A Zhejiang–Columbia study found that adding a generative-AI assistant before checkout in an online retail platform:
- Increased sales by 16%
- Increased conversion rates by 22%
Hybrid systems — where AI handled routine queries and humans took the complex ones — outperformed pure human teams by 11%.
These are encouraging results. But they are also not surprising.
Retail is Digital by Design
Online retail happens online. Its processes — from product listings to customer support — already live in structured, software-mediated environments. In these contexts, introducing AI is relatively straightforward. The system can:
- Observe data in real time
- Generate outputs
- Immediately measure results
AI thrives where data, context, and user interaction are already digitized. It is much easier to deploy AI and harvest its uplift in a process that was born digital than in one still dependent on fragmented systems, emails, and spreadsheets.
Beyond Revenue: Cost and Hybrid Models
As McDonald rightly notes, the real story is not just about increased sales but also reduced operational costs. The study reinforces what we’re seeing across industries:
- The most consistent gains come from hybrid models, where AI complements and amplifies human work rather than replaces it
- Hybrid is quickly becoming the new keyword of AI adoption
Translating to Financial Services
For financial institutions and asset managers, the path from AI to ROI will depend on identifying the digital equivalents of retail processes — the workflows that are already structured and carried out through digital platforms with minimal human intervention.
These include:
- Client servicing and onboarding, where interactions already pass through CRMs and compliance systems
- Operational reconciliation, reporting, and control, where inputs and outputs are deterministic and rule-based
- Market and risk analytics, where AI can accelerate data processing but humans still govern the final judgment
By contrast, use cases that depend on subjective decision-making — investment allocation, governance approval, client suitability — remain far more complex. These are not yet ready for full AI automation.
The Bottom Line
AI’s measurable gains in retail remind us that success depends on digital readiness. The same logic applies to financial services: before AI can lift productivity, the underlying processes must be digital, structured, and measurable.
The winners will be those who start by identifying the “online retail moments” inside their organizations and redesign those processes with AI augmentation in mind
Hybrid systems — not full automation — will define the first real wave of ROI in asset management
