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 continuously monitors sources for new information on defined topics and alerts users to significant developments. Includes automated literature watch and competitive signal monitoring; distinct from deep research which conducts one-off investigations rather than ongoing surveillance.
Continuous research monitoring and alerting has transitioned from second-wave pilots into standardised operational practice across pharma, healthcare, and enterprise contexts. The ecosystem shows clear signs of maturation: major platforms (DistillerSR, Sorcero, Nested Knowledge, MadeAi) have moved past experimental deployments to governance-focused standardisation, regulatory authorities (FDA) have operationalised automated signal detection, and case studies document substantial productivity gains (50-95% reduction in manual labour, 98% faster screening cycles). Yet the transition remains incomplete and uneven. Alert fatigue and threshold-tuning overhead continue to dominate deployment friction; most organisations still lack the operational discipline to sustain continuous monitoring at scale. The technology is proven. The limiting factor is not capability but organisational willingness to invest in the governance, training, and workflow redesign required to make alerts valuable rather than noisy.
Ecosystem maturity is advancing rapidly across pharma, regulatory affairs, medical device, and financial services with clear market scaling. Pharmavigilance deployments define the maturity curve: DistillerSR operates at significant scale with 250+ customers (80% of top pharma and medical device companies) delivering 70% screening time reductions. Sorcero's $42.5M Series B, one-third of top-30 global pharma customer base, and Medical Affairs suite expansion confirm sustained enterprise momentum. Market analysis shows 20% adoption increase in automated signal detection systems in 2025, with pharmacovigilance market scaling from $5.3B (2025) to $12.8B (2034) at 11.2% CAGR, backed by $600M+ annual R&D investments. Regulatory intelligence platforms mature: Flinn.ai (100+ MedTech manufacturers) continuously monitors 120+ jurisdictions with AI-driven impact assessment; Qoniq case study documents EU IVDR-compliant automation reducing weekly manual burden from 25-30 hours to automated cadence. Real-world pharmacovigilance deployments show proactive signal detection: STAR Systems AINE ingests MedWatch, AERS, and literature feeds to flag safety signals months earlier than manual processes, with CFR 21 Part 11 compliant audit trails for regulatory inspection readiness. CoVigilAI covers 154+ countries; systems screen 2.5M+ articles annually with 99% adverse event capture. FDA's Elsa tool and MOSAIC-NLP programme operational. Healthcare validation: Houston Methodist's 8-hospital rollout achieved 95% device use and 4-hour shift time savings; Cleveland Clinic's COSMOS study optimised threshold tuning to 1 alarm per 8 hours. Academic deployments expanding — Oncoscope-AI automates oncology monitoring of 3,898 studies with 98.6% faster daily cycles. Medical writing consolidating around continuous platforms (CiteMed, BiomchBERT).
Governance and regulatory validation now define maturity inflection. ISPOR 2026's "Beyond the Bots" panel (DistillerSR, Nested Knowledge, MadeAi) signals transition to second-phase standardisation: reproducibility, auditability, version control, and governance frameworks. European Medicines Agency's 2025 AI Observatory Report recognises continuous monitoring for signal detection, social media surveillance, and automated ICSR processing as mainstream practice, identifying priorities: explainability, model validation, and governance infrastructure. UK MHRA's June 2026 regulatory sandbox tests AI for continuous safety assessment and risk prediction, signalling government-level validation of continuous monitoring at governance tier. Professional maturity codified: Whitehall Training's certification course (97% adoption rate, 536 reviews) covers automated case triage and signal detection, indicating practitioner-ready tooling and standardised operational patterns.
Domain expansion accelerates: venture capital continuous startup scouting (Lyzr AI) achieves 70% faster identification and 40% deal flow improvement; financial services (70% in POC/active deployment per Global Relay 2026) deploy continuous e-comms/voice surveillance with multi-agent architectures; intelligence applications show 5-layer OSINT monitoring systems detecting geopolitical weak signals. Commercial continuous monitoring platforms now standard: Fullintel serves Top 100 pharma with 24/7 analyst review across 300,000+ sources and 50% cost reduction; market-wide adoption signals 40% of insights leaders continuously review CI using monitoring tools.
Adoption remains unevenly distributed and alert design barriers persist. Alert fatigue and threshold-tuning overhead dominate friction outside pharma/healthcare/regulatory. A 2025 telemonitoring study showed 5-10 mmHg adjustments halve manual processing; simultaneously, 73% of organisations experience outages from ignored alerts, 59% report excessive volume, and 40% of alerts are never investigated. Alert system design analysis reveals five systemic failures: diagnostic context absent, severity tiering missing, ownership undefined, response protocols absent, static thresholds. The binding constraint remains operational: technology capability is proven at scale, but sustained governance discipline, alert calibration, and human-in-the-loop oversight are absent in most organisations outside regulated pharma and healthcare. Platforms expand into new domains (buyer-intent, competitive signals, congress monitoring, venture monitoring) but each reveals the same pattern — threshold tuning and signal-to-noise filtering demand organisational discipline most teams lack.
— Commercial continuous monitoring platform serving Top 100 global pharma companies; delivers real-time regulatory alerts, adverse event escalation, and 24/7 analyst review across 300,000+ sources with 50% cost reduction.
— Professional certification course (97% adoption rate, 536 reviews) covering AI/ML applications in continuous drug safety signal detection, false positive reduction, and automated case triage; indicates codified practitioner maturity.
— Market report documents 20% increase in automated safety signal detection system adoption in 2025, with $5.3B market growing to $12.8B by 2034 (11.2% CAGR) and $600M+ annual R&D investments in predictive analytics.
— Regulatory leading-edge signal: UK MHRA launches sandbox to test AI for continuous safety monitoring, risk prediction, and effect detection; government-funded programme validating AI-driven continuous safety assessment at governance level.
— European Medicines Agency recognition of continuous AI monitoring for signal detection, social media surveillance, and automated ICSR processing as mainstream practice; identifies regulatory priorities including explainability, validation, and governance frameworks.
— Real deployment of AINE continuous signal detection system ingesting MedWatch reports, AERS, and literature feeds with regulatory-compliant audit trails; flags safety signals earlier than manual review, detecting patterns months before manual detection.
— Lyzr AI Startup Scouting Agent continuously monitors funding platforms and startup ecosystems with automated alerts, achieving 70% faster identification, 50% reduction in manual research, and 40% increase in deal flow quality through real-time evaluation.
— Documents evolution from batch quarterly cycles to continuous real-time signal detection in pharmacovigilance. Four pipeline responsibilities: multi-source data integration, statistical signal detection, clinical triage, and audit trail support. Case example: 8-month confirmation lag reduced via continuous pipeline architecture.