How Avallone took non-engineer contributions from 2% to 22% in one summer

Avallone is a 360° KYC platform from Copenhagen: software plus expert services that let corporates, banks and funds collect, answer and manage KYC at scale. The company is SOC 2 Type II certified and lives inside compliance workflows all day, which means the product cannot afford sloppy changes and cannot afford slow ones either.

Ahmad Nazir Raja, Avallone's CTO, did not go looking for a vendor. He was already building the tool himself: Merlin, an internal platform to let everyone in the company contribute to the product. Then he met Kosuke, found the thing he was trying to build already running, and made the decision most CTOs hate making. He stopped building.

KYC product work is a stream of small, concrete tickets: a field to add, a flow to adjust, a template to fix, a customer request that takes ten minutes to understand and three weeks to reach production. Every one of them queued behind engineering.

The people closest to the customer, in product, customer success and compliance services, could see exactly what needed to change and could not touch it.

Ahmad's answer was the one a good CTO gives: build the missing piece. Avallone started by making it easy for engineers to spin up the stack, so that prompting was easy. The next step was the cloud, so non-engineers could use it without fighting a local environment.

When he saw Kosuke, the comparison was uncomfortable and quick. He did what few founders manage to do with their own code: he stopped building and started onboarding. The strongest argument for Kosuke at Avallone is not a feature.

Within four months, 40% of Avallone's product and tech team was building on Kosuke. Of the active builders, 37% are not engineers: they are product managers, and they ship their own changes, in a live sandbox, in the browser, with no local setup.

What did not change is the gate. Every change lands as a pull request, and an engineer reviews and merges it.

The share of Avallone's merges built with Kosuke grew roughly 10x between June and August. The merge rate tells the other half of the story: half of the first month's Kosuke-built pull requests got merged, then 90% or better every month since.