Databutton Reports Thousands of Hours Saved at Aker QRILL After Three Months.

Databutton has released early production metrics from its deployment at Aker QRILL Company, the world's leading krill producer, showing substantial time savings across operations after three months of use.

The Norwegian company deployed five Riff-built applications across Order Management, Finance, and Supply Chain functions, built in collaboration with domain experts at Aker QRILL. The results show supplier invoice checks dropping from hours to approximately 2.5 minutes, statement reconciliations completing in roughly 2 minutes against live ERP data where most previously went unanswered, and shipping document mismatches being caught before they trigger customs holds.

Databutton's team is preparing an intensive 48-hour build session next week at Aker QRILL's Ålesund operations office, targeting three additional solutions for deployment. Two weeks after that session, the company plans to roll out batch picking and shipping document verifier apps to the customer care team, with training delivered alongside the technology.

The update, shared by Databutton co-founder Andrzej Golebiowski on LinkedIn, frames adoption as the harder challenge. "Building the apps is no longer a blocker," Golebiowski wrote. "Getting the team to actually use them is the harder half, and the part most AI projects stall. We make sure that does not happen."

The deployment follows Databutton's recent work with Glamox, where the company embedded Riff agents into industrial supply chain workflows. That partnership focused on procurement and vendor management across Glamox's European operations.

The Aker QRILL deployment represents a concrete test case for Databutton's thesis that AI-powered app builders can deliver measurable operational improvements when paired with domain expertise and sustained implementation support. The company has been positioning Riff as a platform for vertical-specific applications, particularly in finance and operations.

Aker QRILL's use cases center on document processing and data reconciliation, areas where manual work scales poorly and errors create downstream costs. Customs holds caused by shipping document mismatches, for instance, can delay deliveries by days and incur storage fees. Invoice processing delays affect cash flow visibility and vendor relationships. Statement reconciliations left unanswered create audit risks and financial reporting gaps.

The shipping document verifier showing a trend "from one-in-ten toward zero" on mismatch rates suggests the system is learning patterns specific to Aker QRILL's workflows, though Databutton has not disclosed technical details on how the models are trained or updated.

Databutton recently opened Riff Automations to teams and added Gemini 3 support to the platform. The company has been targeting finance and supply chain teams as primary users, competing in a space where Zapier and Make dominate workflow automation but lack the vertical specificity Databutton is pursuing.

The next phase of the Aker QRILL deployment will test whether Databutton's model of embedded build sessions and managed rollouts can maintain momentum as the scope expands beyond the initial power users to broader teams like customer care.