# Gartner Data & Analytics London May 2026: My Honest No-BS Takeaways ## Knowledge Graph Analysis — Juan Sequeda **Source:** [juansequeda.substack.com](https://juansequeda.substack.com/p/gartner-data-and-analytics-london) **Author:** [Juan Sequeda](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23juanSequeda) **Event:** [Gartner Data & Analytics London May 2026](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23conference) **Generated:** 2026-05-20 · Model: Claude Sonnet 4.6 **Companions:** [RDF/Turtle](gartner-da-london-2026-juan-sequeda-claude_sonnet4.ttl) · [HTML Explorer](../web%20pages/gartner-da-london-2026-juan-sequeda-claude_sonnet4.html) --- ## Overview Juan Sequeda's field dispatch from his approximately tenth [Gartner D&A London](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23conference) conference delivers six themed sections of honest, practitioner-grounded takeaways. The article covers the widening AI ROI gap, the mainstreaming of context layers and knowledge engineering, a CDMP vendor strategy framework, governance singularity, practitioner methodologies for avoiding waterfall anti-patterns, and community observations. This knowledge graph encodes the article's entities, relationships, statistics, methodologies, and conceptual insights using a custom ontology atop schema.org, SKOS, PROV, and OWL. --- ## People | Entity | Role | |--------|------| | [Juan Sequeda](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23juanSequeda) | Principal Scientist, ServiceNow; article author; iron-thread methodology creator | | [Mircea Danciulescu](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23mirceaDanciulescu) | WPP Global Data Manager; co-presenter with Juan at Gartner London 2026 | | [Andres Garcia-Rodeja](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23andresGarciaRodeja) | Speaker on context layer; drew packed room with 30 min of basic foundational Q&A | | [Ehtisham Zaidi](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23ehtishamZaidi) | Gartner analyst, CDMP vendor strategy panel | | [Robert Thanaraj](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23robertThanaraj) | Gartner analyst; flagged ISV acquisition risks | | [Roxane Edjlali](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23roxaneEdjlali) | Gartner analyst; highlighted app vendor domain-semantics advantage | | [Daniel Cota](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23danielCota) | Gartner; vendor strategy panel moderator | | [Anurag Raj](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23anuragRaj) | Speaker on governance singularity and policy as code | | [Mark Beyer](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23markBeyer) | Gartner analyst; known to Juan since 2016 | | [Sanjeev Mohan](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23sanjeevMohan) | Independent analyst (SanjMo) | | [Malcolm Hawker](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23malcolmHawker) | Data community leader; Honest No-BS dinner group | | [Ole Olesen-Bagneux](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23oleOlesenBagneux) | Data community figure | | [Samir Sharma](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23samirSharma) | Data community figure | | [Guido De Simoni](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23guidoDeSimoni) | Data community figure | | [Derek Birdsong](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23derekBirdsong) | Data community figure | --- ## Organizations | Entity | Description | |--------|-------------| | [Gartner](https://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fdbpedia.org%2Fresource%2FGartner) | Global technology research and advisory; conference organizer | | [ServiceNow](https://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fdbpedia.org%2Fresource%2FServiceNow) | Enterprise software; Juan Sequeda's employer | | [WPP](https://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fdbpedia.org%2Fresource%2FWPP_plc) | Global marketing services group; Mircea Danciulescu's employer | | [Capsenta](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23capsenta) | Juan Sequeda's previous startup; 2016 Gartner Cool Vendor in Data Integration | | [Substack](https://linkeddata.uriburner.com/describe/?url=http%3A%2F%2Fdbpedia.org%2Fresource%2FSubstack) | Newsletter platform hosting Juan's publication | --- ## Takeaway 1: Basics and Foundations **[AI ROI Crisis](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23aiRoiCrisis)** — 4 out of 5 AI initiatives are failing to achieve ROI in 2025. Root causes are foundational: data quality, architecture gaps, governance immaturity, and the disconnect between CEO expectations and CDO capabilities. **[CIO-CDO Alignment Gap](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23ciocdoAlignmentGap)** — CEOs are spreading AI leadership bets across CISO, CIO, and CDO. CDOs rank 3rd in CEO-perceived AI savviness. CDOs not working closely with their CIO are at a competitive disadvantage. **[Perennial Year of the Foundation](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23foundationYearIrony)** — Gartner again calls 2026 "the year of the foundation." Juan's observation: when has it NOT been the year of the foundation? It is always the year of the foundation — a running industry joke worth examining seriously. ### Statistics | Metric | Value | |--------|-------| | [AI ROI Rate 2025](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23stat1) | Only 1 in 5 AI initiatives achieving ROI | | [AI Agent Cost Concern Gap](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23stat2) | 6/10 IT leaders concerned; only 1/10 D&A leaders | | [AI FinOps Adoption](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23stat3) | 47% of European orgs have financial guardrails | | [Data Engineering Effectiveness](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23stat4) | Only 26% rate data engineering as highly effective | | [Governance Readiness](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23stat5) | Only 13% say D&A governance can fully lead AI governance; 90% say architecture needs overhaul | ### AI Archetypes | Archetype | Description | |-----------|-------------| | [AI-Cautious](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23archCautious) | Gartner-defined; Juan's critique: caution is a posture, not a strategy | | [AI-Opportunistic](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23archOpportunistic) | AI pursued selectively where opportunities are clear | | [AI-First](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23archFirst) | AI embedded as core operating principle across all functions | --- ## Takeaway 2: Context, Context, Context **[What Is Context?](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23contextDefinition)** — Context = Semantics + Operational State + Provenance. Juan extends this to also include Users, Access, and Assets. **[Knowledge Engineering Goes Mainstream](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23knowledgeEngineeringRise)** — Gartner 2026: "Semantics, knowledge graphs and ontologies are no longer niche technologies but essential components of AI-ready data infrastructure." Live poll: 60% starting with semantic layers for BI; 40% going straight to ontologies and knowledge graphs. **[You Cannot Buy Context](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23cantBuyContext)** — Context already exists in every organization. It is fragmented and poorly managed, not absent. No vendor can sell it to you. **[Deterministic Semantics for Trust](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23deterministicSemantics)** — When reliability and trust are required, use semantics-powered deterministic reasoning, not probabilistic LLM inference alone. ### Context Components | Component | Description | |-----------|-------------| | [Semantics](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23contextSemantics) | Ontologies, business glossaries, metrics definitions | | [Operational State](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23contextOperationalState) | Entities, activities, real-time business conditions | | [Provenance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23contextProvenance) | Decision traces, data lineage, process history | --- ## Takeaway 3: Vendor Strategy **[Strategic vs Critical Vendor Tiers](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23vendorTierInsight)** — A vendor can be Strategic (aligned with future direction) without being Critical (mission-critical today), and vice versa. These are independent dimensions. **[Metadata Ownership as the Next Frontier](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23metadataOwnershipInsight)** — The next competitive frontier is not who owns your data, but who owns your metadata. App vendors already hold their domain's semantics; hyperscalers and ISVs are trying to infer what app vendors already know. ### CDMP Strategy Options | Option | Description | |--------|-------------| | [CSP-Centric CDMP](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23cspOption) | Best for time-to-market within one cloud; multi-cloud still needs ISVs | | [ISV-Centric CDMP](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23isvOption) | Best-of-breed depth; monitor acquisition instability risk | | [App-Centric CDMP](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23appOption) | App vendors hold domain semantics best; multi-app challenge | --- ## Takeaway 4: Governance is King and Queen **[Governance Singularity](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23governanceSingularity)** — AI is forcing convergence of data, AI, analytics, cybersecurity, MDM, decision, corporate risk, and IT governance into a single blurring scope. Coined by [Anurag Raj](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23anuragRaj). **[Operational Governance Already Exists](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23operationalGovernanceInsight)** — Operations teams perform governance daily — managing data entry rules, maintaining source correctness, notifying teams of changes — but are not calling it governance. The D&A world has ignored this upstream activity. **[Data Contracts as Bridge](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23dataContractBridge)** — Data contracts bridge upstream operational governance and downstream analytical governance, carrying schema, lineage, SLAs, and quality metrics. **[Policy as Code](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23policyAsCodeInsight)** — Six policy types (definitions/models, quality, security, privacy, ethics, lifecycle) translated into enforceable code. Open question: who is the enforcement engine at scale? ### Governance Domains (Singularity Scope) [Data Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govData) · [AI Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govAI) · [Analytics Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govAnalytics) · [Cybersecurity Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govCyber) · [MDM Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govMDM) · [Decision Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govDecision) · [Corporate Risk Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govCorporate) · [IT Governance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23govIT) --- ## Takeaway 5: Methodologies **[Iron-Thread / Pay-As-You-Go Methodology](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23ironThreadMethod)** — Start from a business outcome, identify the minimum foundation needed, execute end-to-end delivering value while building the foundation simultaneously. Repeat per use case. Avoids big-bang waterfall foundation projects. Published 2019 for knowledge graph design by Juan Sequeda. **[Day in the Life of a Piece of Data](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23dayInLifeMethod)** — Follow one specific piece of data through its complete journey: operational origin → system flows → governance checkpoints → analytics use → decisions → feedback to operations. Forces cross-team dialogue; makes abstract governance concrete. **[Foundation-First Waterfall Anti-Pattern](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23waterfallAntiPattern)** — Big foundation project → next layer → next layer → nothing delivered → everyone loses patience. The industry desperately needs use-case-by-use-case methodologies to counter this cycle. --- ## Takeaway 6: Awesome Community Juan's approximately tenth Gartner D&A conference, with connections going back to the 2016 Capsenta Cool Vendor award. The "Honest No-BS dinner group" included: [Mark Beyer](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23markBeyer), [Ehtisham Zaidi](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23ehtishamZaidi), [Sanjeev Mohan](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23sanjeevMohan), [Ole Olesen-Bagneux](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23oleOlesenBagneux), [Samir Sharma](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23samirSharma), [Malcolm Hawker](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23malcolmHawker), [Guido De Simoni](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23guidoDeSimoni), [Derek Birdsong](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23derekBirdsong). --- ## FAQ **Q: Why are 4 out of 5 AI initiatives failing to achieve ROI?** Data availability and quality remain the #1 barrier. Only 26% of D&A/AI leaders rate their data engineering as highly effective, only 13% say their governance can fully lead AI governance, and 90% say their architecture needs an overhaul. Organizations are deploying AI without the underlying data readiness. **Q: What should a context layer actually include?** Semantics (ontologies, business glossaries, metrics definitions), Operational State (entities, activities, real-time business conditions), and Provenance (decision traces, data lineage, process history). Juan extends this to also include Users, Access, and Assets. Data catalogs rebranded as context platforms may not cover everything needed for AI-first use cases. **Q: What is the governance singularity?** A term from Anurag Raj's session: AI is forcing the convergence of data governance, AI governance, analytics governance, cybersecurity, MDM, decision governance, corporate risk governance, and IT governance into a single blurring scope. These boundaries are dissolving whether organizations plan for it or not. **Q: What is the iron-thread / pay-as-you-go methodology?** Start from a business outcome, identify the minimum foundation needed, execute end-to-end delivering both value and foundation simultaneously. Repeat per use case. The opposite of waterfall-first-then-value. Originally published 2019 for knowledge graph design. **Q: When should I choose CSP-centric vs ISV-centric vs App-centric CDMP?** CSP-centric: already on one cloud, time-to-market is priority. ISV-centric: need best-of-breed depth or cloud neutrality; monitor acquisition risk. App-centric: use cases fit within an application's footprint — app vendors already hold their domain's semantics better than anyone else. **Q: How should CDOs and CIOs work together on AI?** Urgently and directly. CEO surveys show CDOs rank 3rd in perceived AI savviness (behind CISOs and CIOs). CIO-CDO alignment on AI strategy is a prerequisite, not a nice-to-have. **Q: Should my organization adopt an AI-Cautious posture?** Juan's view: No. Caution is a posture that leaves you behind, not a strategy. Being cautious about individual AI investments is prudent; positioning the entire organization as AI-Cautious is not. --- ## Glossary | Term | Definition | |------|-----------| | [Context Layer](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termContextLayer) | Platform providing semantics, operational state, and provenance to enable reliable AI | | [Converged Data Management Platform (CDMP)](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termCDMP) | Gartner term for unified data integration, quality, governance, and metadata platform | | [Iron-Thread Methodology](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termIronThread) | Pay-as-you-go: deliver value and build foundation simultaneously per use case | | [Governance Singularity](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termGovernanceSingularity) | AI-driven convergence of eight governance domains into a single expanding scope | | [Data Contract](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termDataContract) | Formal agreement between data producers and consumers; bridges operational and analytical governance | | [Policy as Code](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termPolicyAsCode) | Translating governance policies into enforceable executable code | | [Knowledge Engineering](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termKnowledgeEngineering) | Building ontologies, knowledge graphs, and semantic layers encoding business meaning | | [Semantic Layer](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termSemanticLayer) | Business-meaning layer for BI and reporting; recommended starting point from BI backgrounds | | [AI FinOps](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termAIFinOps) | Financial governance and cost guardrails for AI; 47% European adoption | | [Provenance](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termProvenance) | Tracking of data origins, processes, decisions, actions, and outcomes | | [AI Archetype](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termAIArchetype) | Gartner classification of organizational AI postures: Cautious, Opportunistic, First | | [Knowledge Graph](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23termKnowledgeGraph) | Network of entities, types, and relationships grounded in ontologies; essential for AI-ready infrastructure | --- ## How To: Build AI-Ready Data Foundations **Step 1 — Start with a business outcome, not a technology** Identify the specific business outcome you want to achieve and work backwards to the minimum data, context, and governance foundation needed. Resist defining the full foundation before the first use case. This is the core of the [iron-thread methodology](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23ironThreadMethod). **Step 2 — Audit the context that already exists in your organization** Context cannot be purchased. Inventory your existing ontologies, business glossaries, metrics definitions, operational rules, policies, and tacit institutional knowledge before evaluating any context platform product. **Step 3 — Choose your CDMP strategy: CSP, ISV, or App-centric** Assess where your data gravity lives, your cloud commitment, and the tier of each vendor: Strategic (future direction) vs Critical (mission-critical today). These are independent dimensions. **Step 4 — Bridge operational and analytical governance with data contracts** Map key data flows between operational systems and downstream analytics. Define data contracts specifying schema, lineage, SLAs, and quality metrics. Bring operational and analytics teams into the same conversation. **Step 5 — Run the "day in the life of a piece of data" exercise** Select one important piece of data. Follow it from its operational origin through every system, governance checkpoint, analytics process, and decision it influences. Map the full round trip back to operations. **Step 6 — Match your knowledge engineering starting point to your background** Coming from BI and star schemas? Start with semantic layers. Coming from operational data integration (Customer 360, etc.)? Go straight to ontologies and knowledge graphs. **Step 7 — Use deterministic semantics for high-reliability AI use cases** When trust and auditability matter, use semantics-powered deterministic reasoning grounded in ontologies and knowledge graphs — not probabilistic LLM inference alone. --- ## Knowledge Graph Provenance This document was generated from [RDF/Turtle](gartner-da-london-2026-juan-sequeda-claude_sonnet4.ttl) using the [kg-generator skill](https://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Fjuansequeda.substack.com%2Fp%2Fgartner-data-and-analytics-london%23kgGeneratorSkill) (OpenLink Software) with model Claude Sonnet 4.6 (Cowork) on 2026-05-20. Entity IRIs resolve via [URIBurner](https://linkeddata.uriburner.com/) — click any linked entity above to explore its knowledge graph context.