Skip to content

MCP Roadmap and CorpusIQ

The Model Context Protocol team publishes a roadmap that sets the direction for protocol work. The updated roadmap (August 2026) organizes the work into five priority areas. Several of them describe patterns CorpusIQ already shipped. This page maps each priority area to the CorpusIQ implementation.

Progressive discovery

The roadmap states that connecting to a server with a hundred tools means the model pays for the entire surface before the user asks a single question, and tool selection gets worse as the list grows. The protocol is starting a progressive discovery effort so a server can offer a small entry point and reveal more of its catalog as the conversation narrows.

CorpusIQ ships this pattern today. A single model-facing entry point resolves the question into a governed workflow, and the tool surface is revealed incrementally. The model makes one decision; CorpusIQ does the routing.

Tool result contracts

The roadmap identifies that a tool call response can carry the same output in more than one form, and a server developer has no way to know which form a client will put in front of the model. The goal is to standardize one clear contract.

CorpusIQ already defines a single output contract for business answers: the AnswerSpec. Every answer carries the reconciliation summary, the pinned metric definitions, the evidence chain, and the canonical text. One schema, one contract, every client.

Stateless servers

The 2026-07-28 specification release removed protocol-level sessions and the initialization handshake. Servers can now scale horizontally without holding state.

CorpusIQ builds on this: answers are self-contained. The rendered result is a direct function of the verified tool result, so the same question renders the same answer on any host, after any refresh, without session memory.

Agent identity and enterprise security

The roadmap prioritizes standardized agent identity: DPoP proof of possession, workload identity federation, and standard token exchange, so servers can trust agents that act on behalf of users who are not present.

CorpusIQ connects to this direction with per-connector read-only OAuth, scoped per-account access, and control-plane tooling that never touches customer systems. The roadmap direction and the CorpusIQ security posture converge on the same principle: an agent is a first-class identity with scoped authority.

What this means

The protocol is standardizing the patterns CorpusIQ built first. That is the strongest possible validation of the architecture: the single entry point, the one output contract, stateless rendering, and agent identity are becoming protocol primitives, and CorpusIQ ships them today for governed business answers.

Contact Terms Privacy Pricing