Tab 03 · Assess and Organize · Evidence Collection Phase, 5 to 18 October 2026 · Lead: Rocky Hill

Assess and Organize

The judgment half of the evidence phase: the quality screen every source passes, the Objective Rule that maps evidence to O1 to O5, coverage by objective, and the Harvard references added this phase. The source leads live on the Evidence Collection tab, and the gap tracker now has its own Priority Gaps tab.
AssessCheck each source for quality before it enters the base. Is it credible, current, relevant, and free of unexamined commercial interest?
OrganizeSort the evidence by project objective so it is clear which objectives are well supported and which still need more.
Map: Objective RuleThe rule every lead uses to tag a source to O1 to O5.

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.

CheckQuestion to AskPasses WhenRed Flags
CredibilityWho 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.
CurrencyIs 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.
RelevanceWhich 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 transparencyCan 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 interestDoes 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 secondaryIs 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.

The rule
TEAM QUANTUM OBJECTIVE RULE (v1, 9 Oct 2026)How every source is mapped to the project objectives
Rule 1: One primary, up to two secondary. Every source gets one primary objective (where it carries the most weight) and no more than two secondary objectives. If a source seems to fit all five, it is too general to anchor a finding; tag it to the one it supports best.
Rule 2: Map by what the source shows, not its topic. Ask these five questions in order. The first "yes" is the primary objective.
  1. O1 Does it define, classify, or measure what AI value is (or is not), including financial and non financial forms?
  2. 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?
  3. O3 Does it name a condition (strategy, governance and accountability, data, operating model, workforce, leadership, change, trust) that enables or blocks value?
  4. 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?
  5. O5 Does it offer a model, standard, benchmark, maturity stage, or diagnostic structure the framework can use?
Rule 3: Required tags on every source. Objective (primary + secondary), source type, method, year, region, stakeholder group(s), time horizon (if stated), commercial interest (yes / no + note), primary or secondary source, quality screen result.
Rule 4: Well supported rule (Plan 4.1). A finding is "well supported" only if it appears in at least two different source types, including at least one peer reviewed or methodologically transparent source. Anything less is reported as a proposition.
Rule 5: Conflicting statistics. Log every headline figure with its definition, sample, and timeframe. Tag it "conflict" if another source reports a different figure for the same thing. The conflict itself is evidence for O1 and O4.
Rule 6: Sponsor and team authored sources. Tag "practitioner perspective". Never the sole support for a finding. Authorship is disclosed in the report (Plan 4.3).
Rule 7: International lens is a tag, not an objective. Regional evidence keeps its objective tag and adds a region tag. Use it to show where a barrier or form of value differs by region (O3), not as a separate study.
Rule 8: Unmappable sources. If none of the five questions is a clear yes, the source goes to the parking lot, not the evidence base.
Crosswalk: RFS questions and Research Hub gaps to objectives
Research Hub categoryPrimary objectiveSecondary objectiveWhy
Q1 What is AI valueO1O2Defines value and its forms; time and stakeholder splits feed O2.
Q2 What AI is notO4O1Capability versus delivered value is the O4 gap; "what value is not" completes the O1 definition.
Q3 How AI is perceivedO2O3Perception differs by stakeholder (O2); trust is an enabling condition (O3).
Q4 Why businesses struggleO3O4Barriers are organizational conditions; failure and scaling rates show the O4 gap.
International perspectiveTag onlyO2, O3Region changes which barriers and value types matter most.
Gap 1 Causal, measured valueO4O1Field experiments prove task level capability; the gap is whether it scales.
Gap 2 Trust, readiness, change to outcomesO3O2Conditions that enable value, viewed by stakeholder.
Gap 3 Beyond Western consultancy dataTag onlyO3Regional and policy context for conditions.
Gap 4 Governance benchmarksO5O3External anchors for maturity grades.
Gap 5 Fresher failure dataO4O3Measured abandonment rates and root causes.
Value over time horizonsO2O1Horizon based measurement of value and success.
Maturity model and diagnostic inputsO5O3Model structure and dimensions.
Source mapping: every source on the Research Hub
Filter by primary objective:
SourceTypePrimarySecondaryFlags
Enholm et al. (2022)Peer reviewedO1O3, O5Pre generative AI
Brynjolfsson, Rock and Syverson (2021)Peer reviewedO2O4Economy level
BCG (2025) Widening AI Value GapConsultancyO1O3Commercial, self reported
McKinsey (2025) State of AIConsultancyO4O3, O1Commercial, self reported
PwC (2025) AI Jobs BarometerConsultancyO1O2Commercial
Deloitte (2025) AI ROIConsultancyO2O4Commercial; conflict (15%)
Deloitte AI Institute (2026)ConsultancyO3O4Commercial
Stanford HAI (2026) AI IndexAcademic instituteO2O1International tag
Kokkonen and Krohn (no date)Industry partnerO1O2Practitioner perspective (sponsor)
Kokkonen (no date c) New GovernanceIndustry partnerO2O3Practitioner perspective (sponsor)
Gartner (2025) Agentic 40%AnalystO4O3Prediction, not measured
Gartner (2024) GenAI 30%AnalystO4O3Prediction, not measured
Marketing AI Institute (2025)MediaO4O1Critique only; conflict
California Management Review (2025)Academic practitionerO4O3Not empirical
Technology Magazine (2025) AdvisoryXMediaO4O2Secondary; DXC related
Gillespie et al. (2025)Academic + industryO2O3, O4Perception; KPMG co funded
Glikson and Woolley (2020)Peer reviewedO3O5Experimental settings
KPMG US (2025)ConsultancyO2O3Commercial
IBM IBV (2025)VendorO4O2, O3Commercial; conflict (25%)
MIT NANDA (2025)Academic affiliatedO4O3Contested; conflict (5%); cite primary PDF
DXC Technology (2025a) AI Execution GapIndustry partnerO3O4, O5Method unpublished; commercial
Snyder (2025) ForbesMediaO4O3Secondary
Szczerba (2026) ForbesMediaO3O4Secondary
Hill (2026) AI to Enterprise ProductionPractitioner guideO5O3Team authored; disclose
AXELOS (2020) ITIL 4 OCMPractice guideO3O5Not academic
Microsoft AI Economy Institute (2026)VendorO2O3International tag
Strickland (2026) IEEE SpectrumProfessional magazineO3O2International tag
Dell'Acqua et al. (2026) Jagged FrontierPeer reviewedO4O1Task level
Brynjolfsson, Li and Raymond (2025)Peer reviewedO4O1, O2Task level
Noy and Zhang (2023)Peer reviewedO4O1Task level
HBS (2025) Cybernetic TeammateWorking paperO3O4Not yet peer reviewed
McKinsey (2025) SuperagencyConsultancyO3O2Commercial
Microsoft (2025) Work Trend IndexVendorO2O3Commercial
IMF AI Preparedness IndexIntergovernmentalO3O2International tag
OECD.AI Policy ObservatoryIntergovernmentalO3O5International tag
NIST AI RMFGovernment frameworkO5O3US
ISO/IEC 42001International standardO5O3Certifiable
EU AI ActRegulationO5O3Europe
RAND (2024) Root CausesPolicy instituteO3O4Non commercial
S&P Global Market Intelligence (2025)AnalystO4O3Primary research note (updated 9 Oct)
Weill, Woerner and Sebastian (2024)AcademicO5O4Correlational; on the Hub
Becker, Knackstedt and Pöppelbuß (2009)Peer reviewedO5noneMethod source; on the Hub
Babina, Fedyk, He and Hodson (2024)Peer reviewedO4O1Firm level; observational; added 9 Oct
Klarna public case (2024 to 2025)Public caseO4O5Media documented; includes reversal; added 9 Oct
McDonald's and IBM public case (2024)Public caseO4noneAbandoned pilot; added 9 Oct
JPMorgan public case (2024 to 2025)Public caseO4O5Self reported figures; added 9 Oct
IKEA public case (2021 to 2026)Public caseO4O2Redeployment; added 9 Oct
Commonwealth Bank public case (2025)Public caseO4O2Retracted 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.

Coverage by objective
ObjectiveLeadStatusAlready strongGap to close by 18 Oct
O1 Define sustainable AI valueRoisinAdequateEnholm 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 timeSholaNeeds moreGillespie 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 conditionsMatt (Greg)StrongDXC 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 resultsRockyStrong
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 recommendationsRoisin (Ilja)Needs moreWeill 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 lensAll (Ilja support)Needs moreAdoption 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 NeedStatusWhat Is Already StrongGap to Close by 18 OctNew Leads on This Page
Q1 What is AI valueAdequateSurvey 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 notStrongGartner 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 perceivedAdequateMelbourne 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 struggleStrongDXC 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 horizonsAdequateJ 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 perspectiveNeeds MoreAdoption 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 inputsNeeds MoreDXC 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.

Updated 9 October: the S&P Global Market Intelligence entry now cites the primary Voice of the Enterprise research note, and Babina et al. (2024) was added for firm level evidence.

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).