Pinecone Nexus offers a knowledge engine for agents
Pinecone’s Ash Ashutosh and Edo Liberty say that the primary user is changing.
In their capacities as CEO and founder of Pinecone, the pair remind us that every technological paradigm shift produces a defining data infrastructure category i.e. relational databases for client-server; object stores for clouds; vector databases for assistive AI.
Pinecone works with 800,000+ active developers and 9,000+ paying customers, who run the company’s vector database to power AI that drives semantic search, recommendation systems and retrieval-augmented generation.
That assistive AI category was built for a human user; type a query, get relevant documents back. It worked.
“Now agents are surpassing humans as the primary consumers of knowledge infrastructure. In this agentic AI era, agents are performing tasks, stuck in brute-force loops. [Then they]… retrieve a set of chunks, read them, realize something is missing. Retrieve more. Synthesize. Hit a conflict. Retrieve again. Roughly 85% of an agent’s effort is spent on knowledge retrieval and the output still requires human review before anyone can act on it,” explain Ashutosh & Liberty.
The inevitable result of this work loop is poor task completion rates of 50–60%. Unpredictable completion times. Runaway token costs. This is the “ten blue links” era (i.e. the classic search engine results page displaying a list of website titles, URLs and descriptions for users to click) of agentic retrieval.
Web search moved on
Web search has already made the transition from ranked links to direct answers. Knowledge infrastructure needs the same leap.
The Pinecone pair claim that their vector database is the foundation; vector primitives and their management remain essential. But the retrieval patterns agents need are fundamentally different from what humans need. That is what changed.
Pinecone Nexus
Nexus is a knowledge engine, not a retrieval system. The distinction matters.
“A retrieval system finds documents and hands them to a frontier model at inference time. The model burns tokens, sifting through raw content, introduces latency and risks hallucination. This is reasoning at the retrieval stage. It is expensive, slow and fragile,” states the Pinecone pair. “Nexus moves the reasoning upstream, from retrieval to knowledge compilation. It structures, contextualises and composes specialised contexts (derived artefacts) before the agent needs them. The agent receives trusted knowledge in a context-specific, structured format, not raw documents. It completes the task, not the retrieval. Frontier models are freed to do what they were designed for – intelligent reasoning, not managing knowledge.”
Nexus has two core components: a context compiler and a composable retriever. The context compiler builds and organises knowledge around how your company operates. The composable retriever formats and serves responses precisely for how each agent needs knowledge to complete its task.
The context compiler is the heart of the shift.
Users give it source data and a task spec. It compiles raw data into task-optimised, specialised contexts. These contexts include newly-derived artefacts – the concrete form of information an AI agent acts on. Purpose-built for accuracy and speed, agents consume these artefacts directly.
Burning tokens
Unlike a traditional compiler, it is iterative: it experiments with representations, evaluates them against the task and converges on the precise knowledge structure the agent needs. The work that used to happen at inference time, burning tokens and producing ambiguous results, now happens once at compilation time – and gets better with every iteration.
Take a mid-market SaaS company. Its data lives in a data warehouse, Salesforce, Slack, Gong, Gmail, Jira and Google Drive.
“The current approach is to point a vibe coding tool at all the sources and unleash it to perform the task. It scans everything, retrieves what it can and hopes the right context surfaces. Sometimes it works. Often it hallucinates, misses critical connections, or drowns in irrelevant data,” said the Pinecone team.
The context compiler works differently. It reads the same underlying data but builds specialised context artefacts for each agent’s task.
They say that this is the difference between a system of record and a system of knowledge. The system of record stores what happened. The context compiler does not organise data; it builds what each agent needs to understand your business data – differently for every agent, every task.
Re-used every time.
Early access for Nexus and KnowQL is open now to customers and partners building agent-native applications in financial services, healthcare, legal, enterprise SaaS and any domain where agents reason over complex, proprietary knowledge.