Standard Bank Group
Standard Bank reports about 20% productivity gains in early AI coding use cases on its Bedrock platform
Standard Bank Group runs an enterprise generative AI platform on Amazon Bedrock, with approved tools for staff. The bank reports that early use cases of AI-enabled coding tools show productivity improvements of approximately 20%, with more than 39,000 employees now active generative AI users. The bank gives no method for the 20% figure.
Key figures
Copied exactly from the sources. Copy any figure with its citation.
approximately 20%
Productivity improvement in early use cases of AI-enabled coding tools, as reported by the bank1
Around one-third
Share of technology employees using AI-enabled coding tools1
What it took
Cost, team size and timeline were not disclosed in the sources.
The problem
Standard Bank argues that access to the newest models is not an edge, because many firms will have similar tools. It wants generative AI to become a capability the whole group can reuse and build into how it serves clients and runs its operations, rather than something only a few teams try.
What they built
The group built an enterprise AI platform on Amazon Bedrock, a service from Amazon Web Services that gives access to several AI models. Staff use a portfolio of approved tools, plus specialist tools for functions such as software engineering. The bank describes a multi-model approach, which means it can compare and govern open and closed models.
How they did it
The bank says its approved enterprise tools have been assessed for security, privacy and responsible use. It picked four lighthouse areas: relationship management, servicing, payments and lending. It says the work builds on several years of investment in its technology platform, including moving 78% of migratable compute to the cloud by the end of the first half of 2026.
Results
Standard Bank reports more than 39,000 active generative AI users, 72% of its workforce, and about one-third of technology staff using AI coding tools. For those tools it says early use cases show productivity improvements of approximately 20%. It gives no method for that figure.
Lessons for your project
Give staff approved tools that have been assessed for security, privacy and responsible use. Standard Bank paired a portfolio of enterprise tools with specialist ones, for example for software engineering.
Keep the model choice open. The bank's platform lets it test and govern both open and closed models, so it can switch as the market moves.
Pick a few lighthouse areas. The bank chose relationship management, servicing, payments and lending for client impact, feasibility and reuse across the group.
Treat early productivity figures as a starting point. Ask for the method and the sample before you use a gain such as 20% in your own business case.
Caveats
- All figures come from Standard Bank's own press release; news reports repeat them and add no independent confirmation.
- The bank gives no method, sample size or list of use cases behind the approximately 20% gain, and calls it an early result.
- No model or vendor is named. The bank only says its Amazon Bedrock platform lets it evaluate open and closed models.
- The press release gives no go-live date. The deployed year is the year of the announcement. The user counts are volumes, not outcomes, and only the approximately 20% figure is a reported result.
Sources
- We've positioned AI as a long-term competitive capability
Standard Bank Group · 2026-08-25 · Press release · standardbank.com
- Standard Bank transitions to AI-enabled organisation
ITWeb · 2026-08-17 · News · itweb.co.za
Summarized in our own words from the sources above. Spotted an error? See corrections.