Data and analytics in financial services, week ending 16 October 2026

Week ending 16 October 2026 · 15 sources

Banks widen cloud AI deals as model provenance comes into focus

CaixaBank expanded its Google Cloud agreement to speed AI and data integration 1. Napier AI says financial institutions now ask where compliance models were trained and what data they learned from 2.

Which way the market is moving

Banks deepen cloud and AI data platforms

CaixaBank, NatWest and CLS each moved data or AI work onto large technology partners this week 189.

CaixaBank has expanded its strategic agreement with Google Cloud to strengthen its technology and speed the integration of AI, data management and cloud infrastructure across its business 1. CLS, which provides settlement, processing and data services, said with Kyndryl that it had completed an infrastructure migration 9.

NatWest has used Nvidia speech recognition so its new audio-visual spending insights tool can understand customers' accents and pronunciations 8. Both the CaixaBank and NatWest moves pair the institution's own customer data and services with a large technology provider's AI or platform capabilities 18.

What it means: Institutions planning AI on customer data should expect platform choices to set the limits of their data management and AI options for years. Contract terms on data use, portability and exit matter as much as features.

Questions for vendors:

  1. How is our data kept separate from other clients' data when your AI services are trained or tuned?
  2. What are the exit and data portability terms if we leave the platform before the contract ends?
  3. Which of your AI components can run on our own infrastructure, and which only on your cloud?

In the registry: Alteryx One (Alteryx), AI Lineage Assistant (Solidatus), smartTrade Analytics (smartTrade Technologies)

Model provenance and data supply chain risk

Napier AI says compliance buyers now weigh where AI models were trained and who built them, as hackers claimed a breach of a Snowflake-connected AI tool 210.

Napier AI says financial institutions choosing AI for financial crime compliance now weigh where a model was trained, who built it and what data it learned from, not only how well it performs 2. It frames this as a choice between sovereign AI and shared intelligence 2.

BankInfoSecurity reported that hackers have claimed a breach of Simon AI, a tool connected to Snowflake 10. Separately, the UK NCSC published recommended profiles for configuring TLS to protect data, with guidance on managing older versions 11.

What it means: AI tools that connect to a data platform extend the attack surface of that platform. Due diligence should cover model lineage and every third party with access to the warehouse, not just the warehouse vendor.

Questions for vendors:

  1. Where was your model trained, on what data, and can you document that lineage for our regulator?
  2. Which third-party tools hold credentials to our data platform, and how are those credentials scoped and rotated?
  3. Which TLS versions and configurations do your services support, and when will older versions be retired?

In the registry: AI Lineage Assistant (Solidatus), Data Observability (Ataccama), Immuta Govern (Immuta)

Supervisors adjust reporting and data roles

The ECB gave institutions a choice of IAS 1 or IFRS 18 for profit and loss reporting, and ASIC is seeking a new chief data and analytics officer 34.

ECB Banking Supervision said institutions will have the option to report profit and loss under IAS 1 or IFRS 18 3. APRA published its life insurance claims and disputes statistics 12, and the US Federal Reserve Board released results of its 2025 Survey of Consumer Finances 13.

In Australia, ASIC's security and data chiefs are due to leave in December 4. The corporate watchdog is now seeking a CISO, a chief data and analytics officer and a contract CTO 4.

What it means: Reporting teams in the euro area will need systems that can produce profit and loss under either standard during the transition. In Australia, changes in ASIC's data leadership may shift how it asks for and uses regulated entities' data.

Questions for vendors:

  1. Can your reporting platform produce profit and loss under both IAS 1 and IFRS 18 from the same data?
  2. How quickly can you change regulatory report templates when a supervisor updates its requirements?

In the registry: Analytics and Intelligence (Novidea), AI Lineage Assistant (Solidatus), Exasol Personal (Exasol)

External and alternative data in credit and markets

Credit, payments and markets firms are leaning on outside data sources, and commentators are flagging signal quality risk 56714.

TransUnion said a tie-up with Bankrate will make it easier for millions of Americans to find personalised credit offers through smarter matching 5. A Finextra community piece looked at Stripe and FedEx treating shipping data as a signal of business health 6, and another set out signal quality and risk issues in managing external data 7.

In markets, Bloomberg wrote on using advanced analytics to look past benchmarks on fixed income trading costs 14. At Sibos 2026, Smartstream's Nick Smith said banks are moving at different speeds towards real-time payments, intraday liquidity management and post-trade automation, largely because of culture 15.

What it means: External data can sharpen credit and trading decisions, but it also brings quality, consent and model risk that must be governed like internal data. Institutions should test outside signals before they feed lending or pricing decisions.

Questions for vendors:

  1. How do you measure and report the quality and freshness of the external data you supply?
  2. What consent and permitted use terms apply to the third-party data in your credit matching?
  3. How do you validate that an alternative data signal still predicts outcomes over time?

In the registry: Market Insights (ProductCloud), AI solutions (Synechron), AI Lineage Assistant (Solidatus)

Regulation and supervision

Sources

  1. Finextra (cloud), 2026-10-09: CaixaBank expands Google Cloud collaboration
  2. FinTech Global, 2026-10-15: Sovereign AI vs shared intelligence: AML’s new dilemma
  3. ECB Banking Supervision, 2026-10-15: Regulatory reporting on profit and loss: institutions will have option to report under IAS 1 or IFRS 18
  4. iTnews, 2026-10-15: ASIC's security and data chiefs to leave in December
  5. Finextra (retail banking), 2026-10-15: TransUnion and Bankrate team up for more personalised credit offers
  6. Finextra (retail banking), 2026-10-14: Stripe and FedEx Lock Onto Shipping Data as a Business Health Signal
  7. Finextra (retail banking), 2026-10-15: Managing external data in fintech: signal quality and risk
  8. Finextra, 2026-10-15: NatWest enlists Nvidia to help its AI spending tool understand regional accents
  9. Finextra (security), 2026-10-15: CLS completes infrastructure migration with Kyndryl
  10. BankInfoSecurity, 2026-10-10: Asos Hackers Claim Breach of Snowflake-Connected Simon AI
  11. NCSC, 2026-10-15: Using TLS to protect data
  12. APRA, 2026-10-15: Life insurance claims and disputes statistics
  13. Federal Reserve, 2026-10-09: Federal Reserve Board releases results of the 2025 Survey of Consumer Finances, which provides the public and policymakers with detailed insights into the economic condition of American families
  14. Bloomberg.com, 2026-10-13: Beyond the benchmark: Using advanced analytics to navigate fixed income trading costs
  15. Finextra, 2026-10-13: Culture Shifts: Why the Industry Is Already Moving Beyond T+1

About data and analytics software

Data and analytics vendors store, model and report on an institution's data: data platforms, warehousing, business intelligence and machine learning.

Browse the products in the registry

Get the market overview every Monday: subscribe to the Vederca newsletter.