V7's context graph enhances AI agents with institutional memory

V7, a company co-founded by Rizzoli and Edwardsson in 2018, has developed an agentic platform called V7 Go that enables AI agents to access and utilize institutional memory through a structured Context Graph. This graph connects entities, relationships, and cited evidence, allowing agents to query and act on organized data from millions of files. V7 claims that agents can complete complex workflows in minutes with 99.9% accuracy, while maintaining an auditable trail of decisions. The platform uses GPT-5.6 Luna for structured extraction and GPT-6 Astra for challenging graph-query tasks, achieving 89% accuracy on the hardest tests. V7 reports that GPT-6 Astra outperforms GPT-5.6 Sol, which scored 78% on very-hard queries. The Context Graph also reduces retrieval costs by 78% compared to GPT-5.4 mini and improves workflow efficiency, with some tasks completing up to 50% faster. Rizzoli emphasized that enterprises will benefit most from AI by leveraging high-quality context rather than sheer agent numbers.