Skip to content

Sources

Where our numbers come from

Every study, forecast and number of our own on this website, with publisher, date and link. Forecasts are marked as forecasts.

Go to our own analysis

Industry and AI

  1. UNIDO (data: International Labour Organization)

    Nov 2025

    International Yearbook of Industrial Statistics 2025(opens in a new tab)

    492 million people worked in manufacturing worldwide in 2023, 14.1% of global employment (section 3.4; UNIDO's source: ILO modelled estimates, November 2024).

    Used onOur pathAbout us

  2. World Bank

    8 Oct 2026

    World Development Indicators: Manufacturing, value added (current US$), World (NV.IND.MANF.CD)(opens in a new tab)

    Manufacturing created $17.6 trillion of value added worldwide in 2025 (current US dollars), 14.9% of world GDP.

    Used onOur pathAbout us

  3. United Nations, Department of Economic and Social Affairs

    Jul 2024

    World Population Prospects 2024: Summary of Results(opens in a new tab)

    In the 63 countries and areas whose population has already peaked (28% of the world's population, China, Germany, Japan, Korea and Russia among them), the share of people aged 65 or older nearly doubles, from 17% in 2024 to 33% in 2054. By the late 2070s, 2.2 billion people worldwide will be 65 or older, more than all children under 18.

    Used onOur pathAbout us

  4. OECD

    Nov 2026

    OECD Digital Economy and Society Outlook 2026(opens in a new tab)

    Across OECD countries, 52% of large firms (250 or more employees) used AI in 2025 or the latest year, 17.5% of small firms (10 to 49 employees) and 20% of all firms (size split: OECD ICT Access and Usage by Businesses database, read on 11 November 2026).

    Used onOur pathAbout us

  5. Statistisches Bundesamt (Destatis)

    2025

    Nutzung von IKT in Unternehmen (Erhebung 2025)(opens in a new tab)

    26% of German firms used AI in 2025; 44% of non-users cite data availability or quality as a reason (Destatis).

    Used onBlog

  6. Eurostat

    11 Dec 2025

  7. Statistisches Bundesamt (Destatis)

    23 Jun 2026

  8. Institut der deutschen Wirtschaft (IW), KOFA

    Feb 2026

  9. Statistisches Bundesamt (Destatis)

    15 Jan 2026

  10. Deutsche Bundesbank

    14 Jul 2025

  11. Eurostat

    8 May 2026

  12. Eurostat

    Jan 2026

  13. Official Journal of the European Union (EUR-Lex)

    29 Jun 2023

  14. European Commission

    11 Sept 2026

  15. European Commission

    Sept 2026

Value and success of AI

  1. Boston Consulting Group

    30 Sept 2025

    The Widening AI Value Gap: Build for the Future 2025(opens in a new tab)

    Only 5% of companies achieve AI value at scale; 60% report minimal revenue and cost gains from AI despite substantial investment (survey of 1,250 companies in 68 countries, answered by CxOs and senior executives).

    Used onAbout us

  2. Deloitte

    21 Jan 2026

    State of AI in the Enterprise 2026: The Untapped Edge(opens in a new tab)

    73% of companies name data privacy and security as an AI risk, more than any other risk (3,235 business and IT leaders in 24 countries, August to September 2025).

    Used onAbout us

  3. Deloitte

    Jan 2026

    State of AI in the Enterprise 2026: The untapped edge (global report)(opens in a new tab)

    84% of companies have not redesigned jobs or the nature of work around AI (Deloitte, 3,235 leaders).

    Used onHome

  4. Boston Consulting Group

    30 Sept 2025

    AI Leaders Outpace Laggards with Double the Revenue Growth and 40% More Cost Savings(opens in a new tab)

    70% of AI's potential value sits in core business functions like sales, manufacturing and supply chain (BCG).

    Used onROI calculator

  5. McKinsey & Company, QuantumBlack

    25 Aug 2026

    The state of AI in 2026: On the road to ROI(opens in a new tab)

    Only 37% of firms report any EBIT impact from AI, unchanged from 2025, although more of them now scale AI.

    Used onROI calculator

  6. KfW Research

    11 Feb 2026

    KfW Research: Künstliche Intelligenz kommt im Mittelstand immer häufiger zum Einsatz(opens in a new tab)

    KfW: 20% of German SMEs use AI; use is more likely in firms with a digitalisation strategy.

    Used onROI calculator

  7. Helmut Krcmar, Springer Gabler

    2015

    Informationsmanagement (6th edition)(opens in a new tab)

    Krcmar (2015): ICT offers not only efficiency gains but, above all, room to shape organisation and business models.

    Used onROI calculatorResearch

  8. Statistisches Bundesamt (Destatis)

    29 Apr 2026

  9. National Bureau of Economic Research (Brynjolfsson, Rock, Syverson)

    Oct 2018

  10. National Bureau of Economic Research (Brynjolfsson, Li, Raymond)

    Apr 2023

  11. Salesforce

    15 May 2025

Data

  1. Gartner

    26 Feb 2025

    Forecast
    Lack of AI-Ready Data Puts AI Projects at Risk (Q&A with Roxane Edjlali)(opens in a new tab)

    Through 2026, Gartner expects organisations to abandon 60% of AI projects that lack AI-ready data.

    Used onHomeData LayerResearch

  2. IBM Institute for Business Value

    6 May 2025

    IBM Study: CEOs Double Down on AI While Navigating Enterprise Hurdles(opens in a new tab)

    68% of CEOs call integrated, enterprise-wide data architecture critical; 50% report disconnected, piecemeal technology.

    Used onData Layer

  3. Bitkom e.V.

    14 Sept 2026

    Erstmals nutzt die Mehrheit der Unternehmen KI(opens in a new tab)

    For German AI users, data preparation (50%) and integration (41%) are big cost items; licenses only 21% (Bitkom).

    Used onData Layer

  4. Jennifer Rowley, Journal of Information Science (SAGE)

    Apr 2007

    The wisdom hierarchy: representations of the DIKW hierarchy(opens in a new tab)

    The DIKW hierarchy (data, information, knowledge, wisdom; Ackoff 1989) is a fundamental, widely recognised model.

    Used onResearch

  5. Martin Fowler

    29 Jun 2004

  6. European Commission

    5 Nov 2026

  7. SAP News Center

    7 Oct 2026

Rights and security

  1. OWASP GenAI Security Project

    17 Nov 2024

    LLM02:2025 Sensitive Information Disclosure (OWASP Top 10 for LLM Applications 2025)(opens in a new tab)

    OWASP (2025): protect sensitive data in LLM apps with least-privilege access, only what a user or process needs.

    Used onPermission Layer

  2. Gartner

    20 Oct 2026

    Forecast
    Gartner Identifies the Top Strategic Technology Trends for 2027(opens in a new tab)

    Gartner expects that by 2028, 25% of enterprise breaches will trace back to abused AI agents with unmanaged privileges.

    Used onPermission Layer

  3. Deloitte

    24 Apr 2026

    Business and IT leaders report AI agents are scaling faster than their guardrails(opens in a new tab)

    Only 21% say they have a mature governance model for agentic AI, while 74% expect to use agents by 2027 (Deloitte).

    Used onHomePermission Layer

  4. NIST

    Aug 2020

    Zero Trust Architecture (SP 800-207)(opens in a new tab)

    NIST zero trust: every access is granted per request with the least privilege needed for the task.

    Used onResearch

  5. OWASP GenAI Security Project

    17 Nov 2024

    LLM06:2025 Excessive Agency (OWASP Top 10 for LLM Applications 2025)(opens in a new tab)

    OWASP (2025) names excessive agency a top LLM risk: too many functions, permissions or too much autonomy.

    Used onResearch

AI, agents and oversight

  1. European Commission, AI Act Service Desk

    13 Jun 2024

    Article 14: Human oversight(opens in a new tab)

    AI Act Art. 14: high-risk AI must be built so that people can effectively oversee it while it is in use.

    Used onIntelligence LayerResearch

  2. Gartner

    25 Jun 2025

    Forecast
    Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027(opens in a new tab)

    Gartner expects over 40% of agentic AI projects to be cancelled by the end of 2027: costs, unclear value, weak risk controls.

    Used onIntelligence Layer

  3. Forrester

    Feb 2026

    The AEGIS Framework: Enterprise Guardrails For Securing Agentic AI(opens in a new tab)

    Forrester: Zero Trust limits access; for agents, 'least agency' also limits the decisions and actions they may take.

    Used onIntelligence Layer

  4. NIST

    26 Jan 2023

    AI Risk Management Framework(opens in a new tab)

    NIST's voluntary AI Risk Management Framework builds trustworthiness into the design, use and evaluation of AI.

    Used onResearch

  5. Official Journal of the European Union (EUR-Lex)

    8 Jul 2026

  6. Stanford HAI

    Apr 2025

  7. Epoch AI

    12 Mar 2025

  8. METR (Kwa et al.), arXiv

    18 Mar 2025

  9. Belcak et al., arXiv

    2 Jun 2025

  10. DeepSeek-AI, Nature

    22 Jan 2025

Foundations from research

  1. Claude E. Shannon, Bell System Technical Journal 27(3)

    1948

  2. Richard Y. Wang, Diane M. Strong, Journal of Management Information Systems 12(4)

    1996

All links opened in November 2026. Some publishers show their pages only in a browser, not to automated requests.

Our own analysis

demi's analysis of 196 use cases

We coded the first 196 use cases in our catalogue by fixed rules: which data the case needs, who may see it, who signs off, what it writes back, what fixed rules decide. Then we read through every case by hand.

The results on this website

  • 100%

    196 of 196 join data from at least two systems, with a median of at least 4.

  • 83%

    163 of 196 work with data that only certain roles may see, and 43 of 196 with strictly confidential data such as personnel, contracts and margins.

  • 93%

    In 182 of 196, a person reviews or releases the result before it takes effect.

  • 97%

    143 of 196 write their result back into a business system. 139 of those 143 only after a person's sign-off.

  • 82%

    In 160 of 196, fixed rules, calculations or master data decide the part of the result that matters, not the wording of a language model.

  • 79%

    154 of 196 need all three layers at once: data from several systems, a person's sign-off and fixed rules in the result.

What the numbers do not say

  • The catalogue is built from practice and research: typical tasks in industrial companies, not a random sample of firms. It was built around tasks with connected data, so every case joining several systems is partly true by construction.
  • The codes record what a case's description requires, not what customers run today or how often a case occurs.
  • The number of systems per case is a floor: it counts the principal systems the catalogue names.
  • An AI agent did the coding by fixed rules. On 30 cases also coded blind by hand, 203 of 240 values agreed. A second human coding is still to come.

Go to the use case catalogue

Your configurator

No use cases yet

Your configurator is still empty

Add a use case from any card or case page and see what it brings per year, with the calculation. Free, no sign-up, no commitment: try it out and change any number.

Browse use cases