V7, a London-based startup, has built a platform that gives AI agents access to a company's institutional memory by structuring files, emails, and data into a queryable graph. The system, called V7 Go, uses OpenAI's latest models to cut document-heavy workflows in finance and insurance from days to minutes, with one customer saving $12,000 per task in expert costs.
"To solve hard enterprise use cases across finance and insurance, AI needs to learn how your business operates just as well as it learned from the Internet," said Alberto Rizzoli, Co-Founder and CEO at V7.
Founded in 2018 by Rizzoli and Simon Edwardsson after they built a computer vision accessibility app together, V7 now helps companies teach AI systems their specific business context. That context - which fund report is current, how the same entity is named across three systems - lives scattered across documents, data rooms, spreadsheets, and internal tools. For teams in finance, insurance, and real estate, retrieval accuracy within workflows is non-negotiable.
How the Context Graph organizes company memory
V7 Go connects to repositories like SharePoint and Google Drive, scans them for entities, relationships, facts, and metrics, and populates a Context Graph. This structured record is cheaper and faster to traverse than long-context approaches. When a new file arrives, the system identifies companies, funds, people, or any entity in an ontology, connects each fact to an existing record, and preserves cited evidence linked to the original source. If the graph lacks enough information, V7 Go can still search underlying documents with retrieval-augmented generation.
The structure of that context matters. On HERB, a benchmark for finding and connecting information spread across enterprise systems, V7's retrieval-only system outperformed the official baseline by 69% and reduced hallucinations on unanswerable queries by 38%. That source-linked context keeps complex workflows grounded in each company's own information.
The Context Graph means a model can work with a firm's history without relearning it each time. For long-running agents, V7 Go keeps recent exchanges in the model's active context and stores older material in the graph to be retrieved when needed.
Real-world results across finance and insurance
V7 Go uses this organized context in workflows such as private equity deal screening and insurance underwriting. In one example, a workflow extracts information from a Confidential Information Memorandum, feeds key financials, deal terms, and management details, and produces a screening note with cited risk fields.
Early customer results show concrete time and cost savings. Asset managers now screen deals 21 times faster, reducing a full-day process to 15 minutes. A financial services team cut review time from more than 100 hours to under 10, saving $12,000 in expert costs per task. Insurance teams reduced errors in claims processing by 13.5% after giving agents historical knowledge of all previous claims and existing policies.
Choosing OpenAI for multi-step reasoning
Complex workflows depend on a model reliably following lengthy instructions, running tools, and interpreting results across hundreds of steps. V7 Go maps each step to fast, medium, and smart tiers. GPT-5.6 Luna handles structured extraction and high-volume work, while GPT-5.6 Terra or Sol power steps requiring more reasoning or tool use.
"We chose OpenAI as our default because it performs best on the multi-step tool workflows V7 Go depends on. In our Context Graph benchmark, GPT-5.6 Sol reduced the tool-call error rate from 2.7% with GPT-5.5, to 0.2%," said Simon Edwardsson, Co-Founder and CTO at V7.
V7 also tested GPT-6 Astra on its most difficult queries. On a challenging set of graph-query questions using messy, real-world data across thousands of documents, GPT-5.6 Sol scored 78% on the very-hard level, while GPT-6 Astra reached 89% accuracy. Both models scored close to 100% on easy, medium, and hard levels. In other efficiency gains, key workflows with several external calls now finish up to 50% faster, and GPT-5.6 Luna delivered a 78% lower cost per document than GPT-5.4 mini.
"With GPT-5.6 Terra, we have been able to remove many intermediate workflow stages that previously existed only to simplify the task for the model. It's saved us days of delivery work and often gets things right on the first build of a workflow, thanks to a stronger model and access to more context," Edwardsson said.
Bringing context into ChatGPT and Codex
V7 Go already exposes Context Graph querying and ingestion through its MCP server, so customers can use it from ChatGPT and other compatible clients. They can also create V7 Go workflows through MCP in Codex. Together with simpler workflow design, this has reduced the time required to create a medium-length workflow from around one hour to about 20 minutes.
V7's longer-term goal is to make that shared memory more proactive. The team is working toward workflows that start when facts in the Context Graph change, flag inconsistencies, and show people which analyses need another look. A restated fund report, for example, could prompt V7 Go to flag work that still relies on old figures.
"Our goal is to help enterprises re-tool for the age of AI, with workflows that solve mission critical tasks, and memory that outperforms us humans," Rizzoli said. "Finance firms getting real value from AI will not be the ones with the most agents. They will be the ones with the best context."
Why this matters for operations and management professionals
V7's approach addresses a specific pain point for anyone managing document-heavy, multi-step processes: agents that forget what they learned last time. The Context Graph removes the need for models to rediscover company context on every request, which directly reduces errors, token costs, and time spent on repetitive review work. For teams in finance, insurance, or any field where accuracy and audit trails are mandatory, structured institutional memory moves AI from a helpful assistant to a system that can reliably complete 50-to-100-step workflows with cited evidence for every decision. Professionals looking to build these kinds of automated workflows can explore AI Agent Courses to understand how agentic systems are designed and deployed in enterprise settings.
Your membership also unlocks:
