When AI takes the wheel, are cloud data platform vendors (including hyperscalers) showing the whole road ahead for Wells Fargo’s data architecture?
The Google Developer Expert and the Independent Consultant studying the impacts of AI/ML at scale:
Norbert Kremer is a leading expert in cloud data architecture, FinOps, and AI engineering, with over 25 years of experience designing and optimizing data platforms for some of the most data-intensive industries. A Google Developer Expert in Data Cloud and an Authorized Google Cloud Trainer, he advises global enterprises on achieving scalable, cost-effective, and secure AI-ready infrastructure.
As Director of the WinterCorp Cloud Data Warehouse Lab, Norbert conducts deep-dive evaluations of platforms like BigQuery, Snowflake, Redshift, and Databricks—often uncovering hidden costs, performance limits, and architectural tradeoffs that vendors don’t disclose. His consulting practice focuses on enabling organizations to move from AI aspiration to operational excellence through trustworthy AI practices, grounded in data quality, security, and alignment with business value.
Norbert is also a frequent instructor and speaker at TDWI, Google Developer Groups, and cloud transformation summits.
Richard Winter is an internationally recognized expert in large-scale data architecture, data platforms, and AI-driven analytics. As CEO of WinterCorp, he has advised three of the top ten U.S. banks and led over 50 enterprise-scale platform evaluations for Fortune 500 firms and government agencies. Known for his vendor-neutral, quantitative approach, Richard helps senior executives de-risk high-stakes data platform decisions that directly impact performance, cost, scalability, and business outcomes.
With more than 30 years of experience, including pioneering work in database engine development and AI-scale architecture, Richard brings deep technical insight and strategic clarity to some of the most complex data challenges in the industry.
The Assumptions Driving AI Failures:
Assumption 1
We’ve pressure-tested this platform. We know what it can handle.
Most tests don’t simulate Wells Fargo’s demanding requirements, where scale, complexity, and rapidly changing data collide. GenAI introduces unprecedented workloads that accelerate all three, revealing platform limits no one expected, and few are prepared for.
Assumption 2
The vendor’s roadmap aligns with our enterprise timeline.
AI services evolve at hyperscaler speed, with features refactored or deprecated in <18 months. Enterprises move at compliance speed. You end up locked into yesterday’s version of “AI at scale”.
Assumption 3
Complex data operations won’t be a problem at scale
Joins, predictive analytics and other operations that manipulate large sets of records in complex ways are a major engineering challenge. Most data platforms don’t do them well at scale Workarounds like nested tables break down under party models, risk, fraud, and regulatory reporting, exactly where banks need precision.
In an AI-driven world, where uncertainty is the norm, many executives are turning to independent evidence to strengthen high-stakes platform decisions.
To give Wells Fargo that evidence quickly and decisively, WinterCorp proposes a focused engagement centered on three critical steps:
01
Define Stakeholder Critical Issues & Target Outcomes
We begin by engaging key stakeholders across departments to align on:
- What outcomes matter most, cost predictability, SLA performance, regulatory resilience, or speed to AI value.
- What risks would be unacceptable, runaway costs, data quality failures, SLA breaches, or lock-in.
This ensures the platform is measured against your business reality, not vendor marketing.
02
Define AI- and Data-Enabled Capabilities Needed
Next, we translate Wells Fargo’s AI ambitions and business needs into specific capabilities the architecture must support, such as:
- Handling high-concurrency, multi-join queries central to fraud, risk, and compliance.
- Managing both structured and unstructured data for grounding AI results.
- Delivering predictable cost models under unpredictable AI query patterns.
This step makes the future concrete and measurable before technology commitments are locked in. - Large and dynamic mixed workloads
- Numerous short queries on rapidly changing data
03
Quantify Requirements & Assess Risks
Finally, we test the architecture against those requirements. Using WinterCorp’s independent frameworks, models and benchmarks, we identify:
- Where the platform performs as expected.
- Where it introduces hidden costs, bottlenecks, or risks.
- What mitigations or alternatives exist to prevent surprises at scale.
- Finding gaps and requirements that the platform will not meet.
The Outcome for Wells Fargo based on facts, measurement, and independent analysis.
Independent, vendor-neutral proof of whether the current data architecture will hold under AI at scale.
A clear view of the risks that matter most to your stakeholders.
Absolute confidence to move forward, knowing your architecture matches the realities of Wells Fargo’s business.
Setup a 20-minute call with Richard
Learn more about WinterCorp’s Process or simply ask us any questions you may have.
