Karsten Wenzlaff, Advisor
August 26th, 2025
Karsten Wenzlaff is a Crowdfunding and Fintech Expert from Germany. He is Secretary-General of Digit[...]
Why Fintechs Are Rebuilding How They Measure Work
AI Learnings | January 8, 2026
AI generated
What KOHO Learned When AI Forced It To Rethink How Work Gets Measured
KOHO's CEO Daniel Eberhard is publicly sharing its AI Playbook and the practical learnings from a two year effort to integrate AI into how the company measures work, allocates resources, and runs teams (see Linkedin post).
The Playbook covers KOHO’s core beliefs, communications approach, resourcing model, telemetry, and execution learnings. It reads like an internal operating guide and what stands out is not excitement about AI but how quickly measurement became the central issue once AI entered daily work.
1. AI Didn't Drive The Biggest Gains. Visibility Did.
AI created focus on telemetry. We got at least as much juice from just better tracking.
This insight changes the AI discussion. Productivity gains didn't come only from automation. They came from finally seeing how work actually happens. Telemetry, in this case, means tracking real workflows using data rather than assumptions or anecdotes.
See: KOHO Secures $190M to Move Towards Banking Power
Once KOHO put systems in place to measure how teams worked, productivity improved even in areas where AI usage was limited. AI forced management to see exposed blind spots that leadership could no longer ignore.
Takeaway: If AI adoption feels underwhelming, the issue may not be the tools. It may be that the organization lacks visibility into where time, effort, and value actually flow.
2. Code Stopped Being The Rate Limiter
We’re now at a place that code is not the rate limiter at KOHO.
When AI enters engineering workflows, producing code becomes easier so the bottleneck moves upstream. Decision making, prioritization, and product clarity start to matter more than execution speed.
Eberhard also flags that traditional measures like lines of code become low signal in this environment. More code does not necessarily mean more value.
Takeaway: Teams built around shipping velocity often struggle when judgment becomes the constraint. Incentives and planning processes must adapt.
3. Pull Requests Became About Value, Not Volume
Lines of code is much steeper year over year than PRs per dev. This is low signal.
Pull requests (PRs) are the units of work developers submit when proposing changes to production code. KOHO found that raw PR counts and code volume stopped telling a meaningful story once AI entered the workflow.
See: KOHO Launches Low Cost Global Money Transfer Service
So KOHO's AI Playbook describes a move toward weighted PR throughput to adjust for complexity, so harder, higher impact work counts more than simple changes.
Takeaway: If teams are still rewarded for volume, AI will amplify the wrong behaviour. Weighting work by impact protects critical but less visible work.
4. Telemetry Exposed Differences That Already Existed
Power users are averaging 8 PRs a week. Low cohort users are 1.5 PRs a week.
This is an uncomfortable but important insight. AI didn't create uneven contribution. It revealed it. Once telemetry grouped developers into cohorts based on usage and output, contribution gaps became visible.
Takeaway: Avoiding measurement doesn't preserve culture. It only delays accountability and clarity.
6. Adoption Was A Change Problem, Not A Tool Problem
People need more hand holding than I thought.
Too busy to learn AI is the most common thing I hear.
When workers are under immense time pressure, curiosity wanes. Even for highly capable teams that might otherwise adopt better tools.
KOHO set expectations without direct mandates. AI use was expected, but people could challenge tools if they could explain why. Judgment and agency remained central.
See: Agentic AI in Banking From Pilots to Real Impact
Takeaway: AI adoption stalls when teams are left to figure it out on their own, and improves once structure, support, and clear expectations are added.
7. Quality Was Defined By Cost To Fix
How we think about quality was anchored in the cost attached to fixing things.
The Playbook references deployment guardrails such as canary deployments (changes are released to a small group first), and blue green deployments (two versions run in parallel to allow quick rollback).
The emphasis isn't on the techniques themselves, but on the economics of failure.
Takeaway: Speed only works when recovery is fast and inexpensive, especially in regulated environments.
What Boards Should Be Asking Now
The most important lesson from KOHO’s AI Playbook is not that AI boosted productivity. Instead, it's that leadership rebuilt how the organization understands work.
- Are we measuring activity or value delivered?
- Which metrics stopped working once AI entered workflows?
- Do we know where today’s real bottlenecks are?
- Can we explain productivity differences honestly?
- Do incentives reward visible work or valuable work?
- Is failure cheap to detect and fix?
- Do we have clear visibility into how work actually flows?
Closing Thoughts
What KOHO’s AI Playbook ultimately shows is that AI does not simply change how much work gets done. It changes what leaders can no longer ignore. Once AI enters daily workflows, old metrics break, contribution differences surface, and decision making becomes the real constraint. Fintechs that treat AI as a tool upgrade will miss this insight. Those that treat AI as an opportunity to rethink measurement, incentives, and accountability are more likely to benefit far more over time.
See: How Fintech Teams Move From Tools To Agents
Kudos to Daniel Eberhard and KOHO for opening up an internal AI Playbook and shared learnings of tradeoffs, frictions, and course corrections. It's a working account of what changed once assumptions met reality. That kind of transparency strengthens the ecosystem. It gives other fintech leaders a more honest starting point, grounded in practice rather than hype.
AI doesn't reduce the need for judgment but it raises the cost of avoiding it.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
 |  |  |
|---|---|---|
 |  |  |
Leave a Reply