AI capability at startups is multidimensional: firms that combine data/tech stacks, AI talent and agile risk-taking report stronger performance in Qatar's nascent ecosystem. The evidence is correlational — based on a cross-sectional founder survey and PLS-SEM — so linkages should be interpreted as associations, not proven causal effects.

AI Capability of Startups in Qatar

Savanid Vatanasakdakul, Deema Al-Mohanadi, Chadi Aoun · July 05, 2026 · Journal of the Association for Information Systems

openalex correlational low evidence 7/10 relevance Summary only summary available; pdf_status=paywall Source PDF

Using survey data from early-stage AI startups in Qatar and PLS-SEM, the study finds that tangible (data/stack/funding), human (technical skills/leadership), and intangible (agility/risk proclivity) resources jointly form an AI capability that is positively associated with self-reported venture performance.

This study develops and empirically examines a startup-specific model of Artificial Intelligence (AI) capability in emerging innovation ecosystems. Although AI capability is increasingly recognized as a driver of innovation and competitive performance, prior research has largely focused on established firms, with limited attention to startups. Drawing on Resource-Based Theory (RBT), this study adapts Mikalef and Gupta’s (2021) AI capability framework to reflect the structural fluidity, resource constraints, and experimentation-driven nature of early-stage ventures in Qatar. The proposed model conceptualizes AI capability as a multidimensional construct comprising tangible (AI data modeling, AI stack, funding), human (technical AI skills, leadership in AI), and intangible (agile execution, risk proclivity) factors. Using survey data from AI startups, the study will employ PLS-SEM to examine the relationships between these factors, AI capability, and venture performance. The findings contribute to AI and digital entrepreneurship research and provide practical implications for strengthening AI readiness in emerging ecosystems.

Summary

Main Finding

This paper develops a startup-specific, multidimensional model of AI capability tailored to emerging innovation ecosystems (Qatar) and proposes to empirically test how that capability links to venture performance. Rather than reporting finalized empirical estimates, the study’s primary contribution is a theory-driven adaptation of Mikalef & Gupta (2021) that reconceptualizes AI capability for early-stage ventures as composed of tangible, human, and intangible factors, and outlines a PLS-SEM strategy to evaluate their direct and mediated relationships with performance.

Key Points

Theoretical framing: Builds on Resource-Based Theory (RBT) and adapts an established AI capability framework to startups’ structural fluidity, resource constraints, and experimentation-driven behavior.
Multidimensional AI capability:
- Tangible: AI data modeling, AI stack (infrastructure/tools), funding.
- Human: Technical AI skills, leadership in AI.
- Intangible: Agile execution, risk proclivity.
Hypothesized structure: These three factor groups jointly form startup AI capability, which in turn affects venture performance (presumably directly and possibly mediating effects).
Contextual focus: Early-stage AI startups operating in an emerging ecosystem (Qatar) — emphasizes local ecosystem constraints and opportunities.
Empirical approach: Planned use of survey data from AI startups and Partial Least Squares Structural Equation Modeling (PLS-SEM) to estimate relationships.
Contribution: Extends AI capability literature beyond established firms to startups; offers operationalizable dimensions for measuring AI readiness in nascent ecosystems; yields practical guidance for founders, investors, and policymakers.

Data & Methods

Data source: Cross-sectional survey of AI startups (sample and response details not provided in the summary).
Constructs / measures (as described):
- Tangible resources: metrics for data modeling capacity, AI stack maturity, and funding level.
- Human resources: measures of in-house technical AI skills and AI leadership.
- Intangibles: measures of agile execution practices and organizational risk proclivity/attitude.
- Outcome: venture performance (likely financial and/or innovation/performance proxies—details not specified).
Analysis plan:
- PLS-SEM to assess measurement model (reliability, validity) and structural paths between dimension groups → overall AI capability → venture performance.
- Potential for mediation/moderation tests (implied by model structure).
Methodological considerations (implicit/likely):
- Cross-sectional survey design — limits causal claims.
- Risk of common-method bias if independent and dependent variables collected from same respondents.
- Generalizability constrained by geographic/contextual focus (Qatar).
- Recommended controls (not specified in summary but advisable): firm age, team size, sector, funding stage.

Implications for AI Economics

Theory:
- Extends RBT-based conceptions of AI capability to startup contexts, highlighting how resource endowments and organizational behaviors shape capability formation in nascent firms.
- Provides a parsimonious, empirically testable taxonomy of AI capability dimensions suitable for micro-level economic analysis.
Empirical measurement:
- Offers operational constructs that can be used in cross-firm analyses of AI adoption, productivity, and innovation at the startup level, improving granularity beyond firm-level binary measures of AI use.
Policy and ecosystem design:
- Suggests targeted interventions for emerging ecosystems: subsidize access to AI stacks and data, build technical talent pipelines, support agile management and risk-tolerant experimentation.
- Informs ecosystem actors (incubators, public funds) on which resource gaps (tangible vs. human vs. intangible) most limit AI readiness among startups.
Investment and firm strategy:
- Signals to investors which combinations of tangible, human, and intangible attributes are theoretically most important for startup AI capability and hence potential performance.
- Guides founders on resource allocation (e.g., balancing investment in tooling vs. hiring technical talent vs. fostering agile culture).
Macroeconomic / regional effects:
- Better measurement and support of startup AI capability could accelerate innovation-led growth in emerging economies, affecting employment, productivity, and technology diffusion dynamics.
Limitations for economic inference:
- Context specificity (Qatar) and cross-sectional design mean findings must be validated across regions and with longitudinal data before strong causal or general equilibrium claims are made.

Assessment

Paper Typecorrelational Evidence Strengthlow — Findings are based on cross-sectional, self-reported survey data and PLS-SEM associations, which cannot establish causality; results are vulnerable to common-method bias, reverse causality, selection into the sample, and unobserved confounders. Methods Rigormedium — The study builds on an established theoretical framework and uses PLS-SEM (appropriate for latent constructs and smaller samples) to test a multidimensional construct, but relies on cross-sectional self-reports, likely limited controls, and does not use identification strategies (e.g., instruments, experiments, panel methods) that would strengthen causal claims. SampleCross-sectional survey of early-stage AI startups in Qatar (respondents likely founders/managers), measuring tangible (data/modeling, AI stack, funding), human (technical AI skills, AI leadership) and intangible (agility, risk proclivity) resources and self-reported measures of venture performance; exact sample size and sampling frame not reported in the abstract. Themesinnovation org_design IdentificationNo causal identification — the paper uses a cross-sectional survey and PLS-SEM to estimate associations between measured AI capability dimensions (tangible, human, intangible) and self-reported venture performance; identification rests on theoretical model specification and observed covariation rather than exogenous variation or quasi-experimental design. GeneralizabilitySingle-country study (Qatar) — limited geographic/general institutional external validity, Focus on early-stage startups — not directly generalizable to established firms or other firm sizes, Possible sectoral heterogeneity within 'AI startups' reduces broad applicability, Cross-sectional and self-reported data limit inference about dynamics and true performance outcomes, Potential non-representative sample or small sample size (common in startup surveys)

Claims (6)

Claim	Direction	Outcome	Confidence & Evidence	Details
This study develops and empirically examines a startup-specific model of Artificial Intelligence (AI) capability in emerging innovation ecosystems. Other	positive	AI capability (as a construct representing startups' AI readiness)	Reading fidelity high Study strength medium	not reported 0.3
AI capability is increasingly recognized as a driver of innovation and competitive performance. Innovation Output	positive	innovation and competitive performance	Reading fidelity high Study strength medium	not reported 0.3
Prior research has largely focused on established firms, with limited attention to startups. Other	negative	research coverage of firms (established vs. startups)	Reading fidelity high Study strength medium	not reported 0.3
The proposed model conceptualizes AI capability as a multidimensional construct comprising tangible (AI data modeling, AI stack, funding), human (technical AI skills, leadership in AI), and intangible (agile execution, risk proclivity) factors. Other	positive	AI capability (multi-dimensional construct)	Reading fidelity high Study strength medium	not reported 0.3
Using survey data from AI startups in Qatar, the study will employ PLS-SEM to examine the relationships between these factors, AI capability, and venture performance. Firm Productivity	mixed	venture performance	Reading fidelity high Study strength low	not reported 0.15
The findings contribute to AI and digital entrepreneurship research and provide practical implications for strengthening AI readiness in emerging ecosystems. Adoption Rate	positive	AI readiness / adoption in emerging ecosystems	Reading fidelity high Study strength speculative	not reported 0.05