The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.
A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.
Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail
AI that automatically drafts policy updates in response to regulatory changes for human review and approval. Includes tracked-change policy revision and compliance mapping; distinct from impact assessment which analyses but doesn't draft remediation.
A small but growing number of forward-leaning organisations now use AI to draft policy updates in response to regulatory changes -- moving from horizon scanning into actual document generation. Production platforms from Regology, FinregE, Scytale, and others, alongside emerging RegTech startups (AscentAI, 4CRisk.ai), demonstrate real-world deployments with measurable impact. Early adopters report striking efficiency gains: peer-reviewed research documents 96% recall and 3.1x analyst efficiency in 4-month production deployments; multi-site case studies show 98% reduction in data discrepancies and $250M revenue impact across 16 facilities; single-deployment instances dropped policy drafting time from five days to one. Market consolidation (CUBE's acquisition of Acin, Zango AI's funding round for regulation-specific LLMs) signals genuine commercial momentum. Yet adoption remains constrained. Only 60% of the most mature compliance organisations have adopted AI-powered regulatory change monitoring; at lower-maturity levels the figure falls to 48%. Data quality and integration challenges affect the majority of adopters, and false negatives in AI-drafted outputs demand multi-layered human review. Every documented deployment keeps humans in the approval loop -- AI drafts, people decide. For most compliance teams, the eight-week manual cycle per regulatory change is still the reality.
Deployment is accelerating but constrained by governance maturity. Vendor ecosystem has matured from RegTech specialists to mainstream GRC platform coverage: Regology, FinregE (FCA-contracted), Scytale (1000+ deployments, 4.9/5 G2 rating), and ServiceNow RCM join AscentAI and Clausematch in offering integrated regulatory change monitoring with automated policy drafting. Enterprise adoption is real: Ascent and Clausematch deploy for Goldman Sachs, Citi, and JPMorgan, while DSALTA reports production policy generation at 98% time reduction (20-25 policies drafted in 2-3 hours vs. 2-3 weeks manually). Independent third-party reviews document 12 verified Regology deployments across Fortune 500 banks and healthcare systems, and peer-reviewed research confirms 4-month production deployments achieving 96% recall on gap detection. Yet infrastructure gaps remain acute. Only 24% of organisations have AI governance frameworks in place, and agentic compliance systems fail 15-30% of the time without strict constraints. Saifr CEO documents critical risks: cascading hallucinations where a single flaw propagates through downstream systems, memory poisoning from incorrect data retention, and the "confused deputy" problem where AI systems become reliable targets for adversarial exploitation. EU AI Act enforcement, originally due August 2026, slipped to 2027-2028 due to missing harmonized standards and untrained conformity assessors -- a signal that regulatory infrastructure itself lags deployment readiness. The result: 78% of enterprises have AI pilots but fewer than 15% reached production scale. Compliance teams continue to spend 70% of their time on manual regulatory monitoring, processing each change over an eight-week cycle. Vendors claim 50x acceleration and near-instantaneous policy updates, but deployed systems retain humans in the approval loop. For those who solve governance and integration, the payoff is dramatic. For the majority of compliance teams still establishing basic AI controls, the gap between promise and practice remains a structural barrier.
— FinregE's machine-readable rulebook framework enables AI to detect and respond to regulatory changes; FCA production deployment of structured-data approach addresses hallucination risk in policy-generation systems.
— Major compliance publisher Bloomberg's acquisition of regulatory change management platform Regology signals market consolidation; validates automated policy-update workflows as strategic, durable market category.
— Anthropic's GA AI Governance Legal Plugin includes 'policy-monitor' and 'reg-gap-analysis' skills that identify regulatory changes and trigger policy drafting; 'policy-starter' generates first-draft policies from regulatory sources.
— AscentAI survey reveals critical adoption bottleneck: monitoring regulatory updates ranks top, but gap between alert and action (policy updates, system changes) remains unresolved at scale; 80% of compliance teams still use spreadsheets.
— Expert analysis documents regulatory fragmentation blocking automated policy updates: EU AI Act delayed to Dec 2027, accountability frameworks for AI-agent decisions unsettled across jurisdictions, limiting deployment feasibility at scale.
— Deloitte survey of 400+ large enterprises: 67% report multi-jurisdiction tracking requires more engineering than building their AI systems; confirms enterprise adoption of policy monitoring tools and identifies bottleneck in governance infrastructure.
— Galileo Labs playbook documents architectural patterns for decoupled policy engines that absorb regulatory change without code rewrites, eval engineering for audit evidence, and compliance documentation requirements across EU AI Act enforcement timeline.
— Real-world deployment across hundreds of HR teams demonstrates regulatory-change-triggered policy generation: system automatically generates state-specific handbook language when employment law changes, with attorney verification confirming 95%+ accuracy.