Models are often not good enough yet, or the shape of the problem simply doesn't lend itself to AI

Scope: especially enterprise use cases

51:09 20VC: Why 90% of Founders Build Startups Wrong | Why AI Growth Rates are Sustainable & Remote Work is BS and the AI Talent War | Competing with Brett Taylor and Sierra: Who Wins the Customer Service War with Jesse Zhang, Decagon

Mike Cannon-Brookes · Oct 13, 2025

AI being understood and successfully deployed by companies and customers will take much longer to play through than expected

The delta between magical demos and actual value delivered is quite high, and closing it is just more work

Scope: applies to enterprise/customer deployment rather than model capability

54:59 20VC: Atlassian CEO on Why Everything is Overvalued & Are We in an AI Bubble | Do Margins Matter & Does Defensibility Exist in an AI World | Is Per Seat Pricing Dead & The Future of Vibe Coding with Mike Cannon-Brookes

Harry Stebbings · Dec 1, 2025

All knowledge work will not be automated within ten years — twenty is more plausible — because large companies' internal data and processes are far too primitive

From spending time with very large companies, the state of internal data and internal processes is laughable; they are far from adopting even Slack and Notion, let alone building custom models

Scope: concedes the pace of AI technological progress is underestimated; framed as respectful pushback inviting correction

18:13 20VC: Scale, Surge, Turing, Mercor: Who Wins & Who Loses in Data Labelling | Is Revenue in Data Labelling Real or GMV? | Why 99% of Knowledge Work Will Go and What Happens Then? | Why SaaS is Dead in a World of AI with Jonathan Siddharth @ Turing

Gustav Söderström · Dec 20, 2023

The bottleneck in AI adoption keeps moving up the stack — technical talent scarcity has eased and the hardest thing now is retooling the organization to rethink product and design around what the model needs to serve the user

Initially finding technical talent was hard, then making data useful, and the problem keeps shifting

19:47 20VC Roundtable: Spotify, Adobe & Linkedin CPOs on How AI Changes The Future of Product, Why AI is Now the Product, How TikTok Changed Product, Why Cost is the Biggest Barrier to LLM Usage & Why Incumbents Can Adopt AI Faster Than Any Prior Innovation Cyc

Aaron Levie · May 22, 2024 · hedged

The long pole of AI adoption is implementation and human change management, not technical breakthroughs

Scope: 'probably'

38:23 20VC: Box's Aaron Levie on Predictions for the Next Wave of AI: Will Foundation Models Be Commoditised | How the Business Model of SaaS Changes Forever | Startups vs Incumbents: Who Wins | App vs Infrastructure Layer: Where is the Value?

Harry Stebbings · Jun 19, 2024

We drastically overestimate how much company employees know about integrating AI into their business units

Accenture just posted $2.4B in generative AI revenues, and Accenture is a consulting business, not a software business — companies are paying for the knowledge they lack

Scope: not meant disparagingly

19:00 20VC: Foundation Models are the Fastest Depreciating Asset in History, Lina Kahn is a Threat to American Capitalism, PE is Not Coming to Save the M&A Market & How China Could Overtake the US in the AI Race with Michael Eisenberg

AJ Tennant · Oct 23, 2024

Enterprise AI is falling down at implementation — buyers are sold a dream and then face six to twelve month rollouts or a $5M services bill attached, and adoption and usage are not where they need to be

The bottleneck is not product but change management and security, which vendors and buyers both underestimate

23:37 20Sales: Biggest Lessons Scaling Slack from $6M to $1BN in ARR | How to Build a Customer Success Machine and Where Most Go Wrong | The Framework to Hire All Sales Reps: Take-Home Assignments, Hiring Panels and more with AJ Tennant @ Glean

Elias Torres · Mar 21, 2025

Humans are the biggest blocker to AI adoption, so enterprises will need extensive hand-holding and services help

The agents, workflows, new products and models needed for self-service adoption don't exist yet, so in the meantime people have to be walked through the step function by humans

Scope: framed as an interim condition until better agents/workflows/models exist

58:42 20VC: Selling Drift for $1.2BN is the Biggest Failure: What No One Tells You About Selling Your Company | Why Incumbents Are Slower & Worse Than Ever | Why the Most Valuable Companies in a World of AI Will Not Have More Than 100 People with Elias Torres

Jonathan Siddharth · Dec 1, 2025

Enterprise AI pilots fail not because models are weak but because enterprises skip the prerequisites: structuring their data, building the right agentic scaffold (prompting, context engineering, internal tool calls), good evals, and workflows designed for partial autonomy

From deploying AGI in enterprises, these are the steps most companies haven't done — the 'first mile schlep'

Scope: responding to the widely cited MIT finding that 95% of pilots fail

41:40 20VC: Scale, Surge, Turing, Mercor: Who Wins & Who Loses in Data Labelling | Is Revenue in Data Labelling Real or GMV? | Why 99% of Knowledge Work Will Go and What Happens Then? | Why SaaS is Dead in a World of AI with Jonathan Siddharth @ Turing

Harry Stebbings · Mar 21, 2026

Consulting businesses focused on AI implementation and adoption will do very well over the next few years, and adoption capability is one of the biggest things a company can improve at

Most companies, including sophisticated sales orgs, lack the internal expertise to implement and adopt new agentic tooling

43:26 20Sales: Inside Figma's $1BN ARR Revenue Machine | Why We Do Not Have Customer Success or SDRs | Why I Do Not Believe in Sales Quotas with Shaunt Voskanian, CRO @ Figma

Richard Socher · Apr 18, 2025

AI adoption will not happen as quickly as people think, despite being delivered as software

Changing processes takes people time, and around 60% of US adults have never even talked to a chat model — the figure is likely higher in Europe

Scope: concedes it is technically a software update

35:41 20VC: Foundation Models: Who Wins & Who Loses | How Economies and Labour Markets Need to Change in a World of AI | China vs the US in an AI Race: What You Need to Know | Rich Socher, Founder @ You.com

Anton Osika · Aug 18, 2025

The biggest bottleneck for large companies adopting AI is change management for the humans in the organization, and leaders should be studying how comparable companies changed rapidly rather than only picking tools

Change management is the binding constraint, so leadership should bring examples of successful rapid change into the leadership room and then across the organization before deciding which AI tools to use

40:10 20VC: Lovable CEO Anton Osika on $120M in ARR in 7 Months | The Honest Truth About Defensibility and Unit Economics for AI Startups | The State of Foundation Models: Long Grok, Short OpenAI, Why | Replit vs Lovable vs Bolt: What Happens

Jason James · Aug 29, 2025

The hardest part of building an autonomous AI recruiter is not the technology but change management — getting humans comfortable with the paradigm shift

They were fixated on whether the technology could do the job rather than what it would take for people to accept a change that dramatically affects their working lives

Scope: obvious in hindsight

51:32 20Product: Why Most CPOs are Bad | Why You Do Not Need PMs in a World of AI | Why the Design Stage is Dead and How to Use Vibe Coding to Replace It | The Three Roles All Founders End Up Firing on Repeat with Jason James @ Tezi

Joelle Pineau · Nov 3, 2025

The biggest enterprise AI adoption challenge is integrating AI with workflows and information systems accumulated over decades, alongside getting people through the change

Enterprises need to exploit information systems they have built over decades, and compatibility plus data confidentiality and security are prerequisites for that

Scope: Cohere's on-premise deployment focus shapes this view

20:43 20VC: Cohere's Chief AI Officer on Why Scaling Laws Will Continue | Whether You Can Buy Success in AI with Talent Acquisitions | The Future of Synthetic Data & What It Means for Models | Why AI Coding is Akin to Image Generation in 2015 with Joelle Pineau

Sebastian Siemiatkowski · Feb 16, 2026

The AI transformation will take longer than he previously expected, because the bottleneck is habit and workflow change rather than model capability

He tends to overestimate pace; it takes time for people to change habits and ways of working, so adoption lags capability

Scope: a revision of his own earlier faster-timeline view

70:14 20VC: SaaS is Dead: Why Systems of Record Will Die in an Agentic World | What Revenue Multiple Will Software Companies Trade At? | From 7,000 to 3,000: We Need Less People Than Ever with Sebastian Siemiatkowski

Harry Stebbings · May 17, 2023

AI will diffuse far faster than previous technology revolutions because the learning curve for an end user is nearly zero and integration takes a day

Industrialisation took thirty-plus years and PCs and the internet took ten-plus years, but with AI a marketer needs no training and integrations are immediate

Scope: says he is both excited and terrified

36:43 20VC: Why the AI Bubble Will Be Bigger Than The Dot Com Bubble, Why AI Will Have a Bigger Impact Than COVID, Why No Models Used Today Will Be Used in a Year, Why All Models are Biased and How AI Kills Traditional Media with Emad Mostaque, Founder & CEO @

Jeff Seibert · Nov 22, 2023

AI adoption will be radically faster than prior technology waves like mobile, and industries will be disrupted much more quickly

Mobile took five to seven years partly because it required buying new hardware and learning a new UX pattern; AI requires no new hardware purchase and chatbots are a fluid, familiar interface, so barriers are very low

32:20 20VC: Why OpenAI Will Become an Infrastructure Play, Why Apple Will Win in an AI World, Why Google is the Most Vulnerable Incumbent, Will LLMs Be Commoditised, Which Startups Are Thin vs Thick Wrappers on Top of LLMs with Jeff Seibert, Founder @ Digits

Harry Stebbings · Jan 22, 2025

Unlike past technological transitions such as mechanized agriculture or computing, which took at least ten years, AI adoption is effectively instant, which changes how enterprise value accumulates

You buy AI tools today and use them today, so a business can die instantly if it doesn't adopt

45:34 20VC: Why All AI Companies Are Under-Valued | The Future of Foundation Models: Scaling Laws, Generalised vs Specialised, Commoditised? | From Unable to Afford Rent to Raising $130M From Index and Peter Thiel with George Sivulka @ Hebbia

Harry Stebbings · Apr 18, 2025

The AI transition will be far faster than past technological transitions because it is a software update rather than physical machinery deployment

Agricultural mechanization and PC adoption took decades because physical equipment had to be moved into farms and offices; software has no such constraint

35:18 20VC: Foundation Models: Who Wins & Who Loses | How Economies and Labour Markets Need to Change in a World of AI | China vs the US in an AI Race: What You Need to Know | Rich Socher, Founder @ You.com

Harry Stebbings · Feb 23, 2026

Analogies to agricultural and industrial revolution adoption timelines are misleading for AI, because physical technologies required purchase, shipping, assembly and training whereas AI capability ships to users instantly.

A French farmer had to buy a tractor, wait a year, assemble it and train 75 people; Gemini ships Nano Banana Pro and you're using it tomorrow.

56:28 20VC: Inside Coatue's $70BN Machine: Why Price Matters Least | Why Mega Markets are the Most Important | How to Assess Durability of Revenue and Margins in AI with Lucas Swisher

Harry Stebbings · Jun 19, 2023 · hedged

Adoption of new technology tends to take longer than people expect

25:34 20VC: Why No Models Today Will Be Used in a Year, Why Open Will Always Beat Closed in AI, Why Proprietary Data is Less Important Than Ever And Why EU AI Regulation is a Disaster with Alex Lebrun, Founder & CEO @ Nabla

Harry Stebbings · Sep 22, 2023

Speed of enterprise technology adoption always takes longer than people expect

Many European enterprises still have no idea what Slack is

11:19 20VC: Are Foundation Models Becoming Commoditised? Do OpenAI and Anthropic of the World Have a Sustaining Moat? Why Smaller Models May Work Better? Why Incumbents with Data Power Win the AI War with Christian Kleinerman, SVP Product @ Snowflake

Brad Lightcap · Apr 15, 2024 · hedged

Enterprise adoption of AI will be far faster than people realize, bucking the convention that enterprises are slow adopters of technology.

Scope: stated as a recent change of mind; not varying by geography

42:29 20VC: OpenAI's Sam Altman and Brad Lightcap on The Future of Foundation Models: Will They Be Commoditised | How to Solve the Problem of Compute | Open vs Closed: Which Dominates and Why | Which Companies and Verticals Will Be Steamrolled by OpenAI

Harry Stebbings · Jan 22, 2025

We consistently underestimate how long enterprises take to adopt new technology and get comfortable with data security and process change

Scope: relayed from a prior show appearance, offered as a challenge rather than Harry's own view

29:48 20VC: Why All AI Companies Are Under-Valued | The Future of Foundation Models: Scaling Laws, Generalised vs Specialised, Commoditised? | From Unable to Afford Rent to Raising $130M From Index and Peter Thiel with George Sivulka @ Hebbia

Emad Mostaque · May 17, 2023

When AI adoption starts in earnest inside enterprises it will be overwhelming, because enterprise work is mostly services and information flow that can be replicated a thousandfold at the push of a button

So much of enterprise is services and information flow, and spinning up a thousand agents changes the economics entirely

37:06 20VC: Why the AI Bubble Will Be Bigger Than The Dot Com Bubble, Why AI Will Have a Bigger Impact Than COVID, Why No Models Used Today Will Be Used in a Year, Why All Models are Biased and How AI Kills Traditional Media with Emad Mostaque, Founder & CEO @

Douwe Kiela · Jun 30, 2023

The blockers to enterprise AI adoption are hallucination, attribution, compliance, up-to-dateness, data privacy and latency — and the field is solving them, so the tidal wave of adoption is coming.

The whole field is moving toward making models ready for enterprise usage.

28:40 20VC: Why Data Size Matters More Than Model Size, Why The Google Employee Was Wrong; OpenAI and Google Have the Advantage & Why Open Source is Not Going to Win with Douwe Kiela, Co-Founder @ Contextual AI

Roman Chernin · Jun 8, 2026

Non-AI-native companies look slow early because of a cold-start problem, but once they solve the foundational shipping system their AI adoption grows exponentially — so expect explosive AI growth from cloud-native enterprises

Solving the foundations lets a strong team ship fast, evolve models and make decisions, and then build many more AI products internally

Scope: requires a strong team

37:39 20VC: Nebius Co-Founder on AI Infrastructure Bubbles | The Real Impact of Open Source on OpenAI & Anthropic | How Price Elastic is Demand for Compute | Could Nebius Sell 10x More Compute If They Had It & more with Roman Chernin

Elias Torres · Mar 21, 2025

The AI transition period will be on the longer side (years, not twelve to twenty-four months), and there is a lot of hype

Even the industrial revolutions produced only 1-2% GDP growth, so a sudden 10-20% jump is implausible; most industries adopt even slower than tech, and tech itself has been extremely slow to adopt this technology

59:09 20VC: Selling Drift for $1.2BN is the Biggest Failure: What No One Tells You About Selling Your Company | Why Incumbents Are Slower & Worse Than Ever | Why the Most Valuable Companies in a World of AI Will Not Have More Than 100 People with Elias Torres

Zach Lloyd · Oct 17, 2025

AI's transformation is a ten-year story rather than a three-year one because the technology will run far ahead of its deployment

Deployment takes a long time and is fast only where barriers are low; hard-to-change institutions like the DMV will still be running old systems

33:51 20VC: The Startup Adding $1M ARR Every Week | Competing Against OpenAI's Codex and Claude Code: Who Wins | Why Gemini is Failing and GPT-5 Is Winning | Do Margins Matter in a World of AI | The Ugly Truth About AI Coding with Zach Lloyd, Warp

Andrew Ng · Nov 17, 2025

AI adoption will deliver tremendous GDP growth but take far longer than the hype suggests — a decade from now we will still be identifying and building valuable enterprise applications.

By analogy to the cloud era, where plenty of on-prem workloads still remain years in.

Scope: a lot of progress still expected in the next one to two years

42:26 20VC: Andrew NG on The Biggest Bottlenecks in AI | How LLMs Can Be Used as a Geopolitical Weapon | Do Margins Matter in a World of AI? | Is Defensibility Dead in a World of AI? | Will AI Deliver Masa Son's Predictions of 5% GDP Growth?

Victor Riparbelli · Jan 15, 2025

Enterprise AI buyers do not know what they actually want — they are mandated to have an AI strategy and willing to spend innovation budget, which is both an opportunity and a trap for startups

Buyers don't understand the technologies well enough to determine their own needs, so they'll sign pilots and POCs to satisfy their boss, which makes ROI impossible to prove

Scope: a bigger problem for startups lacking customer obsession

18:32 20VC: Why Scaling Laws Will Not Continue | OpenAI vs Anthropic vs X.ai: Who Wins and Why | How Far Will Model Providers Go Into the Application Layer | The End State for Models: Many Specialised or Few Generalised with Victor Riparbelli @ Synthesia

Bucky Moore · May 5, 2025

There is unprecedented appetite among enterprise CIOs and CTOs to adopt AI solutions because their CEOs and boards have told them the company's fate depends on it and they'll be fired otherwise

Senior-most decision makers are furiously seeking ways to apply AI under existential pressure from above

30:31 20VC Exclusive: Why Mega Platforms Will Win in VC | Why You Cannot Do VC If You Do Not Do Pre-Seed | Why Market Sizing is BS | Where Will Foundation Models Build/Buy Apps vs Where Will They Not with Bucky Moore

Eran Zinman · Mar 2, 2026

Companies today have no idea how to adopt AI, but they will buy AI capability regardless because they won't survive competitively without it

The tech industry lives in an echo chamber; the wider market is far less sophisticated but faces competitive pressure

31:49 20VC: Monday.com CEO on Is SaaS Dead: Will Everything Be Vibe Coded | Will Systems of Record Become Valueless Databases in an Agentic World | Will LLMs Own the Value in the Application Layer with Eran Zinman

Harry Stebbings · Feb 16, 2026

We systematically underestimate how fast consumers adopt AI and overestimate how fast enterprises do

The usage numbers for consumer AI products like ChatGPT surprised him with how quickly consumers moved

70:38 20VC: SaaS is Dead: Why Systems of Record Will Die in an Agentic World | What Revenue Multiple Will Software Companies Trade At? | From 7,000 to 3,000: We Need Less People Than Ever with Sebastian Siemiatkowski

Lucas Swisher · Feb 23, 2026

AI is being adopted materially faster than SaaS or cloud was, but full enterprise labor displacement will still take a long time — SaaS companies won't evaporate overnight.

Anthropic is at ~$9BN ARR growing 800% where the three hyperscalers averaged 60% growth at the same scale; but enterprises are slow and sticky and integration and deployment work is complex.

Scope: people extrapolate consumer adoption speed to enterprise, which is a mistake

55:32 20VC: Inside Coatue's $70BN Machine: Why Price Matters Least | Why Mega Markets are the Most Important | How to Assess Durability of Revenue and Margins in AI with Lucas Swisher

Daniel Khachab · Oct 28, 2024

AI is a better fit for traditional, non-tech industries than for startups and tech companies, and adoption there will be faster

The real barrier to adoption in traditional industries was never valuing digital but having to learn new interfaces; with AI you just express what you want in natural language, like using WhatsApp, so there is nothing new to learn

12:59 20VC: Why SaaS is Dead | Why AI First Companies Will Win | We are in the Middle of a Cold War for AI Talent | Why Europe is F******* and We Need to Stop Whining with Daniel Khachab, Co-Founder @ Choco

Matan Grinberg · Jun 13, 2026

Accounting firms are the legacy incumbents that have most successfully embraced AI, and are now more agent-native than some startups

They carry scars from being late to cloud, and had engineering leaders who committed to making the org agent-native early despite internal upset

Scope: based on his own customer base

76:41 20VC: Who Wins the Model War: OpenAI, Anthropic or Open-Source | Token Maxing, AI Hangovers & The Coming ROI Reckoning | Labour Displacement Fears are BS & Overblown | From Physicist to Sequoia Founder with Matan Grinberg, Founder @ Factory

Emad Mostaque · May 17, 2023

Enterprise adoption will take much longer because enterprises need auditable, standardized models and, in regulated sectors like financial services, cannot have any web-crawl data in the model at all

Investment banks and asset managers say they cannot use a black box, regulators are asking what data is in there, and they fear out-of-sample behaviour like the model repeating something rude it saw on Reddit

Scope: financial services and regulated enterprises

25:43 20VC: Why the AI Bubble Will Be Bigger Than The Dot Com Bubble, Why AI Will Have a Bigger Impact Than COVID, Why No Models Used Today Will Be Used in a Year, Why All Models are Biased and How AI Kills Traditional Media with Emad Mostaque, Founder & CEO @

Zach Lloyd · Oct 17, 2025

AI's near-term impact depends on industry structure rather than the technology: in competitive, unregulated sectors like SaaS and non-regulated knowledge work it changes everything, while healthcare and government will lag

The binding constraint is whether incentives exist to deploy the technology, not whether the technology is transformational — regulated sectors like healthcare still run on paper forms

Scope: varies by industry

32:25 20VC: The Startup Adding $1M ARR Every Week | Competing Against OpenAI's Codex and Claude Code: Who Wins | Why Gemini is Failing and GPT-5 Is Winning | Do Margins Matter in a World of AI | The Ugly Truth About AI Coding with Zach Lloyd, Warp

Harry Stebbings · Nov 15, 2023

Disruption of tools like the ad optimization engine will take five to ten years, because their customers are large slow-moving enterprises.

Many of those customers are companies like Danone and L'Oreal who are still asking what ChatGPT is; adoption cycles are far slower than the excitement implies.

27:47 20VC: How to Survive and Thrive in a World of OpenAI, Are LLMs Being Commoditised, Where Does the Value Lie; Infrastructure or Application Layer, How Apple Could Win in a World of AI, How Amazon Could Threaten OpenAI and Why Google Struggle with Des Trayn

Daniel Dines · Dec 18, 2024

With current state-of-the-art LLMs, wide-scale deployment of agentic automation will take another five to ten years

Corporate inertia is underestimated; you can't just show an agent a job once and have it reliably run — it takes a full deployment program, same as RPA

Scope: assumes LLMs do not reach AGI in the meantime; with the current state of the art

32:07 20VC: UiPath's Daniel Dines on Why Agents Do Not Mean RPA is F***** | Why We Have Reached the Upper End of Scaling Laws | The Future of Work in an Agent World and What Everyone Misunderstands About Enterprise AI

AJ Tennant · Oct 23, 2024

Enterprises do have real urgency to buy AI now, driven by herd mentality as boards and CEOs see a handful of big brands already getting top- or bottom-line impact

A few companies doing AI right are already impacting their P&L and spending $5-10M annually, which pressures other boards and CEOs to move quickly

22:14 20Sales: Biggest Lessons Scaling Slack from $6M to $1BN in ARR | How to Build a Customer Success Machine and Where Most Go Wrong | The Framework to Hire All Sales Reps: Take-Home Assignments, Hiring Panels and more with AJ Tennant @ Glean

Bucky Moore · May 5, 2025

The conventional wisdom that enterprises adopt new technology slowly does not hold for AI; this cycle is fundamentally different and faster

There is broad-based consensus that failing to embrace AI is existential for a company, which creates voracious, at-all-costs adoption appetite in every corner of the business; this explains unprecedented company growth rates and investors' bullishness on app-layer companies, with CIOs saying they'll get fired if they don't adopt everywhere

Scope: contrasted with cloud, which was gradual then sudden

56:20 20VC Exclusive: Why Mega Platforms Will Win in VC | Why You Cannot Do VC If You Do Not Do Pre-Seed | Why Market Sizing is BS | Where Will Foundation Models Build/Buy Apps vs Where Will They Not with Bucky Moore

Ethan Mollick · Jul 31, 2024

Almost nobody in organizations actually uses frontier AI systems seriously, despite nearly everyone having tried ChatGPT

In any room — banks, innovation conferences, even Silicon Valley people outside labs — only about 5-10% have used the frontier models and only 2-3% have used them for ten hours, his minimum threshold

Scope: based on his own informal show-of-hands sampling across audiences

26:10 20VC: Is More Compute the Answer to Model Performance | Why OpenAI Abandons Products, The Biggest Opportunities They Have Not Taken & Analysing Their Race for AGI | What Companies, AI Labs and Startups Get Wrong About AI with Ethan Mollick

Roman Chernin · Jun 8, 2026

Almost every large company, even technologically advanced ones, is only in the first few percent of AI adoption by volume and by use case

Looking at actual enterprise deployments shows they are just starting, which implies enormous headroom independent of any grand futurist claims

Scope: excludes the fastest-moving startups

4:58 20VC: Nebius Co-Founder on AI Infrastructure Bubbles | The Real Impact of Open Source on OpenAI & Anthropic | How Price Elastic is Demand for Compute | Could Nebius Sell 10x More Compute If They Had It & more with Roman Chernin

Brad Lightcap · Apr 15, 2024

Enterprises genuinely want to move at AI's rate of change, but at 100,000–200,000 person scale it is very hard, and helping them do it is OpenAI's central challenge for the next few years

Desire exists but organizational scale makes execution extremely difficult

Scope: about very large enterprises

34:30 20VC: OpenAI's Sam Altman and Brad Lightcap on The Future of Foundation Models: Will They Be Commoditised | How to Solve the Problem of Compute | Open vs Closed: Which Dominates and Why | Which Companies and Verticals Will Be Steamrolled by OpenAI

Matan Grinberg · Jun 13, 2026

AI tools will produce tremendous productivity growth, but the gains take time to permeate because resource allocation inside companies adjusts slowly

Problems can already be solved faster on a problem-by-problem basis, but companies organized around headcount-to-problem ratios take time to reallocate people and dollars

Scope: gains lag adoption

5:23 20VC: Who Wins the Model War: OpenAI, Anthropic or Open-Source | Token Maxing, AI Hangovers & The Coming ROI Reckoning | Labour Displacement Fears are BS & Overblown | From Physicist to Sequoia Founder with Matan Grinberg, Founder @ Factory

Aaron Levie · Apr 20, 2026

Startups and most podcast guests misunderstand enterprise AI adoption because greenfield companies face no regulators, no legacy workflows and no fragmented data, whereas Fortune 1000 firms face all three.

When you start from scratch you can design workflows however you want with no downside risk or scale; a Fortune 1000 pharma company or bank is regulated, has data fragmented across the org, and employees wired to work a particular way.

11:59 Structural constraints invisible to greenfield builders

Andrew Feldman · Oct 6, 2025

Electricity produced almost no manufacturing productivity gains until factories reorganized the shop floor around it, and computers likewise produced little gain while merely replacing typewriters and ledgers, with the jump arriving in the mid-nineties once machines were networked.

Paul David's study of electricity adoption in manufacturing from 1880 to 1955, following Robert Solow's productivity paradox.

24:10 Productivity gains lag until work reorganizes around the technology electricity computer precedent

Jeff Seibert · Nov 22, 2023

Enterprise resistance to sending sensitive data to AI providers will be overcome the same way cloud resistance was, via clear data-use guidelines and earned trust

Ten years ago enterprises said they'd never put data on AWS or Azure; now they all do, and it's better because running data centers is those companies' core competency while most enterprises have no idea how to operate one

30:01 Enterprise data resistance to ai will be overcome like cloud resistance was via trust and guidelines

Arthur Mensch · Apr 29, 2024

The main enterprise obstacle is prioritization rather than capability, and adoption will accelerate as off-the-shelf solutions and developer platforms remove the need to hire scarce in-house AI scientists

The technology can take many forms, so choosing what to bring to market is a strategic challenge that is not easy for enterprises generally

29:59 Prioritization not capability is the main enterprise obstacle and off the shelf platforms will accelerate adoption

Arthur Mensch · Apr 29, 2024

Generative AI has already moved into core budgets where the application is obvious such as customer support, while remaining experimental in other functions and still at playground stage for core telecom and healthcare applications

31:05 Adoption stage varies widely by function and industry from core budget to experimental to playground

Hemant Taneja · Sep 22, 2025

The MIT finding that most enterprise AI efforts fail has merit, because real transformation requires four hard things at once: data/infrastructure readiness, models trained on the company's own business, workforce and org-chart redesign, and courage from the CEO

It is very difficult for all four to be in place simultaneously, so most companies stall at prototyping an OpenAI or Anthropic model rather than changing the business

16:11 Four simultaneous prerequisites data models org redesign ceo courage explain failure

Harry Stebbings · May 12, 2023

Large enterprises have generally been slow and bluntly ignorant toward new waves of innovation, and if offered a bundled, verified, 'safe' service they will go for it

Observed enterprise buying behaviour and education levels in Europe; a bundled service with a verification stamp is easy to say yes to.

15:57 Enterprises prefer bundled verified safe ai services over tailored solutions

Mamoon Hamid · Oct 21, 2024

The current wave of enterprise AI proof-of-concepts mirrors the late-1990s internet build-out, where companies first spent internally and then turned to outside providers

Every CIO is spending real money on POCs and finding it hard to build custom things inside; twenty-five years ago the same dynamic sent companies to consultancies like Razorfish and Sapient

11:56 Poc then outside provider pattern mirrors late 1990s internet buildout

Douwe Kiela · Jun 30, 2023

Enterprise AI adoption is already happening today rather than starting next year.

At a Google exec event this week, C-level leaders from companies worldwide were all experimenting with AI and putting it into production in places.

29:15 Enterprise ai adoption is already underway not a future event

Douwe Kiela · Jun 30, 2023 · hedged

Enterprise AI adoption will be gradual rather than sudden, because finding the right use cases is the current bottleneck.

People now assume GPT-4 can be used for anything, which isn't true, so work is going into matching use cases to model types.

29:49 Use case matching not raw capability is the adoption bottleneck

Christian Lanng · Sep 27, 2023

Enterprise AI adoption can move very fast when trust guardrails and human validation loops are strong, because buyers see no downside

He saw a pilot go from trial to a 3,000-seat deployment request in sixty days precisely because humans validated everything

55:18 Adoption is fast when guardrails remove downside

Anastasios Angelopoulos · Aug 3, 2026

Every business in the world will need evaluation, and evaluation is the single biggest bottleneck to deploying AI

Businesses struggle to define value — cutting cost is easy, but defining what performance is worth is not

46:43 Evaluation difficulty is the binding bottleneck on deployment

Max Junestrand · Aug 15, 2025 · hedged

Law firms' sophistication about AI systems is increasing faster than vendors can build, moving them from box-ticking to fundamentally rethinking processes and pricing

Firms now believe software will be an integral part of their entire service offering, so they evaluate vendors critically

64:44 Buyer sophistication about ai is outpacing vendor build speed

Jonathan Siddharth · Dec 1, 2025 · hedged

AI adoption will be slow in back office automation but fast in the front office, especially financial services, life sciences and pharma

It is far easier to convince people to adopt technology that makes money than technology that saves money; in efficient markets like financial services you get killed if you're not at the bleeding edge, and in pharma accelerating time to a drug or molecule helps them win at the thing they care about; back office change management will be too slow

19:59 Front office revenue generating adoption is fast back office slow

Aidan Gomez · Aug 19, 2024

Enterprise AI has shifted out of experimental/proof-of-concept budgets into an urgent push to get the technology into production

After a year of running POCs, enterprises are afraid of being caught flat-footed and are now sprinting to put AI into products and augment their workforce

38:41 Enterprises shifted from poc experimentation to urgent production push fearing being left behind

Christian Kleinerman · Sep 22, 2023

The generative AI stack is still unsettled — there is no clean canonical set of three or four components mapped to use cases — so the constraint is stack immaturity as much as enterprise education

Companies vary widely on which model, which vector database, and how much to prompt versus fine-tune; the stack will keep evolving and maturing before best practice per use case is clear

10:30 Genai stack immaturity not just enterprise education is the adoption bottleneck

Christian Kleinerman · Sep 22, 2023 · hedged

Financial services will be at the forefront of AI adoption, followed by retail and CPG, while the public sector will be slowest

Financials and retail/CPG have long figured out how to organize and exploit data for advantage (e.g. hedge fund sophistication); the public sector is data-savvy but constrained by regulation

11:41 Financial services leads adoption followed by retail cpg then slower public sector

Christian Kleinerman · Sep 22, 2023

Companies should push AI adoption as fast as they can because productization difficulties will naturally throttle the pace anyway

Demos are awesome but productization takes far longer than people realize

33:04 Push adoption fast since productization difficulty naturally throttles pace

Daniel Dines · Dec 18, 2024

The inertia of corporations is underestimated — even RPA, a pretty good technology, is nowhere near fully penetrated

RPA technology is already pretty good and still nowhere near fully penetrated

31:50 Corporate inertia is underestimated even well proven rpa remains far from full penetration

Howie Liu · Aug 25, 2023

Enterprise AI is still in the education phase — customers are limited by their own understanding of what Gen AI can do rather than by demand.

Enterprises are only now learning basic primitives like what an LLM or vector database is; once that baseline understanding exists, they will start applying the technology to specific problems.

8:37 Customers limited by own understanding of genai not by demand education phase

Howie Liu · Aug 25, 2023

Unlike traditional enterprise tech, Gen AI will not be confined to the enterprise — mainstream awareness and scale have come through consumer applications like ChatGPT and MidJourney, not just Silicon Valley early adopters.

Awareness came through consumer applications like ChatGPT and MidJourney that already reached real scale beyond Silicon Valley elites.

10:02 Genai awareness spread through consumer apps unlike traditional enterprise tech adoption path

Shyam Sankar · Jan 17, 2024

Technologists should be focused on proof of value rather than proof of concept, because impressive demos can be decades away from production adoption

A self-driving car drove 132 miles through the desert in 2005, yet only in 2023 was there a reliable commercial city self-driving offering; no enterprise wants to adopt something that might reach production in twenty years

55:27 Impressive demos can be decades from production so prioritize proof of value over proof of concept

Eran Zinman · Mar 2, 2026

Technology is moving fast but organizational AI adoption will take much longer, because context — not intelligence — is the binding constraint

No matter how smart a model is, it cannot do a job without context, and roughly 90% of any company's context is undocumented and floats in the air

35:00 Undocumented organizational context is the binding constraint

Des Traynor · Nov 15, 2023

Investing in an AI startup today means betting on adoption at least four to seven years out, and revenue streams may not have crossed the incumbent's even by then.

The right move when you see a real disruption is to sell the incumbent's stock over two years and plant seeds in the upstart with a payday seven years away; you're betting on future market appetite, not next year's big bang.

28:16 Investing in ai disruption means betting on a four to seven year adoption horizon

Elias Torres · Mar 21, 2025

Elephants can't dance because of headcount — people are the blockage to adopting AI and technology

In big companies nobody gets fired for saying no, so it's easier to coast; time goes to promotions, approvals and weekly status emails, and public-company CEOs spend their time on reporting, analysts and press instead of customers, product and vision

18:05 Internal bureaucracy promotions and reporting overhead not technology block ai adoption

Elias Torres · Mar 21, 2025

AI adoption in enterprises will be slow because middle layers block it, and CEOs are making a mistake by delegating the decision instead of owning and driving it

In his own selling, CEOs say yes enthusiastically then pass him down the chain where he gets resistance all the way down, and the CEO later says the team is too busy and to come back next quarter; if a 2,000-person company behaves like this, larger ones will be worse

21:54 Ceo delegation of the ai adoption decision rather than personal ownership causes slow adoption

George Sivulka · Jan 22, 2025

Finance is the worst, most lethargic customer base to sell to unless you provide outsized alpha or real value — in which case it moves faster than any other industry

Excel reached ~90% penetration in finance in eighteen to twenty-four months (1985-86) as everyone dropped the HP 12c, and finance adopted credit card data for valuing public companies within a roughly two-year window

30:11 Finance adopts slowly except when shown real alpha then fastest of all

Emad Mostaque · May 17, 2023 · hedged

There is roughly a six-month window while everyone gets to grips with AI, after which design patterns standardize and spread and laggards are forced to implement to catch up with those outpacing them

Nothing has been standardized yet and everyone is figuring it out simultaneously; once some players visibly outpace others, competitive pressure forces adoption

38:16 Six month standardization window after which laggards forced to catch up

Brad Lightcap · Apr 15, 2024

The question big companies fail to ask is how steep the rate of model improvement is; they wrongly assume the technology is static and that GPT-4 is as good as models will get.

Every technology they've had to adopt before was relatively static — mobile in 2009 versus today is basically the same technology and application development pattern, and the same is true of cloud

32:46 Enterprises wrongly assume ai is static like past platform shifts

Harry Stebbings · Apr 15, 2024

Large corporates, especially European ones, are structurally too slow to absorb OpenAI's rate of model change, because each update invalidates the workflows and processes they just built

Corporates get used to their workflows and processes, and then a model update throws them out the window

34:10 European corporates too slow for ais rate of model change

Harry Stebbings · Jun 27, 2025

There will come a point where public markets penalize companies for not embracing AI, and that market-cap pressure is what will force enterprise adoption

Once market cap is impacted, companies change behavior

55:16 Public market penalty for non adoption will eventually force enterprise ai uptake

Aatish Nayak · Apr 11, 2025

Wide-scale AGI adoption will be bottlenecked by cultural, legal and regulatory barriers rather than arriving smoothly once the technology exists

No governance framework exists for an AGI running a company, and law firms and customers he's asked all say liability and indemnification questions are unresolved for autonomous AI action; culturally there are domains where humans won't want AI involved

49:26 Cultural legal and regulatory barriers not technology are the agi adoption bottleneck

Andrew Feldman · May 26, 2026

Industry adoption tips when leaders personally weigh productivity gains against unseen risk and decree adoption over their own internal legal teams

The blocker is an unseen boogeyman of risk that is occasionally real but usually not, so it takes a leader's judgment call to override

37:13 Leader override of internal legal is what unlocks adoption

Richard Socher · Apr 18, 2025

Large enterprise AI deployments most often fail on adoption, because AI turns every employee into a manager and most individual contributors lack the skill to specify their knowledge unambiguously to an agent

Customers bought thousand-seat OpenAI licenses and found only ~6% weekly usage six months later; people used to doing specific work well aren't used to managing another entity

15:03 Ai turns every employee into a manager and most lack that skill

Gustav Söderström · Dec 20, 2023

AI adoption will move much faster than the move to cloud, producing a sharper S-curve

On-prem companies couldn't partially try the cloud — it was incredibly hard — whereas AI tools can be tried in parallel with existing systems

39:11 Trialability in parallel with existing systems makes ai adoption faster than the all or nothing cloud switch

Alex Lebrun · Jun 19, 2023

Doctors have become far more receptive to AI in the last three years, and the core reason has nothing to do with AI itself but with health systems collapsing everywhere

Until three years ago doctors told him to go away because they were already fighting their EHR; now the NHS is struggling, France is worse, and major US hospital groups are losing money

35:40 Healthcare system collapse not ai quality explains growing doctor receptivity to ai

Sridhar Ramaswamy · Feb 10, 2025

Enterprise AI adoption will grow more gently than a steep hockey stick, but AI is already creating real and enduring value

Many previously very hard tasks are now easy — dictation and transcription, one-line meeting summaries from 25 pages of notes, internal chatbots over structured data that replace clicking through dashboards

32:54 Adoption grows gently not a hockey stick but real value is already here

Sridhar Ramaswamy · Feb 10, 2025

Large-enterprise CEOs are not rampantly skeptical of AI; their ask is 'help us create utility, tell us what is possible'

In 30 meetings at Davos, CEOs immediately grasped chatbots on document corpora and structured data, and got excited about agentic platforms and use cases like automating insurance underwriting

34:15 Enterprise ceos are eager not skeptical asking how to create utility