Framboise AI

AI workflow automation

Make work flow. With AI.

Connect incoming events, AI, business rules, and human expertise. Framboise brings your tools and processes into a workflow you can follow.

Tell us about your project Explore the platform

In the cloud. On your infrastructure. Connected to your business.

Framboise AIExample workflow

From freeform order to SAP.

AI classifies the order and extracts its details. A person reviews and approves the data before the workflow creates the order in SAP.

  1. Trigger

    Freeform order arrives

    A customer sends an order by email or document.

  2. Step 1

    AI Classification and Data Extraction

    Identify the order and extract customer, item, and quantity details.

  3. Step 2

    Human review

    Check, correct, and approve the extracted order data.

  4. Step 3

    SAP Order creation

    Create the order in SAP after human approval.

AI prepares the data. Your team approves. SAP receives the order.

What is Framboise?

Automate the known. Escalate the unknown.

Framboise is a gated workflow engine where AI and people collaborate. AI interprets incoming information, business rules control what happens next, and configured review points bring in a person when judgement is needed.

  • Known work, automated

    When a case meets your defined rules, the workflow carries out the next action automatically.

  • Unknown cases, escalated

    Missing information, uncertain results, and exceptions are routed to a person through the review gates you configure.

  • Enterprise-level monitoring

    Follow workflow progress, inspect step results, and track failures and pending human reviews with visibility across your operations.

  • Configured by your experts

    A power user or system integrator connects your systems and defines the business rules, automation steps, and review points.

  • No training needed for end users

    End users work through guided review and approval screens alongside familiar email and business tools, without learning to build or configure workflows.

From input to outcome

A connected process. A clear next step.

Move beyond isolated AI prompts. Build workflows that respond to events, work with information, and bring people in when their judgement matters.

  1. Start with an event

    An email, a file, a message, or a schedule can set your workflow in motion.

  2. Make sense of information

    Classify content and extract structured fields with AI, ready for the steps that follow.

  3. Keep people in the loop

    Add human review and approval where a workflow needs your team’s judgement.

  4. Connect the next action

    Work with email, file storage, productivity tools, and APIs in the same flow.

Gated flows

AI interprets. Your rules decide.

Some decisions need a predictable answer. Define those decisions with algorithms and explicit conditions, while AI interprets information inside the steps you configure.

  1. Interpret the input

    Use AI to classify a message, extract fields, or prepare information for the next step.

  2. Apply a defined gate

    Check the result against your rules. Conditions determine which configured branch the workflow follows.

  3. Take the permitted action

    Connect each branch to the next action, or to a review step when the result needs a person’s judgement.

Put rule-sensitive decisions in the workflow itself, with checks between AI output and the actions it can trigger.

Human–AI collaboration

Make room for human judgement.

Every process has corner cases. Build review into the workflow so your team can handle exceptions, add context, and approve the decisions that need their expertise.

Your workflow defines where review happens. AI handles the configured work; people contribute where judgement matters.

  1. Choose the review points

    Define when a missing field, uncertain result, or business condition needs human attention.

  2. Bring in your team

    Use a configured review step to let a person assess the information and provide a decision.

  3. Continue the process

    Use the review outcome to guide the next configured step, keeping people and automation in the same flow.

Enterprise connectivity

Connected to the tools you work with.

Build across model providers, business applications, storage, and messaging. Framboise adapters connect the systems that carry your processes.

  • OpenAI
  • Anthropic
  • xAI
  • Azure
  • SAP
  • Jira
  • Google Workspace
  • Outlook / Office 365
  • Trello
  • Shopify
  • Meta

OpenAPI integration

Connect any custom enterprise application with an OpenAPI-compatible interface. The OpenAPI connector turns its API operations into asynchronous workflow steps, so your own systems can take part in the same process.

Data security

Your data. Your deployment choice.

Keep everything in-house with a fully local setup, or choose the external models and cloud services that fit your data-protection requirements. Where your data goes depends on the complete configuration.

  • Fully local. All data in-house.

    Run Framboise and a local AI model on a secured server in your own infrastructure. Documents, prompts, results, and workflow records stay within your environment.

    Keeping all data in-house requires local models, integrations, storage, logs, and backups. Your team controls access and server maintenance.

  • Local engine. External models.

    Framboise can run on your server while calling an external AI model. Prompts and supporting content sent to that model leave your environment for processing.

    The provider’s processing locations, retention terms, and data protections apply to the data you send. Connected external services have their own data boundaries.

  • Cloud deployment.

    The workflow engine and its stored data run in the selected cloud environment. Your hosting region, model deployment, and integrations determine where data is stored and processed.

    Choose hosting and model-processing locations together, with access controls and data-handling settings that meet your requirements.

Why Framboise

Built around the way work happens.

Bring rules, AI, people, and your existing systems into one connected process. Each has a clear role in getting the work done.

  • Rules where they matter

    Use explicit conditions and algorithms for decisions that need to be deterministic.

  • AI with a clear role

    Apply AI to interpreting content and preparing information inside a defined workflow.

  • People in the process

    Design review points for exceptions, approvals, and the context only your team can provide.

  • Your infrastructure choice

    Start in the cloud or run on-premises with supplied hardware or the infrastructure you maintain.

  • Connected business tools

    Bring data and actions together through adapters for the systems your business uses.