ConvOpsConvOps
ConvOps Brain

The way we work now.

Not another blank chat. ConvOps starts each AI session inside your process, with the rules and remembered context your team already works from.

Start with ConvOps See how it works
Claude · new session
Y

Let's continue the vendor security review.

tasks_searchVendor security review · in progress
context_query4 memories · 3 routes

Picking up where we left off. No re-brief needed:

  • decisionRisk tiering follows the v2 rubric.
  • gotchaTheir SOC 2 was outdated, current one requested.
  • nextDraft the DPA addendum. docs/security/dpa-addendum.md
3
Layers present when work starts: process, rules, context
2
Brain layers underneath: WHAT memories and WHERE routes
0
Blank re-briefs when relevant context is stored
WHAT knowledge

Memories preserve decisions, learnings, preferences, and gotchas.

Brain stores durable knowledge as memories: the decisions a team made, the correction that should not be forgotten, the customer preference that changes the answer, or the implementation gotcha that saves a future agent from rediscovering it.

Memories · WHAT knowledge
  • decisionRefunds over $500 need manager sign-off.
  • learningEnterprise trials convert better with a guided data import.
  • preferenceThis account prefers Slack over email for updates.
  • correctionStore this customer's data in the EU region.
  • gotchaTheir SSO fails on a trailing slash in the callback URL.
WHERE knowledge

Semantic routes tell agents where the work lives.

Routes are not memories. They are the map: file paths, code locations, MCP resources, docs, commands, or agent handoffs that tell an AI where to go when a topic comes up again.

Routes · WHERE knowledge
  • Refund policyfile
    docs/policies/refunds.md
  • Onboarding flowfile
    src/onboarding/flow.ts
  • Plan catalogresource
    api://billing/plans
  • Reset sandboxcommand
    make reset-sandbox
Session start

Session start loads the context layer in one move.

At the start of a session, context_query searches both Brain layers: memories for what matters, routes for where to look next. The mechanism stays technical. The experience is simple: work starts with the right background already present.

Claude · new session
Y

Let's continue the vendor security review.

tasks_searchVendor security review · in progress
context_query4 memories · 3 routes

Picking up where we left off. No re-brief needed:

  • decisionRisk tiering follows the v2 rubric.
  • gotchaTheir SOC 2 was outdated, current one requested.
  • nextDraft the DPA addendum. docs/security/dpa-addendum.md
Cross-session continuity

A different AI can pick up with the same remembered context.

ConvOps workflows preserve process state. Brain preserves the context around that state. Together they make handoffs less dependent on copy-pasted summaries and one person remembering every prior detail.

In practice
  • Resume later without rebuilding the whole brief.
  • Hand work to another AI with the relevant context available.
  • Keep workflow state and remembered context working together.
Grounded, not omniscient

Brain makes saved context findable. It does not pretend to know everything.

The value is not omniscience. Brain helps agents retrieve context that has been stored or routed on purpose, then continue through the workflow with better grounding and less repeated discovery.

In practice
  • Stored memories are explicit knowledge, not mystical awareness.
  • Routes point to known places; they do not replace verification.
  • Agents still follow workflows, gates, and policies when context is loaded.
See it in motion

See a session start with its operating layer attached

A task resumes with process state, rules, memories, and routes already available. The AI is not waiting for a human to rebuild the whole brief.

1
2
3
4
5
6
Claude
Watch the workflow in action
0 / 6

Blank chats vs the way work happens now

Without Brain

  • Every new session starts by asking the operator to rebuild the brief.
  • The AI has to rediscover process, rules, files, and prior decisions.
  • Important corrections stay trapped in old chats and summaries.
  • Work depends on whoever remembers what happened last time.

With Brain

  • The session opens with process, rules, memories, and routes attached.
  • Agents start from the team operating model instead of a blank prompt.
  • Decisions, gotchas, and preferences become reusable work context.
  • Workflows, policies, and remembered context move together.
Questions

Frequently asked questions

Memories are WHAT knowledge: decisions, learnings, preferences, corrections, and gotchas. Routes are WHERE knowledge: file paths, code locations, MCP resources, docs, commands, or handlers that tell an AI where to go.

context_query searches both layers at session start. It recalls relevant memories and finds semantic routes, giving the AI useful context before it explores files or asks the operator to re-explain.

No. Workflows still define process and movement. Brain preserves context and routing around that process so each step can run with the right background.

Policies are reusable behavioral rules injected into the active step. Brain is persistent context and routing. Policies tell the AI how to behave; Brain helps the AI remember what matters and where to look.

Yes. Brain belongs to ConvOps, not a single AI vendor. Any AI using ConvOps can benefit from the stored memories and routes available to the workflow.

No. It retrieves context that has been stored or routed intentionally. Agents still need to verify current facts, follow the workflow, obey gates, and apply active policies.
Continue the tour

Explore the rest of the platform.

01

Policies

See how reusable behavioral rules are injected into the active workflow step.

Read more
02

Workflow Engine

See how remembered context supports the process that drives each task.

Read more
ConvOps

The way we work now.

Not another blank chat. ConvOps starts each AI session inside your process, with the rules and remembered context your team already works from.

Start with ConvOps Learn how it works