3 Data points that are resetting expectations for AI, treasury, and payments
What's happening
Three new studies are shifting expectations in AI, treasury, and payments. UC Berkeley's Center for Responsible, Decentralized Intelligence benchmark found AI agents complete only about 1 in 4 real-world work tasks (top performer OpenAI's Codex finished about 26.2%). Tearsheet summarizes these three data points as evidence that AI still falls short while payments are becoming more strategic and customer expectations change. In the UK, the FCA's Mills Review says AI money managers could affect more than 10 million adults as one in five are open to AI-managed finances, and regulators warn autonomous AI risks may even void cyber insurance in 2026 as firms deploy AI across investment management.
Why it's trending
These findings converge now because new benchmarks, regulator reviews, and industry risk notices landed in quick succession, reframing realism about AI capability and its commercial and regulatory consequences.
SignalHolding at its usual pace, confirmed across 1 independent source type.
Story volume
Stories per dayAngles you could write
If AI agents only finish one in four real tasks, maybe 'AI will replace your job' was never the real threat, the real risk is misallocated trust and capital.
+2 more angles for this topic with an account — all it takes is your email.
Original sources4
- 3 Data points that are resetting expectations for AI, treasury, and payments
Three recent studies reveal meaningful shifts across financial services: where AI still falls short, why payments are becoming strategic, and how customer expectations are being reset. Here’s the narrative behind those numbers. 26.2% of AI agents completed just one in four real-world work assignments The number: OpenAI’s Codex, the top-performing system in UC Berkeley’s Agents’ &
TearsheetJul 28 - Why autonomous AI could void your cyber insurance in 2026
Adoption of AI across the investment management industry has never been higher, with firms deploying the technology to summarise research, monitor threats, automate workflows and bolster cyber defences. Yet this rapid uptake is generating fresh risks, from autonomous systems drifting beyond their intended boundaries to AI-generated errors, hallucinations and generative AI-powered attacks, alongsid
FinTech Global RegTechJul 28 - AI money managers loom for 10 million UK adults
More than 10 million UK adults could soon see their finances managed by an AI agent, according to a landmark review published this week by the Financial Conduct Authority (FCA). The Mills Review, which examines the role of AI in retail financial services, found that one in five adults are already open to letting AI […] The post AI money managers loom for 10 million UK adults appeared first o
FinTech GlobalJul 28 - AI Agents Fail 3 Out of 4 Real Job Tasks
Artificial intelligence (AI) agents can complete only about 1 in 4 real-world work tasks, according to a new benchmark from UC Berkeley’s Center for Responsible, Decentralized Intelligence that pitted leading AI systems against actual assignments from working professionals. The findings test a bold prediction: that AI agents will replace human workers across most knowledge jobs […] The post
PYMNTSJul 24
More rising in Fintech
- Visa Plans 2,600 Job Cuts With Tech and Product Teams Most AffectedSurging4.7Surging4.7 momentum
- Circle Snaps Up IBM's Blockchain Patents In Move That Could Reshape Digital PaymentsClimbing3.3Climbing3.3 momentum
- X Money Launches with Ambitious Features—and Lingering QuestionsClimbing3.3Climbing3.3 momentum
- FDATA calls for a principles-led framework for the next generation of Open FinanceSteady0.9Steady0.9 momentum
- Thredd’s McCarthy Says Payments Will Govern the Agentic EnterpriseSteady1.0Steady1.0 momentum
- Keenai appoints Singapore head for broader push into South East AsiaSteady0.9Steady0.9 momentum