Nov 6, 2026 · 52 min
Enterprise AI remains stuck in its first inning
E452: Why 95% of AI Projects Fail, Anthropic’s Valuation, and Why AI Is Still in the First Inning
The episode offers a framework for judging AI businesses when adoption, pricing, data advantages, and valuations remain highly uncertain.
- 1Enterprise AI adoption depends on organizational, technical, and cultural foundations—not simply access to powerful foundation models.
- 2Vertical applications can build proprietary data, domain trust, and outcome-based advantages that horizontal AI utilities may lack.
- 3Extreme AI valuations require disciplined assumptions about market capture, competition, switching costs, and multiple adoption curves.
Don't miss
Coyne’s discussion of Inspiron shows how enterprise software can create proprietary, newly digitized data while improving senior-living and healthcare outcomes.
The brief
Joseph Coyne and David Weisburd frame enterprise AI as being in its first inning, with adoption unfolding across developers, enterprises, consumers, and eventually agents.
The discussion shifts from models to infrastructure: successful deployment requires organizational readiness, technical foundations, cultural change, and a clear path to measurable outcomes.
Inspiron, Bark, and Evercoast illustrate how AI can solve physical-world problems while generating newly digitized data that may become valuable training material.
Coyne expects software economics to move gradually beyond SaaS toward usage, tokens, and outcomes, while consultative implementations help startups learn repeatable enterprise workflows.
The Anthropic debate tests the framework: enormous upside must be weighed against competition, pricing pressure, switching costs, and uncertain market capture.
What was said on this episode
29 statements · 16 positive · 5 negative · 1 mixed · 7 neutral
Enterprise AI adoption remains at an early stage with substantial capabilities still unrealized.
“enterprise AI is really just starting to scratch the surface of what it can do within enterprise.”
Listen at 1:50
- Joseph Coyneon enterprise AI adoptionNeutral3:45
Preparing enterprise data infrastructure is half the challenge of successful AI adoption.
“getting the data to a point where from an infrastructure standpoint, it's ready to adopt AI is half the battle.”
Listen at 3:45
- Joseph Coyneon senior leadership adoption of AIPositive4:52
Senior leadership adoption helps AI reach mission-critical enterprise functions.
“If senior leadership is embracing it, it has to be in a lot of cases top down to reach sort of the most mission critical parts of the enterprise.”
Listen at 4:52
- Joseph Coyneon Inspiron longitudinal patient dataPositive10:57
Inspiron’s longitudinal data can help predict dementia and Alzheimer’s from behavior patterns.
“What that can do is actually do predictive analytics and start to predict things like dementia and Alzheimer's based on activity levels, behavior levels.”
Listen at 10:57
- Joseph Coyneon Inspiron healthcare data setNeutral14:31
Inspiron’s newly created healthcare data set has not yet been monetized.
“the data set they're creating hasn't even been monetized.”
Listen at 14:31
Bark predicts depressive episodes and self-harm or harm-to-others signals from children’s online activity.
“We can then predict with very high accuracy when somebody is experiencing depressive episodes or specifically has flagged, even if they're using tons of slang, self-harm, harm to others.”
Listen at 16:34
- Joseph Coyneon EvercoastPositive18:56
Evercoast has the world’s largest dataset of captured human 4D assets.
“They have the world's largest data set of human 4D captured assets.”
Listen at 18:56
- Joseph Coyneon Evercoast 4D human-capture datasetPositive19:34
Evercoast’s 4D human-capture dataset is highly valuable to potential partners.
“that data set is very valuable to a lot of folks.”
Listen at 19:34
- Joseph Coyneon enterprise software pricingNeutral24:08
Enterprise software pricing will gradually shift away from traditional SaaS models.
“I think we'll see a slow ramp away from it.”
Listen at 24:08
- Joseph Coyneon AI software pricing modelsPositive24:12
Usage-, consumption-, and token-based pricing models are gaining consideration.
“the usage-based, consumption-based, token-based models are all things that are in the conversation right now.”
Listen at 24:12
- Joseph Coyneon agentic enterprise workflowsPositive26:39
Startups can turn early enterprise consulting experience into repeatable agentic workflows.
“Once they've felt that and solved that, they can then replicate it and even create agentic workflows to replicate it across other use cases”
Listen at 26:39
- Joseph Coyneon AI startup product-market maturityPositive29:51
AI startups are reaching product-market maturity faster than in prior periods.
“that curve is happening shorter and faster than it's ever happened.”
Listen at 29:51
- Joseph Coyneon AI startup financing valuationsNeutral29:54
Series A and B valuations are now reaching levels formerly associated with Series D and E rounds.
“you see series A's and series B's being priced. where series Ds and Es were prior.”
Listen at 29:54
- Joseph Coyneon high-valued early-stage AI startupsNegative30:50
Very high early-stage startup valuations still imply extremely low-probability outcomes.
“you are still betting on an outcome that is extremely low probability.”
Listen at 30:50
- Joseph Coyneon venture-backed AI outcomesPositive33:42
Successful venture outcomes have shifted from billion-dollar to tens-of-billions valuations.
“the big outcomes are now tens of billions, not a billion.”
Listen at 33:42
- Joseph Coyneon foundation-model market entrantsNeutral36:02
New competitors may enter the foundation-model market.
“new entrants are not a foredrawn conclusion.”
Listen at 36:02
- Joseph Coyneon consumer willingness to pay for AI modelsNegative36:34
Consumer willingness to pay for AI models will probably decline over time.
“Over time, that probably goes down.”
Listen at 36:34
Switching costs between large language models are currently low.
“switching costs are low.”
Listen at 36:48
- Joseph Coyneon LLM market structureNegative38:41
The LLM market is unlikely to settle into only two dominant providers.
“I think it's very unlikely that there'll be two dominant players in the LLM world.”
Listen at 38:41
- Joseph Coyneon large language model marketPositive38:51
The market for large language models will continue growing over the next three to five years.
“the market size, of course, it's going to continue to grow.”
Listen at 38:51
Anthropic deserves a higher revenue valuation multiple than OpenAI because of enterprise exposure.
“I give a little boost to Anthropic in that regard.”
Listen at 39:29
- Joseph Coyneon AI model providersNegative41:06
New AI entrants could erode incumbent pricing power.
“that could erode, you know, pricing power”
Listen at 41:06
- Joseph Coyneon enterprise AI pricingMixed41:39
Enterprise AI buyers will pay more, but the eventual premium remains uncertain.
“enterprises are going to be willing to pay more, but how much more will they be willing to pay?”
Listen at 41:39
- Joseph Coyneon vertical AI companiesPositive42:44
Industry trust and domain knowledge can give vertical AI companies an early competitive lead.
“building that trust within an industry as esoteric as airlines or insurance underwriting or banking underwriting lending, that requires trust, understanding the customer process, the legacy process, understanding the data, the KPIs, et cetera, can give you an early lead”
Listen at 42:44
- Joseph Coyneon vertical AI companiesPositive43:09
Winning vertical AI companies combine deep domain expertise with rapid product development.
“companies that have domain expertise plus the ability to build product very, very fast.”
Listen at 43:09
Engineers may eventually replace salespeople, but the transition could take 10–20 years.
“It might take 10, 20 years.”
Listen at 44:42
- Joseph Coyneon salespeoplePositive45:38
Salespeople provide customer insights that inform engineering and product roadmaps.
“the salesperson can also be sort of a customer solutioning specialist that brings back unique insight or understanding to the engineers that informs product roadmap”
Listen at 45:38
- Joseph Coyneon top salespeopleNeutral48:29
AI has not yet displaced the performance advantage of top salespeople.
“The best salespeople generally are still the best salespeople.”
Listen at 48:29
- Joseph Coyneon AI-native founders doing salesPositive48:47
AI-native founders can conduct sales and business development at unprecedented scale.
“A founder doing sales can do things that they never could have done at a scale they've never done.”
Listen at 48:47
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.