Rohlik Group, an online grocer in five countries, describes its engineering organization as "agent-mostly." Cognition reports about 90% of code is agent-generated, with agents in planning, code generation, code review, and incident response.
Reported outcomes
Outcomes are reported exactly as stated by the source; we have not independently audited them.
How it works
- Agents operate in every SDLC step: planning, coding, review, incident response.
- Engineers supervise agents.
Risks & compliance
- Agent-written code requires human review and testing
- Vendor-published results; not independently audited
Tools mentioned & how to build it
Disclosure: Some links may earn us a commission; it never affects grades. Tools are listed by fit to this case, never by payment.
Tools mentioned
- Devin (Cognition)
Autonomous AI software engineer from Cognition.
Build a version yourself
- n8n
Open-source workflow automation with AI agent nodes; self-host or cloud. - Relevance AI
Build an 'AI workforce' of agents, e.g. BDR agents that research and book meetings.
Sources
“About 90% of code is agent-generated, and agents operate in every step of the SDLC: planning, code generation, code review, and incident response.”Rohlik Group and Devin
Grade rationale: Named organization with a quantified outcome from the vendor or the organization itself (self-reported). Automated check on Oct 17, 2026: 1/1 quotes and 1/1 outcome claims matched the source. Compiled by the Agent Casefile research desk (AI-assisted; see methodology). Spot an error? Request a correction.
