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
Robots that work alongside humans or autonomously to perform assembly, manipulation, and pick-and-place operations. Includes force-sensitive cobots and high-speed delta robots; distinct from welding and finishing automation which handles specialised manufacturing processes. Scope covers AI/ML-enabled perception, grasp planning, and adaptive manipulation; pre-programmed fixed-trajectory industrial robots are out of scope.
Collaborative and articulated robot assembly is now firmly in the good-practice tier, with production deployments scaling across electronics, automotive, food and beverage, and FMCG sectors. The category has transitioned from "does it work" to "how fast can we deploy it at scale." Force-sensitive cobots with AI-driven perception and adaptive manipulation are becoming the standard for assembly operations, particularly in mid-range payload (3-20 kg) scenarios where flexibility and rapid reconfiguration matter. Assembly remains the largest cobot application, with 23-25% of global market share. The economic case is now empirically validated: 18-24 month payback periods, 30-40% changeover reduction, and 70% deployment-time savings versus traditional robots justify adoption even at mid-size manufacturers. However, the adoption curve is stratifying. Electronics and food production are moving fast (AGIBOT G2 at 310 units/hour, 77% changeover reduction in FMCG). Traditional automotive and discrete manufacturing are scaling steadily (IFR: 64,500+ cobots in 2024, 13% YoY growth, 12% of new industrial robot installations). The constraint is no longer technology or economics—it's organizational readiness. Reprogramming overhead, safety certification complexity, workforce skills gaps, and integration costs (often 2-5x hardware price) remain binding adoption barriers, particularly for high-variability, low-volume production. AI advancement is beginning to shift this dynamic: foundation-model-driven imitation learning (UR AI Trainer, announced May 2026) now enables learning from human demonstration on production cobots, potentially reducing reprogramming burden and lowering the skill floor for deployment.
The vendor ecosystem is deep and production-hardened. Universal Robots, KUKA, ABB, Techman, Doosan, FANUC, and newer entrants (AGIBOT, Kassow) now compete across 3-30 kg payloads with UR20, UR30, KUKA LBR iiwa, ABB PoWa, and FANUC CRX-3iA anchoring the mid-range assembly segment. Over 420,000 collaborative robot units are deployed globally across 50,000+ production sites (2025-2026), with electronics and automotive accounting for the majority. Deployment scale accelerated in 2024-2026: IFR reported 64,500+ cobots sold in 2024 with 12% market share and 13% year-on-year growth; cobot share of new industrial robot installations has grown from 2.8% (2017) to 12% (2024), demonstrating sustained market acceleration. Recent production evidence shows capability expansion: AGIBOT G2 deployed at Longcheer Technology for tablet assembly achieves 310 units/hour throughput, 99%+ success rate, and 36-hour integration time—demonstrating embodied AI systems operating reliably in high-volume production. FMCG deployments validate changeover speed: Berlin confectionery plant reduced chocolate line changeover from 47 to 11 minutes (77% reduction) with 6 cobots, though revealed safety training gaps as a critical implementation barrier. Economic validation continues: market research confirms 18-24 month payback periods, 30-40% changeover reduction, 70% deployment time reduction, and 80% cost reduction versus traditional robots. Labor shortage pressures sustain adoption drivers (1.9M US jobs projected unfilled by 2033, 60% of German cobot deployments target roles vacant 6+ months), and 69% of manufacturers are investing in robotics per CADDi 2026 survey. AI advancement is shifting the deployment model: Universal Robots and Scale AI launched UR AI Trainer (May 2026), enabling imitation learning for contact-rich assembly tasks (screwing, pressing, inserting) via leader-follower demonstration on production cobots. Foundational technology maturity signals strong: TU Delft research demonstrates €10 3D-printed calibration tools achieve 0.2mm positioning accuracy (50-fold improvement), democratizing precision assembly for SMEs previously limited to €50k-500k measurement infrastructure. Force-torque control research shows 60% of patents filed 2022-2026, indicating ISO/TS 15066 compliance infrastructure maturing across vendors. However, systemic barriers persist despite economic validation. PatSnap landscape analysis ranks reprogramming overhead, safety compliance, dynamic task allocation, economic justification, and workforce skills as the five principal constraints to scaling across high-variability manufacturing. Safety governance remains underresourced: plants without documented risk assessments report cobot incident rates 4.7x higher than structured facilities (78% reduction with formal protocols), yet safety retrofitting into completed cell designs costs 2-5x the design-phase integration cost. Total cost-of-ownership studies estimate $200k-300k per cell when integration, tooling, and training are included—a significant adoption friction point for organizations without existing robotics infrastructure. Geographic bifurcation accelerates: Asia-Pacific captures 70% of incremental deployment growth while EMEA and Americas lag due to weak automotive cycles and higher labor costs dampening ROI urgency. Market forecasts project cobot market growing from $2.8B (2026) to $13.3B by 2034 (CAGR 21%), driven primarily by SME adoption patterns and RaaS models expanding from $1.7B to $5B+ by 2036. Category remains materially constrained by 3-25 kg payload ceiling and ~250 mm/s speeds, making cobots unsuitable for high-throughput heavy-load manufacturing—a scope boundary clarified by successful high-speed, high-payload competitors emerging in humanoid assembly (BMW 30,000+ units) and specialized high-performance platforms (Kassow, ABB PoWa). The practical question is no longer "does it work?" but rather "how do we scale organizational readiness faster than deployment capability advances."
— IFR official 2024 data: 64,500+ cobots sold with 12% market share and 13% YoY growth; cobot share of new installations grew from 2.8% (2017) to 12% (2024).
— Industry analysis: cobot market $2.8B (2026) to $13.3B by 2034 (CAGR 21%); RaaS segment expanding; CADDi survey: 69% of manufacturers investing in robots to address 1.9M US job vacancies by 2033.
— Market research: 18-24 month payback periods, 30-40% changeover reduction, 70% deployment time reduction vs traditional robots; labor constraints structural (60% of German cobot deployments target roles vacant 6+ months).
— FMCG deployment case: 6 cobots reduce chocolate line changeover from 47 to 11 minutes (77% improvement); safety protocols reduce cobot incidents by 78%, revealing real-world deployment barriers and safety-training criticality.
— BMW deploys humanoid robots (Hexagon Robotics AEON, Figure 02) for assembly with 30,000+ cars manufactured; marks production transition from collaborative industrial cobots to embodied humanoid assembly systems.
— UR AI Trainer launched at GTC 2026 enables imitation learning for contact-rich assembly tasks (screwing, pressing, inserting) on production cobots; foundation-model-driven manipulation with 100K+ cobot installed base.
— TU Delft peer-reviewed research: 3D-printed €10 calibration tool reduces cobot positioning error from 10mm to 0.2mm (50-fold improvement), democratizing precision assembly for SMEs without €50k-500k measurement infrastructure.
— Hannover Messe 2026: AGIBOT G2 achieves 310 units/hr in electronics assembly (tablet production) with 99%+ success rate; ABB PoWa, FANUC CRX-3iA, Kassow platforms expand collaborative assembly into high-speed production.