Understand before acting.
Agents should work from the knowledge and instructions a team approves, and surface uncertainty when that context is incomplete.
More room for
meaningful work.
Inori is an early-stage product initiative building AI agents for the everyday work of growing teams. Our focus is practical: bring useful context to repetitive workflows and make the next step easier to review.
The work between work
deserves better tools.
A lead waits for a follow-up. A support question needs three searches. A campaign brief lives in one place and its context in another. An internal request moves slowly because nobody has the full picture.
We believe agents can help close those gaps. Inori's product direction is to connect relevant knowledge with focused workflows across sales, customer support, marketing and operations.
We're currently in development and inviting early-access conversations. We have not announced general availability, customer counts or independently verified performance results.
“Useful automation gives people a clearer next step — and the room to choose it.”
THE IDEA BEHIND INORI
WHAT GUIDES THE PRODUCT01 / CONTEXT 02 / CONTROL 03 / FOCUS
Thoughtful by design.
Keep judgment with people.
Permissions, clear proposals and review points are central to our design, especially when a step affects customers, money or business records.
Earn the next workflow.
We prefer a narrow, useful pilot over a broad promise. The goal is to learn from real work before expanding the product.
OUR TECHNOLOGY DIRECTIONClaude at the reasoning layer.
Claude at the reasoning layer.
Inori around the workflow.
Inori is being developed to use Claude, through Anthropic's API, for interpreting context, planning next steps and proposing structured tool calls. Our application layer is intended to control the tools available to an agent, execute permitted actions and return results to the model.
Approved contextKnowledge & instructions
Claude reasoningInterpret & propose
Inori orchestrationPermissions & workflow
Human reviewConfirm sensitive steps
Using a model with tool-use capability does not itself make an action authorized or correct. Each workflow still needs explicit access boundaries, validation and appropriate human oversight. These are product design goals; final capabilities will be confirmed during pilots.
Inori is independently developed. Claude and Anthropic are trademarks of Anthropic. No partnership, certification or endorsement is claimed.
Read Anthropic's tool-use documentationWHERE WE ARE TODAY
Building with early teams.
We're using early-access conversations to understand recurring workflows, integration needs and review requirements. Selected pilots will have a defined scope and separate participation terms.
The website shows our intended product direction and illustrative workflows. It does not provide a live agent workspace, connect to your business tools or start a subscription.
Explore the solutionsA USEFUL EARLY-ACCESS CONVERSATION
Tell us about the work.
Which task repeats every week? Where does its context live? What should always stay with a person?
Those are the questions that help us build a useful pilot.
Share your workflowHELP SHAPE INORIBuild the next step
Build the next step
with us.
We're looking for teams with a clear workflow and a thoughtful approach to automation.
Request Early Access