AI is the Strategy: From Agentic AI to Autonomous Business Models onto Strategy in the Age of AI

Mickie de Wet; Ren\'e Bohnsack

REVIEW 4 major objections 6 minor 8 references

AI is the Strategy: From Agentic AI to Autonomous Business Models onto Strategy in the Age of AI

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Agentic AI is not a tool that supports strategy; it can become the strategy, as Autonomous Business Models let AI agents execute value creation, delivery, and capture with minimal human oversight.

desk verdict A conceptually useful but empirically thin paper: the ABM definition overreaches its own evidence, but the framework deserves a serious referee and a careful revision. read the letter →

arxiv 2506.17339 v2 pith:J4XCTFWB submitted 2025-06-19 cs.CY econ.GNq-fin.EC

René Bohnsack , Mickie de Wet This is my paper · ORCID

classification cs.CYecon.GNq-fin.EC

keywords AutonomousBusinessModelsagenticAIsyntheticcompetitionmodelinnovationstrategyAI-augmentedstrategicmanagementgovernance

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that agentic AI—systems that initiate, coordinate, and adapt actions on their own—can take over the core work of a business model, so that AI is not a tool supporting strategy but the strategy itself. It defines the endpoint as an Autonomous Business Model (ABM), in which AI agents are the primary actors executing value creation, delivery, and capture, with humans only setting goals and handling exceptions. The authors argue this is a distinct managerial logic, not just automation, and that it leads to 'synthetic competition,' in which AI-run firms observe, learn, and counter each other at machine speed, often beyond direct human awareness. A reader should care because, if true, strategy, leadership, and governance shift from running operations to designing, integrating, and protecting AI systems that run themselves.

What carries the argument

The central object is the Autonomous Business Model (ABM), defined as a business model in which agentic AI systems are the primary agents executing the firm's value creation, delivery, and capture logic with minimal ongoing human intervention. The concept is carried by three mechanisms: agentic execution of core logic (an 'AI Factory' drives customer interaction, operations, and response), minimal human intervention (oversight is exception-based and rare), and adaptive decision cycles (sensing, deciding, learning, and improving in continuous feedback loops). Together these mechanisms convert the firm from a hierarchy of humans and processes into a system governed by AI loops, with humans as goal-setters, governance designers, and fallback supervisors.

What would settle it

Track a firm that claims to run an Autonomous Business Model (for example, Swan AI) and count how often humans must intervene in routine operations versus exceptional cases over a quarter. If a would-be ABM needs human involvement in most novel, high-stakes, or ambiguous decisions to maintain revenue, or if errors from autonomous decisions compound without human correction, the general claim that agentic AI can be the primary agent of value capture fails.

Watch

Extended reading notes

Core claim

The central claim is that the unit of strategy is changing: agentic AI can become the primary agent of a firm's business model rather than an input to it. An Autonomous Business Model is defined as a business model in which agentic AI systems are the primary agents executing value creation, delivery, and capture with minimal ongoing human intervention, leveraging their capacity to sense, decide, and learn adaptively. The paper derives three mechanisms—agentic execution of core logic, minimal human intervention, and adaptive decision cycles—and illustrates them with the real case of Swan AI, an Israeli startup whose AI sales agents run lead generation end to end, and a hypothetical Ryanair whose operations are orchestrated by an 'AI Factory.' From these, it argues that competitive advantage shifts from static resources and positioning to data/feedback loops, and that rivalry evolves into synthetic competition between autonomous systems. If the claim holds, strategy stops being a static plan and becomes a continuous learning system embedded in AI execution.

Load-bearing premise

The framework depends on the assumption that the core activities of a business model—making, delivering, and getting paid for what it offers—can be codified well enough for agentic AI to perform them with humans stepping in only for exceptions.

Editorial extensions

If this is right

Firms can grow revenue without proportional headcount growth: the Swan AI case targets $30 million in annual recurring revenue from a three-person founding team running on agentic sales agents.
Data and learning feedback loops become the main strategic moat, making early ABM leaders progressively harder for rivals to catch because performance is tied to accumulated operational data.
Incumbents can move toward autonomy by building an AI Factory that takes over pricing, scheduling, maintenance, and customer interaction, shifting management to guardrails and exception handling.
Competition becomes synthetic: AI-run firms respond to each other in seconds, so durable positioning becomes harder and advantage depends on algorithm design, data quality, and adaptation speed.
Strategy shifts from formulation to AI stewardship—leadership means configuring and protecting self-managing systems, not micromanaging operations.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

(Editorial inference) If ABMs become common, strategy research and practice may need to treat model architecture, training data, and feedback-loop design as first-order competitive variables, not implementation details.
(Editorial inference) Synthetic competition raises the prospect of algorithmic collusion without human intent, so antitrust and liability frameworks may need to attribute agent behavior to designers and operators.
(Editorial inference) A controlled simulation of rival ABMs with different learning rates and data access could turn the paper's competitive-dynamics claims into testable predictions, e.g., whether machine-speed rivals converge on price levels or destabilize markets.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The thing to know: this paper gives strategy scholars a clean vocabulary for thinking about AI-run firms: Autonomous Business Models (ABMs), synthetic competition, and a stage model from human-driven to AI-driven. The writing is clear, the literature base is genuine (Iansiti and Lakhani, Sjodin, Casadesus-Masanell and Ricart), and the research agenda in Section 7 is genuinely useful. If you work on AI strategy or business model innovation, this is worth a read.

What it does well: the three mechanisms in Section 5 (agentic execution, minimal human intervention, adaptive decision cycles) are a sensible way to separate mere AI augmentation from something more radical. The Ryanair hypothetical is a good thought experiment for how an incumbent might move along the spectrum. The paper also honestly cites Xu et al. showing that today's AI agents fail at common sense and implicit social inference. That is a real limitation, and at least it is not hidden.

The soft spots are real and they cluster around the central definition. The ABM definition in Section 5 requires agentic AI to be the primary agent executing value creation, delivery, and capture with minimal human intervention. But the Swan AI vignette says explicitly that human salespeople step in to close deals. That means value capture is still being done by humans. The Ryanair case is hypothetical, and it mostly covers pricing and operations, not the full capture logic. So the cases do not actually reach the definition. The footnote insisting humans never cede control also sits awkwardly with 'minimal ongoing human intervention.' You can argue the definition is aspirational or ideal-typical, but the paper does not mark it as such. There is also some circularity: the ABM concept is abstracted from Swan, and then Swan is cited as evidence that ABMs exist. That is common in conceptual papers, but it is a stretch to call it independent support.

The stress-test note is on target: the gap between the definition and the evidence is the load-bearing weakness. It is fixable by either narrowing the definition to the parts the cases support (delivery and some creation) or repositioning ABMs as a directional goal and being explicit that no current firm has one fully.

Bottom line: this is a thought piece, not an empirical paper. It will probably start useful debates, and it deserves a serious referee despite the evidentiary gaps. I would send it to peer review with a major-revision request focused on the capture mechanism and the minimal-intervention boundary. Reading-group pick: maybe, for the conversation it can spark.

Referee Report

4 major / 6 minor

Summary. The paper proposes that agentic AI can become the primary executor of a firm's business model, giving rise to Autonomous Business Models (ABMs) in which AI systems handle value creation, delivery, and capture with minimal human intervention. It grounds this proposal in two vignettes: Swan AI, a startup automating B2B lead generation, and a hypothetical Ryanair reconfiguration centered on an AI Factory. The paper defines ABMs through three mechanisms (agentic execution, minimal human intervention, and adaptive learning), introduces the notion of 'synthetic competition' as a machine-speed rivalry among AI-run firms, and lays out implications for strategy, ecosystems, and leadership, together with a future research agenda.

Significance. If the central claim were fully supported, the paper would offer a useful conceptual vocabulary for a debated phenomenon, connecting agentic AI to business model theory and strategic management. Its strengths are integrative: it builds on established frameworks (Casadesus-Masanell and Ricart 2010; Iansiti and Lakhani 2020; Amit and Zott 2001) and it proposes a structured research agenda in Tables 2 and 3. However, the paper is a conceptual essay rather than an empirical test: the central definition is stronger than the evidence adduced, and some of the evidence is self-referential. The contribution is therefore best read as a research agenda or perspective, with the core empirical claim still open.

major comments (4)

[§4–§5] The formal definition in §5 requires agentic AI to be the primary agent executing 'value creation, delivery, and capture,' yet the Swan AI vignette in §4 states that human salespeople step in for complex negotiations and to close deals, and the Ryanair scenario is explicitly hypothetical and confined to pricing, scheduling, and maintenance. Consequently, no case in the manuscript instantiates AI-executed value capture; the definition is therefore stronger than the evidence. Please either narrow the definition to value creation and delivery, or provide an example in which AI agents execute contracting, billing, or revenue recognition end-to-end.
[§5, footnote 2] The definition's 'minimal ongoing human intervention' and Mechanism 2's 'exception-based' oversight sit in tension with footnote 2's assertion that 'they never cede control—human judgement and governance remain the final authority.' If humans never cede control, it is unclear how the ABM differs from a human-in-the-loop model, and the threshold for 'minimal' is not measurable. Please specify the observable conditions under which a model qualifies as an ABM (for example, the frequency or type of human interventions) and reconcile the governance language.
[§4–§6] The ABM concept is introduced by abstracting from the Swan AI vignette, and then the same vignette is used in §4 and §6 as evidence that ABMs exist; similarly, 'synthetic competition' in §6 largely restates what the definition already implies (AI agents competing autonomously). This circularity weakens the empirical footing. Please separate the conceptual proposal from its illustration and, if the claim is empirical, adduce independent examples or an explicit case-study protocol.
[§7] The paper itself cites Xu et al. (2025), who found that AI agents in a simulated software company lacked common sense, background knowledge, and the ability to infer implicit assumptions in social conversations. That finding bears directly on Mechanism 3's sensing-deciding-learning loop and on whether 'minimal ongoing human intervention' can be sustained beyond narrow, well-specified functions. The manuscript should address this as a boundary condition or scope limitation rather than treating it as a peripheral future-research question.

minor comments (6)

[§2] The text contains the typo 'Wiesinger at al. 2024'; it should read 'Wiesinger et al. 2024.'
[References / §7] The text cites Xu et al. (2025) in §7, but the reference list contains Xu et al. (2024); please align the year and ensure the intended work is cited consistently.
[Table 2] Table 2's caption repeats Table 1's caption ('Characteristics of Business Models Across the Autonomy Path'); it should be renamed to reflect the future research agenda it actually presents.
[§4 / References] The paper alternates between getswan.ai and getswan.com when naming the company and its website; please standardize the domain.
[Figures 1–2] Figures 1 and 2 reproduce CEO statements without explicit source or permission details; please add full citations and permissions information.
[Table 1] In Table 1, the ABM row labels its governance logic as 'AI-driven governance,' yet §5 and footnote 2 state that humans set goals and retain final authority; this inconsistency should be clarified.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-definitional loop: ABM is induced from the vignettes and then re-applied to them; no load-bearing self-citation or forced prediction.

self definitional [Section 4 (Swan AI vignette) and Section 5 (ABM definition)]
"Building on the aforementioned vignettes, we define an Autonomous Business Model (ABM) as a business model in which agentic AI systems are the primary agents executing the firm’s value creation, delivery, and capture logic, with minimal ongoing human intervention, by leveraging their capacity to sense, decide, and learn adaptively. ... Swan AI thus operates at the edge of an ABM: its core operational processes—the very thing clients pay for—are autonomous."

The ABM construct is explicitly induced from the Swan AI and Ryanair vignettes ('Building on the aforementioned vignettes'), and the same Swan AI vignette is then classified as an instance ('Swan AI thus operates at the edge of an ABM'). The example is therefore both the input that fixes the content of the construct and the output cited as evidence that the construct exists. This is a definitional loop rather than an independent empirical test. It is not load-bearing for the paper's conceptual contribution because the cases are labeled illustrative and Swan is only 'at the edge' of an ABM (humans still close deals), but as evidence for the reality of ABMs, the case reduces to the definition's source.

full rationale

This paper is a conceptual theory-building piece rather than an empirical prediction exercise, so most of the derivation chain is not circular. The ABM construct is transparently built from the two vignettes ('Building on the aforementioned vignettes...'), and the Swan AI case is then called 'the edge of an ABM.' That is a mild definitional loop in the sense that the same case both fixes the construct and is cited as an instance of it. However, the loop is not load-bearing: the paper labels the cases as illustrative, qualifies Swan as only 'at the edge' (humans still close deals), and the Ryanair case is explicitly hypothetical. The framework's contribution is the generalization to a three-mechanism model, which has independent content beyond any single case. There are no load-bearing self-citations; authority is drawn from external sources such as Iansiti and Lakhani (2020) and Casadesus-Masanell and Ricart (2010). The 'synthetic competition' label is a deductive consequence of the ABM definition rather than an independently tested prediction, and the paper itself cites algorithmic trading as an external anchor. The manuscript's own caveats (footnote 2 on retained human authority; Section 7 citing Xu et al. on agentic AI limitations) reduce the empirical force of the claims but do not create additional circularity. Overall, no significant circularity beyond the minor illustration loop, hence score 2.

Assumptions & free parameters 0 free parameters · 3 assumptions · 2 invented entities

No numerical parameters are fitted or hand-chosen; the paper is purely conceptual. The load-bearing assumptions concern AI capabilities and the decomposability of business execution, both asserted in the text but not independently validated.

assumptions (3)

domain assumption Agentic AI systems are capable of autonomous decision-making and executing actions toward predefined goals with minimal oversight.
Stated in Section 2 as the basis of agentic AI; this capability is not demonstrated in the paper and anchors the feasibility of ABMs.
domain assumption Business model execution can be decomposed into mechanisms (value creation, delivery, capture) that AI agents can operate with only exception-based human intervention.
Introduced via the self-driving car analogy in Section 3 and formalized in Section 5 (Mechanisms 1 and 2); if false, ABMs cannot generalize beyond narrow tasks.
domain assumption Data feedback loops create compounding, defensible competitive advantage.
Section 6 argues that more operational data improves ABM performance and creates moats; this is treated as given, drawing on Iansiti and Lakhani (2020), without independent evidence in this paper.

invented entities (2)

Autonomous Business Model (ABM)
purpose: Labels a hypothesized governance regime in which AI agents execute core business logic with minimal human intervention.
Defined in Section 5 through abstraction from two vignettes; no falsifiable handle outside the conceptual definition is provided.
Synthetic competition
purpose: Describes a predicted competitive regime among AI-run business models acting at machine speed.
Introduced in Section 6 as a consequence of ABMs; no measurement or historical precedent is given beyond an analogy to algorithmic trading.

how reviews work

0 comments

Cite this review

Pith. "Pith review of AI is the Strategy: From Agentic AI to Autonomous Business Models onto Strategy in the Age of AI." pith.science (2026). https://pith.science/paper/J4XCTFWB

@misc{pith2026250617339,
  author       = {Pith},
  title        = {Pith review of: AI is the Strategy: From Agentic AI to Autonomous Business Models onto Strategy in the Age of AI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J4XCTFWB}},
  note         = {Machine review of arXiv:2506.17339}
}

read the original abstract

This article develops the concept of Autonomous Business Models (ABMs) as a distinct managerial and strategic logic in the age of agentic AI. While most firms still operate within human-driven or AI-augmented models, we argue that we are now entering a phase where agentic AI (systems capable of initiating, coordinating, and adapting actions autonomously) can increasingly execute the core mechanisms of value creation, delivery, and capture. This shift reframes AI not as a tool to support strategy, but as the strategy itself. Using two illustrative cases, getswan.ai, an Israeli startup pursuing autonomy by design, and a hypothetical reconfiguration of Ryanair as an AI-driven incumbent, we depict the evolution from augmented to autonomous business models. We show how ABMs reshape competitive advantage through agentic execution, continuous adaptation, and the gradual offloading of human decision-making. This transition introduces new forms of competition between AI-led firms, which we term synthetic competition, where strategic interactions occur at rapid, machine-level speed and scale. It also challenges foundational assumptions in strategy, organizational design, and governance. By positioning agentic AI as the central actor in business model execution, the article invites us to rethink strategic management in an era where firms increasingly run themselves.

Figures

Figures reproduced from arXiv: 2506.17339 by the authors.

**Figure 2.** Report on memo of Shopify CEO These announcements, although different in context—a hyper-growth startup and a tech incumbent—point to a profound shift. Companies are moving beyond simply integrating generative AI to assist employees with tasks like drafting emails, creating AI is the Strategy: From Agentic AI to Autonomous Business Models onto Strategy in the Age of AI René Bohnsack* & Mickie de Wet *Corresponding a… view at source ↗

**Figure 3.** Ryanair’s business model (cf. Casadesus-Masanell and Ricart 2010), illustrating the link between Choices and Consequences [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 4.** Hypothetical Future Ryanair Business Model with AI Factory at the Core [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: The Path towards Autonomous Business Models [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

**Figure 6.** Figure 6: Strategic Loops for Competing with Autonomous Business Models [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

8 extracted references · 5 canonical work pages

[1]

Aguilera RV , Jackson G (2003) The cross national diversity of corporate governance: Dimensions and determinants. Acad. Manag. Rev. 28(3): 447–465. Aldrich H E, Ruef M (2006) Organizations evolving (2nd ed.). SAGE Publications. Amazon Web Services (2019) Ryanair improves customer support using Amazon SageMaker [Case study]. Amazon Web Services. https://aw...

work page 2003
[2]

Beyond data: How AI is both predicting and creating consumer trends. Forbes. https://www.forbes.com/councils/ forbesbusinesscouncil/2025/04/02/beyond-data- how-ai-is-both-predicting-and-creating- consumer-trends/. Markides C (2013) Business model innovation: What can the ambidexterity literature teach us? Acad. Manag. Perspect. 27(4): 313–323. Mohammed IA...

work page 2013
[7]

Google Wright PM, McMahan GC (1992) Theoretical perspectives for strategic human resource management

Agents [White paper]. Google Wright PM, McMahan GC (1992) Theoretical perspectives for strategic human resource management. J. Manag. 18(2): 295–320. Xu FF, Song Y , Li B, Tang Y , Jain K, Bao M, Wang ZZ, Zhou X, Guo Z, Cao M, Yang M, Lu HY , Martin A, Su Z, Maben L, Mehta R, Chi W, Jang L, Xie Y , … Neubig G (2024, December

work page 1992
[10]

Using large language models to generate, validate, and apply user intent taxonomies. arXiv. https:// doi.org/10.48550/arXiv.2309.13063 Shane S, Venkataraman S (2000) The promise of entrepreneurship as a field of research. Acad. Manag. Rev. 25(1): 217–226. Sjödin D, Parida V , Palmié M, Wincent J (2021) How AI capabilities enable business model innovation:...

work page Pith review arXiv doi:10.48550/arxiv.2309.13063 2000
[18]

Theagentcompany: Benchmarking llm agents on consequential real world tasks [Preprint]. arXiv. https://arxiv.org/abs/ 2412.14161 Zeng Y , Guo B, Zhang D (2021) Governance of artificial intelligence: Frameworks and practices. Philos. Technol.34(3): 543–567. Zott C, Amit R (2010) Business model design: An activity system perspective. Long Range Plann. 43(2–3...

arXiv 2021
[20]

TechInformed

WestJet and Ryanair CEOs: How emergencies are forcing airlines to modernise. TechInformed. https:// techinformed.com/technical-emergencies- forced-airlines-to-digitize/. Gal MS, Elkin-Koren N (2017) Algorithmic consumers. Harv. J.L.& Tech. 30(2): 309–353. Gerlich M (2025) AI tools in society: Impacts on cognitive offloading and the future of critical thin...

doi:10.3390/soc15010006 2017
[206]

Competing in the age of AI: Strategy and leadership when algorithms and networks run the world

Iansiti M, Lakhani KR (2020). Competing in the age of AI: Strategy and leadership when algorithms and networks run the world. Harvard Business Review Press. Jacobides, MG, Cennamo C, Gawer A (2018) Towards a theory of ecosystems. Strateg. Manag. J. 39(8): 2255–2276. Jacobides MG, Brusoni S, Cennamo C, Velu C (2021a) The evolutionary dynamics of ecosystem ...

work page 2020
[2453]

Bodonhelyi A, Bozkir E, Yang S, Kasneci E, Kasneci G (2024, January) User intent recognition and satisfaction with large language models: A user study with ChatGPT [Preprint]. arXiv. https://doi.org/10.48550/ arXiv.2402.02136 Bommasani R, Hudson DA, Adeli E, Altman R, Arora S, von Arx S, ... Liang P (2023) On the opportunities and risks of foundation mode...

arXiv doi:10.48550/arxiv.2402.02136 2023

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Aguilera RV , Jackson G (2003) The cross national diversity of corporate governance: Dimensions and determinants. Acad. Manag. Rev. 28(3): 447–465. Aldrich H E, Ruef M (2006) Organizations evolving (2nd ed.). SAGE Publications. Amazon Web Services (2019) Ryanair improves customer support using Amazon SageMaker [Case study]. Amazon Web Services. https://aw...

work page 2003

[2] [2]

Beyond data: How AI is both predicting and creating consumer trends. Forbes. https://www.forbes.com/councils/ forbesbusinesscouncil/2025/04/02/beyond-data- how-ai-is-both-predicting-and-creating- consumer-trends/. Markides C (2013) Business model innovation: What can the ambidexterity literature teach us? Acad. Manag. Perspect. 27(4): 313–323. Mohammed IA...

work page 2013

[3] [7]

Google Wright PM, McMahan GC (1992) Theoretical perspectives for strategic human resource management

Agents [White paper]. Google Wright PM, McMahan GC (1992) Theoretical perspectives for strategic human resource management. J. Manag. 18(2): 295–320. Xu FF, Song Y , Li B, Tang Y , Jain K, Bao M, Wang ZZ, Zhou X, Guo Z, Cao M, Yang M, Lu HY , Martin A, Su Z, Maben L, Mehta R, Chi W, Jang L, Xie Y , … Neubig G (2024, December

work page 1992

[4] [10]

Using large language models to generate, validate, and apply user intent taxonomies. arXiv. https:// doi.org/10.48550/arXiv.2309.13063 Shane S, Venkataraman S (2000) The promise of entrepreneurship as a field of research. Acad. Manag. Rev. 25(1): 217–226. Sjödin D, Parida V , Palmié M, Wincent J (2021) How AI capabilities enable business model innovation:...

work page Pith review arXiv doi:10.48550/arxiv.2309.13063 2000

[5] [18]

Theagentcompany: Benchmarking llm agents on consequential real world tasks [Preprint]. arXiv. https://arxiv.org/abs/ 2412.14161 Zeng Y , Guo B, Zhang D (2021) Governance of artificial intelligence: Frameworks and practices. Philos. Technol.34(3): 543–567. Zott C, Amit R (2010) Business model design: An activity system perspective. Long Range Plann. 43(2–3...

arXiv 2021

[6] [20]

TechInformed

WestJet and Ryanair CEOs: How emergencies are forcing airlines to modernise. TechInformed. https:// techinformed.com/technical-emergencies- forced-airlines-to-digitize/. Gal MS, Elkin-Koren N (2017) Algorithmic consumers. Harv. J.L.& Tech. 30(2): 309–353. Gerlich M (2025) AI tools in society: Impacts on cognitive offloading and the future of critical thin...

doi:10.3390/soc15010006 2017

[7] [206]

Competing in the age of AI: Strategy and leadership when algorithms and networks run the world

Iansiti M, Lakhani KR (2020). Competing in the age of AI: Strategy and leadership when algorithms and networks run the world. Harvard Business Review Press. Jacobides, MG, Cennamo C, Gawer A (2018) Towards a theory of ecosystems. Strateg. Manag. J. 39(8): 2255–2276. Jacobides MG, Brusoni S, Cennamo C, Velu C (2021a) The evolutionary dynamics of ecosystem ...

work page 2020

[8] [2453]

Bodonhelyi A, Bozkir E, Yang S, Kasneci E, Kasneci G (2024, January) User intent recognition and satisfaction with large language models: A user study with ChatGPT [Preprint]. arXiv. https://doi.org/10.48550/ arXiv.2402.02136 Bommasani R, Hudson DA, Adeli E, Altman R, Arora S, von Arx S, ... Liang P (2023) On the opportunities and risks of foundation mode...

arXiv doi:10.48550/arxiv.2402.02136 2023