Ben Thompson · Sep 24, 2024
Ben Thompson argues in Stratechery that AI’s first enterprise wave may replace workers
Markdown · JSON · Original source
Ben Thompson’s article was published on September 24, 2024. He argues that AI’s first enterprise wave may echo back-office computing: its largest gains could come from replacing workers, not making them more efficient. He sees Copilot’s challenge as getting employees to change how they work, while autonomous agents would do work in place of people. But he doubts agents can make companies more efficient without extensive oversight that raises costs, and questions whether the necessary data and infrastructure can be made viable at scale. Thompson sees ’s o1 model as a possible step toward more viable agents and is most intrigued by Palantir’s work on data integration. He expects successful AI companies may need top-down sales and long implementation efforts, with pricing tied to replaced seats. His larger conclusion is that enterprise AI may arrive before broad consumer adoption: users with volition may benefit, while for others the biggest payoff may be entertainment and gaming. Whether agents can deliver business results at a justifiable cost remains unknown.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
Ben Thompson · Article
Source details
Original source date: Sep 24, 2024 (2024-09-24T11:45:52+00:00).
Date basis: Date stated on the publisher’s page.
Added to Escape Meridian: 2026-11-17T03:33:23.210Z. First observed: 2026-11-10T01:55:39.184Z. Retrieved: 2026-11-10T01:57:38.921Z.
Referenced by
Recorded citations in public readings. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on November 10, 2026; originally published September 24, 2024. Thompson’s claims about future adoption, business results, and the viability and costs of agents are forecasts and judgments in the article, not reported outcomes.
Open question
Will enterprise deliver business results at a justifiable cost, given the oversight and data-integration work Thompson says they require?
Source-stated
Ben Thompson argues that AI’s first enterprise wave may focus on replacing workers, rather than making them more efficient, to improve companies’ bottom lines.
I think that benefit will, like the first wave of computing, come from replacing humans, not making them more efficient.
Source-stated
Thompson says Copilot’s challenge is getting employees to change how they work to benefit from the investment.
it’s another thing entirely — a much more difficult thing — to get all of your employees to change the way they work in order to benefit from your investment
Source-stated
Thompson is skeptical that autonomous agents can make a company more efficient without extensive oversight that makes the effort more expensive.
what seems unlikely to me is any sort of autonomous agent actually operating in a way that makes a company more efficient without an extensive amount of oversight that ends up making the entire endeavor more expensive.
Source-stated
Thompson predicts that successful AI companies may need top-down sales, long implementation efforts, and pricing more like seat-replacement licenses.
successful AI companies will need to go in from the top. And, instead of per-seat licenses, we may end up with something more akin to “seat-replacement” licenses
Source-stated
Thompson says the article’s publication date was September 24, 2024.
Tuesday, September 24, 2024
Source-stated
Thompson argues that AI may take longer to become a major consumer technology, while the largest payoff for many could come through entertainment and gaming.
we may be waiting longer than we expect for AI to take over the consumer space, at least at the scale of something like the smartphone or social media. That is good news for the MKBHDs of everything — the users with volition — but for everyone else the biggest payoff will probably be in areas like
Source and contributor context
Ben Thompson · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
AI agent
An AI system that chooses and carries out actions toward a goal, often using software tools.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Editorial provenance
Version 1. Publication ID: 7c90bd6e-a79d-6374-11f0-df91c5716283.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T03:28:45.307Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 312a2d56f3636e48f62ff8da08a9a5f0f64fb65abb9ae3bde0436fdda5017112. Source text hash: d1a91b84488e0a808cefa8b5da1043fc8f3a6125e4ba2537f53c352cc0d32b14.
Revision and source-check history
Version 1 · 2026-11-17T03:33:23.210Z
Ben Thompson argues in Stratechery that AI’s first enterprise wave may replace workers
Ben Thompson’s article was published on September 24, 2024. He argues that AI’s first enterprise wave may echo back-office computing: its largest gains could come from replacing workers, not making them more efficient. He sees Copilot’s challenge as getting employees to change how they work, while autonomous agents would do work in place of people. But he doubts agents can make companies more efficient without extensive oversight that raises costs, and questions whether the necessary data and infrastructure can be made viable at scale. Thompson sees ’s o1 model as a possible step toward more viable agents and is most intrigued by Palantir’s work on data integration. He expects successful AI companies may need top-down sales and long implementation efforts, with pricing tied to replaced seats. His larger conclusion is that enterprise AI may arrive before broad consumer adoption: users with volition may benefit, while for others the biggest payoff may be entertainment and gaming. Whether agents can deliver business results at a justifiable cost remains unknown.
Original source
Ben Thompson · Article
Source details
Original source date: Sep 24, 2024 (2024-09-24T11:45:52+00:00).
Date basis: Date stated on the publisher’s page.
Added to Escape Meridian: 2026-11-17T03:33:23.210Z. First observed: 2026-11-10T01:55:39.184Z. Retrieved: 2026-11-10T01:57:38.921Z.
Limits
Text as retrieved on November 10, 2026; originally published September 24, 2024. Thompson’s claims about future adoption, business results, and the viability and costs of agents are forecasts and judgments in the article, not reported outcomes.
Open question
Will enterprise deliver business results at a justifiable cost, given the oversight and data-integration work Thompson says they require?
Source-stated
Ben Thompson argues that AI’s first enterprise wave may focus on replacing workers, rather than making them more efficient, to improve companies’ bottom lines.
I think that benefit will, like the first wave of computing, come from replacing humans, not making them more efficient.
Source-stated
Thompson says Copilot’s challenge is getting employees to change how they work to benefit from the investment.
it’s another thing entirely — a much more difficult thing — to get all of your employees to change the way they work in order to benefit from your investment
Source-stated
Thompson is skeptical that autonomous agents can make a company more efficient without extensive oversight that makes the effort more expensive.
what seems unlikely to me is any sort of autonomous agent actually operating in a way that makes a company more efficient without an extensive amount of oversight that ends up making the entire endeavor more expensive.
Source-stated
Thompson predicts that successful AI companies may need top-down sales, long implementation efforts, and pricing more like seat-replacement licenses.
successful AI companies will need to go in from the top. And, instead of per-seat licenses, we may end up with something more akin to “seat-replacement” licenses
Source-stated
Thompson says the article’s publication date was September 24, 2024.
Tuesday, September 24, 2024
Source-stated
Thompson argues that AI may take longer to become a major consumer technology, while the largest payoff for many could come through entertainment and gaming.
we may be waiting longer than we expect for AI to take over the consumer space, at least at the scale of something like the smartphone or social media. That is good news for the MKBHDs of everything — the users with volition — but for everyone else the biggest payoff will probably be in areas like
Source and contributor context
Ben Thompson · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
AI agent
An AI system that chooses and carries out actions toward a goal, often using software tools.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Editorial provenance
Version 1. Publication ID: 7c90bd6e-a79d-6374-11f0-df91c5716283.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T03:28:45.307Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 312a2d56f3636e48f62ff8da08a9a5f0f64fb65abb9ae3bde0436fdda5017112. Source text hash: d1a91b84488e0a808cefa8b5da1043fc8f3a6125e4ba2537f53c352cc0d32b14.
uncertain · 2026-11-17T03:28:45.307Z · anthropic/claude-opus-5.5
Publication: 7c90bd6e-a79d-6374-11f0-df91c5716283
attribution — pass: This is a single Stratechery essay by Ben Thompson. The reading attributes its arguments to Thompson, and it attributes the quoted views of Nadella and Benioff appropriately by not presenting them as Thompson's own.
My core contention here, however, is that AI truly is a new way of computing
summary — uncertain: The summary captures the thesis but leaves central actors unexplained. Copilot is never identified as Microsoft's AI assistant for its Office software. The terms 'volition' and 'seat-replacement' pricing are also undefined, and a non-specialist cannot follow the Copilot-versus-agents contrast without them. Smallest repair: add a clause such as 'Microsoft's Copilot, an AI assistant built into its workplace software'. Also note that Thompson calls o1 a 'step change' for agents, which the reading softens to 'a possible step'.
That Update was about OpenAI’s new o1 model, which I think is a step change in terms of the viability of agents
date — pass: September 24, 2024 matches the source dateline and the UTC day of originalPublished. The source is a publisher page, so no archive disclosure is needed. The trailing related-post titles are not used, and forecasts are framed as Thompson's expectations.
Tuesday, September 24, 2024
support — pass: The source supports each claim: worker replacement, Copilot as a change-management problem, skepticism about oversight costs, top-down sales and seat-replacement licensing, and a delayed consumer payoff. The reading keeps these hedged as forecasts. The closing open question follows from Thompson's stated concerns.
Executives, however, want the benefit of AI now, and I think that benefit will, like the first wave of computing, come from replacing humans, not making them more efficient.
media — pass: No media is selected, and no verified candidates were supplied. That is acceptable.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIAn AI system that chooses and carries out actions toward a goal, often using software tools.
Definition referenceOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIAn AI system that chooses and carries out actions toward a goal, often using software tools.
Definition reference