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 accelerates period-end financial close processes and automates revenue recognition across contract types. Includes close task automation and ASC 606 compliance support; distinct from financial reporting which generates outputs from close data rather than managing the close process itself.
AI-driven financial close automation remains a leading-edge practice caught between proven capability and structural execution barriers. The technology works in isolation—named deployments (ChatFin on SAP B1: 72.5 hours → 2.8 hours; Hyundai Mobis via Consark: 40% cycle reduction; $1.2B specialty chemicals: 11→3 days, $850K savings) demonstrate substantial time and cost compression with measurable outcomes. Yet adoption is stalled: only 2% of organizations fully automate close despite 93% planning AI investment within 18 months. The execution gap is not technical but organizational. Recent research exposes the core tension: AccountingBench tested frontier AI models on 12-month close simulation with real transaction data—models achieved 95% accuracy initially but diverged 15%+ by year-end due to missing state management, validation controls, and workflow continuity. Finance ROI measurement remains broken: 87% of CFOs view AI as critical to strategy yet only 21% report tangible value; 95% of AI pilots deliver zero P&L impact, with 37% of claimed time savings consumed by output rework. MIT Sloan research documents adoption stall explicitly: proofs of concept never leave sandboxes, promising pilots sit unused, expected closing-cycle speedup hasn't materialized despite four years of vendor maturity. Revenue recognition adds deterministic precision constraints: ASC 606 requires 100% accuracy for contract modifications and billing reconciliation—a ceiling current AI and ERP systems cannot reach. Data quality emerges as foundational barrier—85% of UK finance teams still use Excel, 67% cite data reliability as top priority, revealing most firms cannot adopt AI until spreadsheet dependencies are replaced. The practice is leading-edge because solutions exist and some organizations deploy successfully, but scaled adoption requires resolving data quality foundations, auditability, ROI measurement, and organizational change management that remain unsolved.
Vendor platform momentum accelerates with production adoption signals cementing in Q2 2026. BlackLine demonstrates category-leader scale: Q1 2026 earnings report Studio360 platform penetration at 13% of eligible ARR (up from 11% prior quarter), remaining performance obligations +18% YoY, 94% of new bookings landing on platform; two-thirds of customers actively deploying Verity AI agents with 183% QoQ usage growth. Verity agents deliver concrete production outcomes: reconciliation time reduced 90%, match rates 80-90%, collections straight-through processing ~90%. Named customer deployments validate measurable outcomes: $1.2B specialty chemicals manufacturer deployed multi-agent close platform on SAP/Oracle/NetSuite, automating 9,400 monthly journal entries, achieving zero SOX 404 findings, and compressing close from 11 to 3 days with $850K annualized savings; $95M professional services firm on SAP Business One reduced active close time from 72.5 hours to 2.8 hours (96% compression) by close cycle 3; multiple financial services firms achieve 40–55% close cycle compression through well-configured AI agents (baseline 6.1 days → 3.4 days, 44% reduction). Big 4 production readiness confirmed: KPMG released GA financial close AI assistant with Workday integration; KPMG survey of 1,013 finance leaders shows 93% of US companies planning AI deployment in finance within 18 months, with 50% planning multi-agent orchestration and agentic deployments showing 32-point outcome improvement advantage (76% vs 44% non-agentic). Revenue recognition automation accelerates: Zuora, RightRev, Workday, SAP, and Oracle all deliver GA ASC 606 solutions targeting consumption-based pricing complexity. These signals confirm leading-edge capability maturity: named deployments, measurable ROI, and ecosystem production readiness among category leaders.
Yet market adoption stalls despite CFO investment intent and mature vendor capability. SSON research (March 2026) finds 97% still manual close, only 2% fully automated; 84% of journal entries remain manual, 86% reconcile in spreadsheets. Enterprise AI agent production-readiness remains critical constraint: consulting firm analysis finds only 3 of 23 typical PoCs reach production, only 11% of organizations actively using agents, and only 25% of AI initiatives delivering expected ROI—three-root-cause failure pattern (no hard business case, poor data foundation, wrong architecture). Deployment intent dropped from 42% (Q3 2025) to 26% (Q4 2025)—signals market pulling back from early enthusiasm. The critical barrier is not technical capability but organizational execution: 87% of CFOs prioritize AI for finance and 68% increase digital transformation spending, yet Zuora survey reveals only 28% with measurable financial impact, 87% perceiving gaps between promise and reality, and only 43% confident in audit/control compliance. Hidden costs undermine ROI: Workday analysis reveals 37% of claimed time savings lost to output fixing; practitioners report software 30-50% of cost, data readiness $5-15K unplanned, integration +30-50%, and 15-25% productivity drop 3-6 months post-launch. Finance AI ROI measurement itself is broken: only 15% report EBITDA lift despite productivity claims; 77% cannot measure AI ROI; 95% of non-specialist AI pilots deliver zero P&L impact. Root causes are structural: 98% data integration challenges, 92% skills shortages, 82% lack AI ROI measurement capability, 73% of firms lack structured end-to-end deployment despite 98% trying AI on isolated tasks. MIT Sloan peer-reviewed research documents adoption stall explicitly: proofs of concept never leave sandboxes, promising pilots sit unused when close-week pressure hits. Data quality emerges as foundational prerequisite: 85% of finance teams still rely on Excel, 67% prioritize data reliability as top investment, revealing most firms cannot adopt AI until spreadsheet dependencies replaced with integrated tools. Only 26% of organizations have achieved full-scale AI production with measurable returns.
Revenue recognition automation faces persistent operational barriers despite vendor innovation and Big 4 ecosystem investment. Vendor specialization signals market maturity, yet major ERPs still cannot match contracts to billing systems deterministically, track modifications without manual intervention, or reconcile at 100% precision ASC 606 requires. Audit compliance emerges as gate: auditors now ask for immutable records of AI inputs, logic, and reviewer sign-off; DIY AI controls fail due to editable chat history, silent model drift, and lack of change governance. Frontier AI models tested on realistic 12-month close scenarios achieved 95% accuracy initially but 15%+ balance sheet divergence by year-end due to missing state management and validation controls. Critical assessment: 95% of AI rollouts fail when layered on fragmented data and untrusted processes; AI amplifies dysfunction rather than fixing it. Practitioner breakdown shows success requires bridging extraction to GL posting, state management, auditability, and treating close as integrated workflow (20h to 1-2h per client end-to-end). Workday and vendor platforms describe vision of continuous close agents with real-time transaction auditing and autonomous journal posting, but execution remains concentrated among governance-first vendors (BlackLine, FloQast, Trullion) and forward-leaning customers with mature data programs. Until data quality foundations, auditability infrastructure, ROI measurement discipline, and process-first (not tool-first) approaches improve, close automation will remain concentrated among category leaders; broader market stuck in pilot purgatory.
— CFO identified $200k annual tax leakage savings via AI; 80% invoice cost reduction, 99.8% accuracy, <1 min processing; industry benchmark: 40–55% close compression via AI deployment.
— Production reconciliation agents: 95% STP, 25% DSO reduction, April 2026 metrics—708k interactions, 361k signals, 108k errors blocked before posting; embedded AI in ERPs drives 30% close acceleration by 2028.
— V7 Go ASC 606 agent: 90% faster contract analysis (2-3 hours → 5-10 minutes), 99% accuracy, audit-ready documentation with AI citations; directly addresses revenue recognition automation.
— Multi-country NetSuite AI automation (UK, Germany, Netherlands): month-end close 4–5 days → 8 hours, 90% STP, LSTM/XGBoost models with SOC 2/FCA/German-Dutch regulatory compliance.
— MIT Project NANDA: 95% AI pilots zero P&L impact; measurement discipline (not tech failure) drives ROI; payback 3-12+ months depending on complexity and data quality maturity.
— Gartner: journal entry automation <5% adoption (accuracy/audit exposure barriers); adoption concentrated in high-volume AP (37%) vs. judgment work (forecasting 12%); data/governance primary differentiators.
— 311 senior finance professionals: 64% using AI but only 12% operationalized in core processes; barriers: 72% skills gap, 80% governance/trust; shadow AI rising (33% report issue).
— Fusion Computing-led 40-person financial planning firm: month-end reporting 2 days → 4 hours (6x acceleration), recovered 15–20 hrs/week, measured ROI in 90 days via governance-first 3-phase deployment.
2020: Major cloud vendors (Oracle, ServiceNow) release dedicated financial close automation products; Forrester study finds 90% of organizations still struggle with manual, error-prone close processes due to ERP fragmentation and lack of standardized procedures. Early case studies show 20% time savings in specific deployments, but barriers remain high.
2021: BlackLine, FloQast, and Trintech expand deployments across sectors (healthcare, finance, real estate); Nucleus Research validates ROI case for Oracle ERP + close automation (payback <1 year, 90% close-time reduction). Trintech surveys confirm 90% of CFOs report close-process barriers despite tool availability; adoption momentum slower than vendor messaging. ASC 606 revenue recognition automation remains partially addressed across industries.
2022-H1: FloQast launches AI-driven Reconciliation Management product with 31% time-reduction metrics from production customers (Twilio). However, Trintech H1 benchmark survey shows 74% of finance organizations still lack established automation; SEC enforcement remains elevated for revenue recognition violations, indicating compliance gaps persist.
2022-H2: Continued vendor momentum with BlackLine/Kyriba integration case (Culligan Water: 99% match rate), Sage survey shows 40% adoption of new close technology (up from 21% prior year) and 21% of larger companies using AI. Gartner data shows 55% of finance executives targeting touchless close by 2025, but Controllers Council survey confirms persistent gaps: 11% no automation, 32% basic levels, with skills and data readiness as blocking factors. AI implementation risks emerge (Unity Software $110M loss from ML model failures) alongside demonstrated ROI.
2023-H1: Production deployments continue with BlackLine and FloQast customers achieving documented time savings and process improvements. However, adoption barriers persist: Trintech benchmark data from early 2023 shows many finance organizations still lack established automation infrastructure. Critical assessment of month-end close reveals severe workforce stress—FloQast survey reports 99% of accountants experiencing burnout with 81% experiencing personal life disruption during close cycles, highlighting staffing and workload challenges limiting broader implementation momentum.
2023-H2: Vendor acceleration (BlackLine 5-day fast-track, FloQast awards) and concrete customer results (Crusoe 10-day close, C3.ai 4-day quarterly close) demonstrate mature production deployments. However, generative AI enthusiasm collides with caution: F&A talent gap (62% lacking deep technical skills) and AI governance risks (hallucinations, auditability concerns) emerge as adoption blockers. Market bifurcates further—Fortune 500 adoption accelerating while 74% of organizations still lack established infrastructure.
2024-Q1: FloQast launches expanded platform with AI-enhanced reconciliation management, transaction matching, journal entry automation, and consolidation capabilities. Concurrent research identifies data quality as the critical adoption barrier: BlackLine survey shows 37% of CFOs lack confidence in financial data accuracy; CFOs embrace automation conceptually but remain cautious about AI risk without foundational data governance. SMB sector shows continued enthusiasm (Bill.com survey), though maturity remains concentrated among enterprises with stable accounting processes.
2024-Q3: FloQast releases auditable AI agents for close automation (Journal Entry, Data Transformation), continuing platform acceleration. Concurrent critical signals emerge: insightsoftware survey reveals 98% of finance teams face data integration challenges and 92% report skills shortages; Gartner predicts 30% of GenAI projects will be abandoned after PoC by end of 2025; FASB post-implementation review documents persistent ASC 606 complexity. AI enthusiasm yields to skepticism about deployment realism and auditability in financial applications.
2025-Q1: FloQast ships AI agents into production (biopharmaceutical company case: 12-20 hours monthly close-task savings). Market research signals continued growth: revenue recognition software market at $3B growing 15% CAGR to $8B by 2033. However, Gartner's January 2025 analysis documents persistent GenAI deployment challenges—project failure rates high, organizations chasing models vs. solving use cases. Data integration (98% of teams), skills gaps (92%), and data quality (82% cite as blocker) remain unchanged as primary adoption barriers; CFO confidence in financial data remains low. ASC 606 complexity continues: FASB and SEC actively issuing guidance and enforcement. Practice remains leading-edge in capability but shows maturation—early adopters see results, structural barriers limit broader transformation.
2025-Q2: Vendor momentum sustained (BlackLine $172M revenue, 4,451 customers, 105% NRR) but broader market skepticism deepens. California Management Review research: only 11% of companies at GenAI scale maturity; Kyriba survey reveals CFO unease about AI risks (accuracy, privacy, security). Job displacement signals (76,440 AI-driven job losses in 2025) and modest ROI reports (<10% cost reduction, <5% revenue increase across most orgs) temper adoption enthusiasm. Structural blockers unchanged: 98% data integration challenges, 92% skills shortages, 82% cite data management as implementation barrier. Market bifurcates further between leading vendors with proven deployments and majority struggling with ROI and governance.
2025-Q3: Academic research validates AI productivity gains in close (MIT/Stanford study: 7.5 days saved, 55% more clients supported per week) but practitioner assessments remain cautious. FloQast releases AI Agent Builder with expanded capabilities (AI Detections, Testing, Variance Analysis). KPMG report positions Intelligent Close as modernization foundation. Critical signals emerge: Penrose Labs study documents AI accuracy failures in real-world close attempts (cascading errors, accrual-accounting misunderstandings); RightRev analysis highlights consumption-based revenue recognition complexity and 100% precision requirements that current AI struggles to meet. Positive vendor momentum (FloQast product GA) offset by rising concerns about AI implementation realism and auditability in high-stakes financial processes.
2025-Q4: Vendor momentum accelerates with BlackLine expanding agentic AI suite across financial workflows (Documents, Journals, Variance, Intercompany, AR agents) and reporting 45% surge in new customer bookings; adoption signals reveal critical execution barriers. Glenn Hopper year-end analysis shows 59% of finance functions use AI but only 10% at enterprise scale, 71% unable to deploy GenAI in workflows, median ROI stuck at 10% vs. 20% targets, two-thirds in "pilot purgatory." insightsoftware survey documents 58% see AI as essential but only 39% feel confident—critical confidence gap. Auditability concerns intensify: Warren Averett and independent analysis highlight AI-generated estimates lacking documentation, ASC 606 automation opacity, and governance risks limiting production deployment. Trintech research confirms 95% of AI pilots fail to deliver measurable returns. Market bifurcates sharply: category leaders (with ISO 42001 certification, auditability controls) achieving results; majority constrained by data quality, skills gaps, governance maturity, and auditability confidence.
2026-Jan: Vendor capability continues with RecVue and FloQast expanding revenue recognition and close automation platforms; practitioners share deployment strategies at TakeControl APAC conference. However, critical adoption barriers intensify: agentic AI deployment slows (42% Q3 2025 to 26% Q4 2025); 77% of organizations cannot measure AI ROI; 95% of enterprise GenAI projects fail to deliver measurable returns within 6 months. Only 14% of CFOs report meaningful AI ROI despite 66% expecting impact within 2 years. Revenue recognition automation faces specific operational challenges: ASC 606 software cannot resolve contract-system mismatches, modification tracking gaps, or billing-revenue reconciliation issues. Market signal: leading-edge capability persists but execution barriers—data quality, governance, auditability, ROI measurement—constrain adoption to category-leading vendors; broader market remains in pilot purgatory.
2026-Feb: Production deployment evidence validates agentic close automation: Liquid AI and an AI startup deploy FloQast Transform agents for discrete task automation (allocations, accruals, reconciliation) with measurable time savings. BlackLine Q4 2025 results confirm 4,394 customers and sustained investment in agentic agents. However, hidden implementation costs emerge: Workday analysis reveals 37% of AI time-savings lost to output fixing, undermining ROI claims. Broader adoption stalls—76% plan investment but only 6% deployed at scale; SMB adoption particularly challenged (12% using AI, 63% evaluating). Revenue recognition automation faces unresolved operational barriers: ERPs and AI systems cannot yet handle contract-system mismatches, modification tracking, or billing-revenue reconciliation at 100% precision ASC 606 requires. Market bifurcates sharply: category leaders achieving measurable results, majority constrained by data quality, hidden costs, governance gaps, and execution barriers.
2026-Mar: Vendor platform acceleration continues: Consark (Hyundai Mobis 40% cycle reduction), Workday Sana (400+ customers, Lights Out Finance bots GA), Oracle NetSuite (Close Manager, exception detection, narrative generation GA). RightRev positioned as Leader in automated revenue recognition. Critical adoption barriers persist: SSON research reveals 97% still manual close, 2% fully automated, 84% journal entries manual, 86% reconcile in spreadsheets. NetSuite AI Connector accuracy failures documented by independent consultants. Deployment intent stalled (26% vs 42% prior quarter). Hidden rework costs confirmed: 37% of AI time savings consumed by output fixing. Category leaders (vendors with governance, auditable AI) achieving results; broader market in pilot purgatory with 95% of non-specialist pilot programs failing. ASC 606 precision gap remains unresolved across ERPs and AI agents—accuracy risks persist despite vendor claims.
2026-Apr (Early): Regulatory pivot signals CFO tailwind: US Treasury released AI Risk Management Framework (March 2026) explicitly reframing non-adoption as financial risk. Trullion launches agentic revenue recognition with auditable AI agents for ASC 606/IFRS 15 compliance; RightRev, Workday, SAP, and Oracle all have GA revenue recognition solutions targeting AI company consumption-based pricing complexity. AccountingBench research delivers a critical precision warning: frontier AI models (GPT-4, Claude) achieved 95% accuracy in 12-month close simulations initially but diverged 15%+ by year-end due to missing state management and validation controls, underscoring the gap between pilot performance and production reliability. Yet execution barriers deepen: Battery Ventures CFO survey (129 executives) finds only 4% pilot success rate above 50% and 71% cite model inaccuracy; Zuora data shows a persistent AI trust gap with 82% of boards lacking ROI measurement capability for finance AI. Root cause analysis: OneTribe Advisory identifies data model fragmentation as core blocker—95% of gen AI pilots show no measurable P&L impact when revenue/cost definitions lack unified computation paths across systems. Structural pattern holds: category leaders (Consark, FloQast, BlackLine, Trullion) achieving measurable results with auditable, governance-first AI; broader CFO market stuck between strategic intent and execution realism.
2026-May (Early): Production adoption signals mature at category leaders. BlackLine Q1 2026 earnings show two-thirds of customers using Verity AI agents with 183% QoQ usage growth; consumption-based pricing model for agents launching in 2027. KPMG releases GA close AI assistant with Google Gemini Enterprise and Workday integration, signaling Big 4 enterprise-grade readiness. MIT/Stanford research documents 7.5-day close-time reduction and shift of 8.5% of time to analytical work, validating productivity thesis. Revenue recognition automation accelerates: Zuora releases GA ASC 606 software as part of multi-vendor specialization (RightRev, Workday, SAP, Oracle). Yet CFO adoption intent remains disconnected from execution: 87% of CFOs prioritize AI for finance and 68% increase digital transformation spending, yet Zuora's survey reveals only 28% with measurable financial impact, 87% perceiving gaps, 43% confident in audit/control compliance. Adoption intent stalled: deployment dropped 42% (Q3 2025) to 26% (Q4 2025). Structural barriers unchanged—98% data integration challenges, 92% skills gaps, 82% lack ROI measurement. Pattern holds: category leaders achieving results with auditable, governance-first architecture; broader market stuck in pilot purgatory despite CFO investment signals and mature vendor capability.
2026-May (Mid): Category-leader production momentum validated with named deployments and ecosystem production readiness. DreamzTech case study documents $1.2B specialty chemicals manufacturer compressing close 11→3 days with multi-agent AI platform, achieving $850K annualized savings, automating 9,400 monthly journal entries, zero SOX 404 findings. BlackLine Agentic Financial Operations GA launch emphasizes governance-first architecture (glass-box auditable AI, immutable audit trails, 90% reconciliation time reduction, 80-90% match rates); Big 4 ecosystem investment (KPMG pilot, Google Gemini partnership, Workday integration) signals enterprise-grade production readiness. Practitioner case studies from Zenskar confirm AI handling of mechanical close workflows in production: bank reconciliation cut from 4 hours to 45 minutes (81%), revenue variance from 2 hours to 20 minutes (83%), ASC 606 schedule from 45 to 10 minutes per contract — using Claude and ChatGPT with human review of exceptions. Adoption survey (KPMG: 1,013 finance leaders) shows 93% of US companies planning AI finance deployment within 18 months, 50% planning multi-agent orchestration; however agentic AI deployments show 32-point advantage vs. non-agentic (76% vs. 44% outcome improvement). Critical execution barriers persist: practitioners report 73% of firms lack structured end-to-end deployment despite 98% trying AI on isolated tasks, with hidden costs consuming 37% of claimed savings and 95% of AI rollouts failing when layered on fragmented data and untrusted processes. Audit compliance emerges as new gate: auditors require immutable input/logic records, change governance, and reproducibility across close cycles. Market signal: production capability mature, ecosystem investment accelerating, but organizational execution barriers (data quality, hidden costs, audit readiness, process-first thinking) remain constraining broader adoption.
2026-Jun: MIT Sloan peer-reviewed research (300+ CFO multiyear study) formally documented the adoption stall: close-cycle speedup has not materialised, proofs of concept sit unused, and pilots never leave sandboxes — despite four years of vendor maturity — attributing the failure to organizational barriers rather than technology gaps. A concrete counter-example emerged simultaneously: ChatFin deployed on SAP Business One for a $95M professional services firm compressed active close from 72.5 hours to 2.8 hours (96%) by cycle 3, while a UK study (Sixthfin x Odoxa) found 85% of finance teams still use Excel and 67% prioritise data reliability as their top investment — confirming that data quality foundation, not AI capability, remains the decisive adoption prerequisite for the majority of the market.
2026-Jun (Mid-Late): Scan confirms dual pattern: production-scale deployments validate capability maturity while governance and measurement barriers constrain broader adoption. New production signals: 40-person Toronto financial planning firm compressed month-end reporting 2 days → 4 hours (6x reduction) via governance-first 90-day deployment; multi-country NetSuite automation across UK/Germany/Netherlands achieved 4-5 day close → 8 hours with 90% STP and FCA/SOC 2 compliance; reconciliation agents in production delivered 95% STP, 25% DSO reduction, 108k errors blocked before posting; SAP S/4HANA AI close automation demonstrated $200K annual tax leakage savings, 80% invoice cost reduction, and 99.8% accuracy at under 1 minute processing. Platform GA announcements: Oracle Fusion 26B embedded four agentic apps (Ledger, Payables, Payments, Expenses); V7 Go delivered 90% faster ASC 606 contract review (2-3 hours → 5-10 minutes, 99% accuracy, audit-ready documentation). Critical barriers identified by comprehensive research: ChapsVision (90% pilot→production failure, governance/trust primary blocker), Gartner (journal entry automation <5% adoption due to accuracy/audit concerns), insightsoftware survey (64% using AI, only 12% operationalized in core processes), EY (21% of CFOs report AI readiness vs 80% expecting impact). Revenue recognition complexity escalates for AI company pricing models: ASC 606 now governs usage-based APIs, prepaid credits, consumption thresholds—emerging complexity distinct from traditional SaaS. Deloitte Big 4 guidance (June 2026) on revenue recognition for outcome-based agentic AI pricing signals practice maturation into audit/regulatory domain. Market signal: capability ceiling reached (demonstrated via named deployments and platform GA); value realization ceiling remains organizational—execution, governance, ROI measurement, and audit readiness determine adoption. Practice remains leading-edge but promotion risks without resolution of operational, governance, and measurement barriers that pervade 64-88% of target market.