Skip to main content

Platform

An AI workflow orchestration platform for building on the systems you run

Kinetic is how organizations build applications, workflows and agent-ready processes on top of their ERP, HRIS, ITSM and CRM. It has three layers: experience, workflow and integration. AI builds the portals and workflows through MCP, people and AI agents work the same requests, and your systems of record stay where they are.

Book a working session What ships today

USDA Proof of concept on vSphere, Active Directory and its cloud platforms in 4 days. Server provisioning went from 3 weeks to 30 minutes. Read the USDA case study

The Kinetic console listing the forms in a constituent CRM application: campaigns, change log, community groups, constituents, data steward tasks, department notifications, households, interactions and others, each with a type, an Active status and a last-updated time.
One application in the Kinetic console. Each form holds a kind of record; workflows attach to the forms. City of Cedar Harbor · fictional organization, sample data

Systems of record are hard to change, so people build around them

Most large organizations run some version of these three. Each gets a job done. None of them shares a record with the others.

  • The departmental app

    Facilities buys its own work order tool because changing the ERP would take a year.

  • The tracking spreadsheet

    A new-hire checklist spans three departments. Someone updates it by hand and chases it by email.

  • The standalone agent

    An AI assistant drafts client replies from last month's export of the CRM.

Kinetic gives that building one home. Every portal and process works from the same records and connections, and the next application reuses them.

How Kinetic is built

Three layers between the people doing the work and the systems holding the data

The experience layer is what people see. The workflow layer moves each request along. The integration layer reads and writes the systems underneath, which do not change. Engineers can read the technical architecture.

Kinetic architecture: people and AI agents work through three layers: the Experience layer, the Workflow layer and the Integration layer. Those layers connect to the systems and data the organization already runs. Native agent work exchange in the Integration layer is in development.

Native agent work exchange. Agents pick up and hand off work in Kinetic as participants in the process, the same way people do. Agents already start work today through MCP and the REST API.

In development

The systems and data you already own · stay in place

  • ERP
  • HRIS
  • ITSM
  • CRM
  • Case and permitting
  • Data warehouse and lake

What ships today

Seven of these eight you can use now

The eighth, native agent work exchange, is in development. We label it that way everywhere and give no date.

Kinetic capabilities by layer and release status
Capability Layer Status
Portals, forms, service catalogs and apps for clients, vendors, residents and employees Experience Available
Workflows across systems, with approvals, deadlines and escalation Workflow Available
AI steps inside a workflow that classify, extract, summarize and draft Workflow Available
MCP server: AI builds forms, portals and workflows, and can create records and start workflows within the signed-in user's permissions Workflow Available
Connectors with read and write access to systems of record, and to any system with an OpenAPI spec Integration Available
Reach into systems on your own network, including on-premises applications and databases Integration Available
REST APIs that applications and AI agents call to start work and check its status Integration Available
Native agent work exchange: agents take work from Kinetic and return it as first-class participants Integration In development

One request, worked by a person, a model and your ERP

Bank detail changes are a common fraud target, so finance teams want a person to approve each one. AI does the reading; a person makes the call.

Agents can start work like this through MCP and the REST APIs today. Native agent work exchange is In development

  1. 01 A supplier asks to change its bank details Person or agent The request comes in through the vendor portal. An AI agent could submit the same request through MCP or the API.
  2. 02 The model reads the attachment AI step It extracts the new account details from the bank letter and notes that they differ from the supplier record in the ERP.
  3. 03 Accounts payable decides Person The AP lead reads the AI summary next to the current ERP record, then approves or rejects the change.
  4. 04 Kinetic updates the ERP Workflow It writes the approved details and opens a callback task to the supplier contact on file. Unapproved after two business days, it escalates.
  5. 05 One status Everyone The supplier and the AP team see the same status on the same record.

Where Kinetic runs, and which model it calls

Both are your decision. The platform is the same in every deployment, and it does not come with a model of its own.

Deployment

  • Kinetic-hosted cloud. We run and patch it.
  • Customer-managed. In your own data center or cloud account.
  • Air-gapped. On networks with no outside connection, including DoD IL5 environments.
Compare the deployment options

Models

Anthropic Claude, Amazon Bedrock, Azure OpenAI, open models you host yourself, or any model that supports MCP. A model is one step in a workflow, so switching providers next year means changing that step, not rebuilding the process.

How an AI step works inside a workflow

Where AI fits in Kinetic, and where it stops

AI builds the applications and does the reading inside them. The rest belongs to the workflow and your people.

AI does

  • Build portals and workflows through the MCP server, for your team to review. See a build step by step.
  • Classify, extract, summarize and draft as a step in a running workflow.
  • Create records and start workflows through MCP, with only the signed-in user's permissions.

AI does not

  • Choose the route at run time. The published workflow does, identically on every request.
  • Keep its own copy of your ERP or CRM data.
  • Approve the steps you have reserved for a person.

What you can build next

Most customers start in one of three areas. Whichever comes first, the second application starts with its forms, connections and approval patterns already built.

Frequently asked questions

Answered by Kinetic Data. Last updated .

  • What is an AI workflow orchestration platform?

    Software that runs a business process across several systems and gives AI a defined job inside it. In Kinetic, AI builds the portals and workflows and then works as a step that reads or drafts. The workflow decides what happens next.

  • Does Kinetic replace our ERP, HRIS, ITSM or CRM?

    No. Those systems stay the source of truth, and Kinetic works with them through their own APIs. Kinetic adds what they were never built for, such as a supplier portal or a process that crosses four of them.

  • Is the MCP server available?

    Yes, it is generally available. Most teams use it to build: an assistant such as Claude creates and edits portals and workflows in Kinetic. It can reach any Kinetic API the signed-in user is allowed to use, so it can also create records and start workflows. The workflows themselves still run the same steps every time.

  • How do AI agents take part today?

    Agents start work and check its status through the MCP server and the REST APIs, with the permissions of the account they use, and their requests go through the same approvals as anyone else's. Native agent work exchange, where agents take queued work and hand it back like a colleague would, is in development.

  • Where can Kinetic run?

    In Kinetic's cloud, in your own data center or cloud account, or on an air-gapped network. It has been deployed in DoD IL5 environments.

Bring one process. Leave with a plan to build it.

Two hours with our team, at no cost. Bring one process and the systems it touches. We sketch which layer does what, where AI belongs in it, and send you a written plan afterward.

Book a working session