Assess: The Quality Screen Every Source Passes Before It Enters the Base
Run each new source through the six checks below before it is added to a tab. A source can still be used if it fails a check, but the weakness gets written down next to it, the way section 3.1 of the Project Plan already does with its Critical Consideration column. The standing rule from section 3.4: consultancy and vendor claims are triangulated against academic sources before they carry weight in findings.
| Check | Question to Ask | Passes When | Red Flags |
|---|---|---|---|
| Credibility | Who produced it and what is their track record? | Named authors or established institution; peer review or published editorial standards. | Anonymous content, AI generated blog farms, no institutional home. |
| Currency | Is it current enough for a field moving this fast? | 2024 or later for market data; older is fine for theory (J curve, trust theory) where the concept, not the number, is being used. | Pre 2023 adoption or ROI statistics presented as current. |
| Relevance | Which research question or objective does it serve? | Maps cleanly to at least one of Q1 to Q4, the international lens, or a maturity model dimension. | Interesting but unmappable; collect somewhere else, not in the base. |
| Method transparency | Can you tell how the numbers were produced? | Sample size, population, and method published; limitations acknowledged. | Headline percentages with no methodology anywhere (the DXC AdvisoryX study itself carries this caution). |
| Commercial interest | Does the publisher sell a remedy for the problem it describes? | No stake, or the stake is declared and the claim is triangulated against an independent source. | Vendor research used alone to justify a finding that benefits the vendor. |
| Primary vs secondary | Is this the original source or a retelling? | Original report, paper, or dataset cited; media retellings used only for critique and color. | Citing a news summary when the underlying report is freely available. |
Map: Objective Rule (Mapping Evidence to O1 to O5)
The objectives locked in Project Plan section 1.5. This is the shared rule every lead uses to tag a source to O1 to O5, the crosswalk from the Hub's earlier categories to the objectives, and the mapping of every source on the Hub. Apply the rule through the tag chips on each source card on the Initial Research and Evidence Collection tabs; the mapping table below is the starting state those chips were seeded from.
- O1 Does it define, classify, or measure what AI value is (or is not), including financial and non financial forms?
- O2 Does it show how value, success, or failure differs by stakeholder group, or over short, medium, or long term horizons, or how success is judged?
- O3 Does it name a condition (strategy, governance and accountability, data, operating model, workforce, leadership, change, trust) that enables or blocks value?
- O4 Does it report on more than one stage of the chain (capability, adoption, outcome, sustained value), or give a failure, abandonment, or scaling rate?
- O5 Does it offer a model, standard, benchmark, maturity stage, or diagnostic structure the framework can use?
| Research Hub category | Primary objective | Secondary objective | Why |
|---|---|---|---|
| Q1 What is AI value | O1 | O2 | Defines value and its forms; time and stakeholder splits feed O2. |
| Q2 What AI is not | O4 | O1 | Capability versus delivered value is the O4 gap; "what value is not" completes the O1 definition. |
| Q3 How AI is perceived | O2 | O3 | Perception differs by stakeholder (O2); trust is an enabling condition (O3). |
| Q4 Why businesses struggle | O3 | O4 | Barriers are organizational conditions; failure and scaling rates show the O4 gap. |
| International perspective | Tag only | O2, O3 | Region changes which barriers and value types matter most. |
| Gap 1 Causal, measured value | O4 | O1 | Field experiments prove task level capability; the gap is whether it scales. |
| Gap 2 Trust, readiness, change to outcomes | O3 | O2 | Conditions that enable value, viewed by stakeholder. |
| Gap 3 Beyond Western consultancy data | Tag only | O3 | Regional and policy context for conditions. |
| Gap 4 Governance benchmarks | O5 | O3 | External anchors for maturity grades. |
| Gap 5 Fresher failure data | O4 | O3 | Measured abandonment rates and root causes. |
| Value over time horizons | O2 | O1 | Horizon based measurement of value and success. |
| Maturity model and diagnostic inputs | O5 | O3 | Model structure and dimensions. |
| Source | Type | Primary | Secondary | Flags |
|---|---|---|---|---|
| Enholm et al. (2022) | Peer reviewed | O1 | O3, O5 | Pre generative AI |
| Brynjolfsson, Rock and Syverson (2021) | Peer reviewed | O2 | O4 | Economy level |
| BCG (2025) Widening AI Value Gap | Consultancy | O1 | O3 | Commercial, self reported |
| McKinsey (2025) State of AI | Consultancy | O4 | O3, O1 | Commercial, self reported |
| PwC (2025) AI Jobs Barometer | Consultancy | O1 | O2 | Commercial |
| Deloitte (2025) AI ROI | Consultancy | O2 | O4 | Commercial; conflict (15%) |
| Deloitte AI Institute (2026) | Consultancy | O3 | O4 | Commercial |
| Stanford HAI (2026) AI Index | Academic institute | O2 | O1 | International tag |
| Kokkonen and Krohn (no date) | Industry partner | O1 | O2 | Practitioner perspective (sponsor) |
| Kokkonen (no date c) New Governance | Industry partner | O2 | O3 | Practitioner perspective (sponsor) |
| Gartner (2025) Agentic 40% | Analyst | O4 | O3 | Prediction, not measured |
| Gartner (2024) GenAI 30% | Analyst | O4 | O3 | Prediction, not measured |
| Marketing AI Institute (2025) | Media | O4 | O1 | Critique only; conflict |
| California Management Review (2025) | Academic practitioner | O4 | O3 | Not empirical |
| Technology Magazine (2025) AdvisoryX | Media | O4 | O2 | Secondary; DXC related |
| Gillespie et al. (2025) | Academic + industry | O2 | O3, O4 | Perception; KPMG co funded |
| Glikson and Woolley (2020) | Peer reviewed | O3 | O5 | Experimental settings |
| KPMG US (2025) | Consultancy | O2 | O3 | Commercial |
| IBM IBV (2025) | Vendor | O4 | O2, O3 | Commercial; conflict (25%) |
| MIT NANDA (2025) | Academic affiliated | O4 | O3 | Contested; conflict (5%); cite primary PDF |
| DXC Technology (2025a) AI Execution Gap | Industry partner | O3 | O4, O5 | Method unpublished; commercial |
| Snyder (2025) Forbes | Media | O4 | O3 | Secondary |
| Szczerba (2026) Forbes | Media | O3 | O4 | Secondary |
| Hill (2026) AI to Enterprise Production | Practitioner guide | O5 | O3 | Team authored; disclose |
| AXELOS (2020) ITIL 4 OCM | Practice guide | O3 | O5 | Not academic |
| Microsoft AI Economy Institute (2026) | Vendor | O2 | O3 | International tag |
| Strickland (2026) IEEE Spectrum | Professional magazine | O3 | O2 | International tag |
| Dell'Acqua et al. (2026) Jagged Frontier | Peer reviewed | O4 | O1 | Task level |
| Brynjolfsson, Li and Raymond (2025) | Peer reviewed | O4 | O1, O2 | Task level |
| Noy and Zhang (2023) | Peer reviewed | O4 | O1 | Task level |
| HBS (2025) Cybernetic Teammate | Working paper | O3 | O4 | Not yet peer reviewed |
| McKinsey (2025) Superagency | Consultancy | O3 | O2 | Commercial |
| Microsoft (2025) Work Trend Index | Vendor | O2 | O3 | Commercial |
| IMF AI Preparedness Index | Intergovernmental | O3 | O2 | International tag |
| OECD.AI Policy Observatory | Intergovernmental | O3 | O5 | International tag |
| NIST AI RMF | Government framework | O5 | O3 | US |
| ISO/IEC 42001 | International standard | O5 | O3 | Certifiable |
| EU AI Act | Regulation | O5 | O3 | Europe |
| RAND (2024) Root Causes | Policy institute | O3 | O4 | Non commercial |
| S&P Global Market Intelligence (2025) | Analyst | O4 | O3 | Primary research note (updated 9 Oct) |
| Weill, Woerner and Sebastian (2024) | Academic | O5 | O4 | Correlational; on the Hub |
| Becker, Knackstedt and Pöppelbuß (2009) | Peer reviewed | O5 | none | Method source; on the Hub |
| Babina, Fedyk, He and Hodson (2024) | Peer reviewed | O4 | O1 | Firm level; observational; added 9 Oct |
| Klarna public case (2024 to 2025) | Public case | O4 | O5 | Media documented; includes reversal; added 9 Oct |
| McDonald's and IBM public case (2024) | Public case | O4 | none | Abandoned pilot; added 9 Oct |
| JPMorgan public case (2024 to 2025) | Public case | O4 | O5 | Self reported figures; added 9 Oct |
| IKEA public case (2021 to 2026) | Public case | O4 | O2 | Redeployment; added 9 Oct |
| Commonwealth Bank public case (2025) | Public case | O4 | O2 | Retracted claim; added 9 Oct |
Organize: Evidence Coverage by Objective
Organized against objectives O1 to O5 from Project Plan section 1.5, now locked, with status as of 9 October. The job of this phase is unchanged: move every objective to Strong by 18 October. Owners follow the confirmed role list.
| Objective | Lead | Status | Already strong | Gap to close by 18 Oct |
|---|---|---|---|---|
| O1 Define sustainable AI value | Roisin | Adequate | Enholm et al. definition; BCG, McKinsey, PwC value data; Stanford consumer surplus; Kokkonen on non financial value; field experiments for measured task value. | Evidence for resilience, adaptability and trust as measurable forms of value; a working definition tested against the field experiments. |
| O2 Value across stakeholders and time | Shola | Needs more | Gillespie et al., KPMG, IBM CEO, Stanford perception gap for employees, leaders and citizens; J curve and Deloitte for horizons. | Investor, regulator and technology partner views; firm level evidence of how long the J curve lag lasts; side by side comparison of how studies define success and failure. |
| O3 Organizational conditions | Matt (Greg) | Strong | DXC five barriers, McKinsey workflow redesign, BCG 10/20/70, Deloitte skills, IBM silos, RAND root causes, Superagency, Glikson and Woolley, AXELOS. | More non commercial confirmation beyond RAND; regional differences in which conditions matter most. |
| O4 Capability vs business results | Rocky | Strong what changed | Field experiments (capability); McKinsey 21% redesign and MIT shadow AI (adoption); MIT primary report, Deloitte, IBM, and the S&P primary research note (outcomes); Babina et al. firm level evidence and the J curve (sustained); five shortlisted public cases tracing all four stages. | Closed 9 October (see the what changed link). Remaining team action: pick 3 of the 5 shortlisted cases at the 14 Oct review. |
| O5 Framework and recommendations | Roisin (Ilja) | Needs more | Weill et al. maturity stages; Becker et al. procedure; NIST AI RMF, ISO/IEC 42001, EU AI Act; DXC barriers; Edge Framework phases. | One or two more published AI maturity models to compare against; shortlist of public cases for the walk through; Weill and Becker added to the Hub. |
| Cross cutting International lens | All (Ilja support) | Needs more | Adoption and trust by country; IMF, OECD, EU AI Act leads. | Non Western business value evidence (most outcome data is North America and Europe). |
Previous version (by RFS question, 3 Oct)
| Project Need | Status | What Is Already Strong | Gap to Close by 18 Oct | New Leads on This Page |
|---|---|---|---|---|
| Q1 What is AI value | Adequate | Survey evidence (BCG, McKinsey, Deloitte) and the academic definition (Enholm). | Causal, measured evidence; a working definition tested against it. | Jagged Frontier, Generative AI at Work, Noy and Zhang, Cybernetic Teammate. |
| Q2 What AI is not | Strong | Gartner agent washing, AI theater, hype critique both directions. | Academic grounding for capability limits. | Jagged frontier concept (capability is uneven, not general). |
| Q3 How AI is perceived | Adequate | Melbourne and KPMG global data, Stanford perception gap, trust theory (Glikson and Woolley). | Perception linked to behavior and outcomes; original stakeholder data. | Superagency, Work Trend Index, pulse survey, practitioner interviews. |
| Q4 Why businesses struggle | Strong | DXC five barriers, MIT learning gap, workflow redesign evidence, CEO testimony. | Non commercial confirmation; measured (not predicted) failure rates. | RAND root causes, S&P 42% abandonment, Gartner 2024 baseline. |
| Value over time horizons | Adequate | J curve theory, Deloitte one to five year agentic expectations. | Firm level evidence of the lag and recovery. | Field experiments above give the early, task level end of the curve. |
| International perspective | Needs More | Adoption and trust percentages by country. | Non Western, non consultancy data; regulation as a regional cost factor. | IMF Preparedness Index, OECD.AI Observatory, EU AI Act. |
| Maturity model and diagnostic inputs | Needs More | DXC five barriers and Enholm enablers as candidate dimensions. | External benchmarks so maturity grades are defensible, not invented. | NIST AI RMF, ISO/IEC 42001, EU AI Act risk tiers, Edge Framework phases. |
New References Added This Phase (Harvard Style)
Only the sources introduced during the evidence collection phase, including the plan sources added 9 October and the primary MIT NANDA report. The foundation references stay on the Initial Research tab. Verify dates and details against each original publication before final submission.
Babina, T., Fedyk, A., He, A. and Hodson, J. (2024) 'Artificial intelligence, firm growth, and product innovation', Journal of Financial Economics, 151. Available at: https://ideas.repec.org/a/eee/jfinec/v151y2024ics0304405x2300185x.html (Accessed: 9 October 2026).
Becker, J., Knackstedt, R. and Pöppelbuß, J. (2009) 'Developing maturity models for IT management: A procedure model and its application', Business & Information Systems Engineering, 1(3), pp. 213 to 222. doi: 10.1007/s12599-009-0044-5.
Braun, V. and Clarke, V. (2006) 'Using thematic analysis in psychology', Qualitative Research in Psychology, 3(2), pp. 77 to 101. doi: 10.1191/1478088706qp063oa.
Brynjolfsson, E., Li, D. and Raymond, L. (2025) 'Generative AI at Work', The Quarterly Journal of Economics, 140(2), pp. 889 to 942. doi: 10.1093/qje/qjae044.
Dell'Acqua, F., McFowland, E., Mollick, E., Lifshitz, H., Kellogg, K.C., Rajendran, S., Krayer, L., Candelon, F. and Lakhani, K.R. (2026) 'Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality', Organization Science. doi: 10.1287/orsc.2025.21838.
European Commission (2024) Regulatory Framework on Artificial Intelligence. Available at: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai (Accessed: 3 October 2026).
Gartner (2024) Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025 [Press release]. Available at: https://www.gartner.com/en/newsroom/press-releases/2024-07-29-gartner-predicts-30-percent-of-generative-ai-projects-will-be-abandoned-after-proof-of-concept-by-end-of-2025 (Accessed: 3 October 2026).
Harvard Business School AI Institute (2025) The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise. Available at: https://aiinstitute.hbs.edu/?p=26305 (Accessed: 3 October 2026).
International Monetary Fund (2024) AI Preparedness Index. Available at: https://www.imf.org/external/datamapper/datasets/AIPI (Accessed: 3 October 2026).
International Organization for Standardization (2023) ISO/IEC 42001:2023 Information Technology, Artificial Intelligence, Management System. Available at: https://www.iso.org/standard/81230.html (Accessed: 3 October 2026).
McKinsey & Company (2025) Superagency in the Workplace: Empowering People to Unlock AI's Full Potential. Available at: https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work (Accessed: 3 October 2026).
Microsoft (2025) Work Trend Index Annual Report. Available at: https://www.microsoft.com/en-us/worklab/work-trend-index (Accessed: 3 October 2026).
Microsoft AI Economy Institute (2026) Global AI Adoption in 2025: A Widening Digital Divide. Available at: https://www.microsoft.com/en-us/research/wp-content/uploads/2026/01/Microsoft-AI-Diffusion-Report-2025-H2.pdf (Accessed: 9 October 2026).
MIT NANDA (2025) The GenAI Divide: State of AI in Business 2025. Available at: https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf (Accessed: 9 October 2026).
National Institute of Standards and Technology (2023) AI Risk Management Framework (AI RMF 1.0). Available at: https://www.nist.gov/itl/ai-risk-management-framework (Accessed: 3 October 2026).
Noy, S. and Zhang, W. (2023) 'Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence', Science, 381(6654), pp. 187 to 192. doi: 10.1126/science.adh2586.
OECD (2026) OECD.AI Policy Observatory. Available at: https://oecd.ai (Accessed: 3 October 2026).
RAND Corporation (2024) The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed. Available at: https://www.rand.org/pubs/research_reports/RRA2680-1.html (Accessed: 3 October 2026).
Snyder, H. (2019) 'Literature review as a research methodology: An overview and guidelines', Journal of Business Research, 104, pp. 333 to 339. doi: 10.1016/j.jbusres.2019.07.039.
S&P Global Market Intelligence (2025) AI Experiences Rapid Adoption, but with Mixed Outcomes: Highlights from VotE: AI and Machine Learning. Available at: https://www.spglobal.com/market-intelligence/en/news-insights/research/ai-experiences-rapid-adoption-but-with-mixed-outcomes-highlights-from-vote-ai-machine-learning (Accessed: 9 October 2026).
Weill, P., Woerner, S.L. and Sebastian, I.M. (2024) Building Enterprise AI Maturity. MIT CISR. Available at: https://cisr.mit.edu/publication/2024_1201_EnterpriseAIMaturityModel_WeillWoernerSebastian (Accessed: 9 October 2026).