BankingNewsAI Daily Brief ·
Incore Bank deploys agentic AI with Kyndryl and Google Cloud for customer onboarding.
Banking AI
Financial institutions & fintech technology
Incore Bank is deploying agentic AI for customer onboarding with Kyndryl and Google Cloud
Kyndryl and Google Cloud announced an agentic-AI customer-onboarding deployment at Incore Bank. The implementation makes onboarding—not internal experimentation—the immediate banking use case, and ties the service to a major cloud vendor US institutions already buy from.
→ Action
Prioritize onboarding as an agentic-AI candidate only with explicit human escalation, identity-verification controls, evidentiary logging, and measurable cycle-time and abandonment targets.
Read article → from Kyndryl
General AI
Large language models & AI infrastructure
Z.ai released an MIT-licensed, 1M-context model priced to undercut frontier-model APIs
Z.ai launched GLM-5.3-Flash, an open-weight multimodal mixture-of-experts model with 320B total parameters, 18B active parameters, and a 1M-token context window. Artificial Analysis reported a 57 Intelligence Index score and roughly $0.09 per task, while Z.ai priced API use at $0.15 per million input tokens and $0.50 per million output tokens. The model is positioned for coding and long-horizon agent workflows, although its reported factual-knowledge accuracy is weaker than leading proprietary models.
→ Action
Benchmark GLM-5.3-Flash against your current approved models for internal coding, document-processing, and agentic workflows; its open license and materially lower API pricing can change the build-versus-buy economics.
Read article → from Smol
Anthropic faces a new large-scale music-copyright suit over Claude training data
Sony Music Publishing, Warner Chappell, and other labels sued Anthropic in federal court, alleging use of thousands of songs to train Claude and seeking damages for copyrighted works and allegedly stripped rights-management information. The suit expands the legal exposure surrounding training-data provenance at a major enterprise AI vendor.
→ Action
Revisit vendor contracts for training-data representations, indemnities, model-use restrictions, and exit rights; do not assume model-provider IP risk remains abstract.
Read article → from The Verge
a16z has raised a $1.1 billion fund focused on AI hardware and infrastructure
Andreessen Horowitz raised $1.1 billion for its Machine Age Fund to invest in AI computing infrastructure, including the hardware layer beneath model deployment. The move signals that capital is shifting beyond model makers toward the supply chain, systems, and capacity constraints that determine enterprise AI economics.
→ Action
Stress-test AI roadmaps against compute availability, hosting concentration, and inference-cost assumptions rather than budgeting only for software licenses and model APIs.
Read article → from PYMNTS