Tour 01
Enterprise knowledge
Find an approved policy, retrieve relevant context and prepare a useful answer.
Take this tourProduct tour
See how Pūnaha handles work you already recognise
Choose a short tour or step through one of the deeper workflow examples. Each one shows where information comes from, what the system does and where a person stays involved. The examples are representative, so there is no invented customer story behind them.
Choose a route
Choose the kind of work or control you want to follow
Each route uses a different mix of knowledge, models, review points and controls, but the way through the work always remains understandable.
Tour 02
Regulated research
Gather evidence, use AI assistance and keep a subject-matter reviewer at the decision point.
Take this tourTour 03
Flexible AI workflow
Use local and external model deployments at different stages, then merge the results.
Take this tourTour 04
Customer support
Retrieve approved guidance and prepare a response for a person to review before use.
Take this tourExplore
Assign and review access
Follow access from a named resource and exact action through to an explainable decision and formal review.
Explore access governanceExplore
Choose a deployment shape
See how one executable can run all services together or support separate service pools.
Explore deployment choicesDeep dive
Three workflow walkthroughs
Inspect seven-stage fictional workflows, then watch or download their narrated videos and scripts.
Open the showcaseAsk Pūnaha
Ask the question you would ask in a first meeting
Choose a starting question and see how Pūnaha uses approved information to build its response.
Guided demonstration · Not a live AI response. No information is sent or saved.
Guided answer
Where can Pūnaha run?
Pūnaha runs on customer-controlled infrastructure. The right service shape depends on your security boundaries, operating model and integration requirements.
How the response is formed
Read the source: Private AI deployment- Understand the deployment requirement
- Choose customer-controlled infrastructure or cloud
- Configure the agreed boundary
Guided answer
How does Pūnaha find reliable answers?
Pūnaha searches approved connected knowledge, retrieves relevant context and provides supporting sources for a person to review. The source remains visible; the answer is not presented as automatically correct.
How the response is formed
Read the source: Enterprise AI search- Ask a question
- Retrieve approved knowledge
- Prepare an answer with sources
Guided answer
Can we use different AI models?
Yes. Local and external model deployments can be configured for different stages of a workflow, with their results brought together for the next step.
How the response is formed
Read the source: Flexible AI workflows- Define the task
- Select a configured model for each stage
- Bring the results together
Guided answer
Where does human approval fit?
A workflow can pause at a review or decision point when human judgement is required. Approval is optional, so an appropriate governed process can continue automatically.
How the response is formed
Read the source: AI orchestration- Complete AI-assisted work
- Pause at the configured review point
- Continue after the decision
Guided answer
What if the next task cannot be known in advance?
Use a CMMN-based adaptive case when evidence, events and professional judgement determine the next appropriate work. Repeatable parts of the case can still call published workflows and decisions.
How the response is formed
Read the source: Adaptive case management- Open a governed case type
- Respond to evidence and events
- Plan appropriate work within the published guardrails
Guided answer
What can developers configure?
Developers can connect knowledge and models, configure visual workflow modules, define BPMN-based processes and CMMN-based cases, and inspect the resulting execution state.
How the response is formed
Read the source: Pūnaha for developers- Connect the required systems
- Configure models and workflow stages
- Test the complete process
Guided answer
Can we control access to one particular workflow or connection?
Yes. Pūnaha can apply an exact action to a named resource for a person, group or service. The effective-access view explains why the final decision was allowed or denied.
How the response is formed
Read the source: Access governance- Choose the named resource and action
- Assign the appropriate person, group or service
- Inspect and review the effective access
Guided answer
Can different Pūnaha services run separately?
Yes. The same executable can run all enabled services together or selected control, workflow, AI and browser-facing services in separate pools. The chosen topology still needs environment-specific acceptance testing.
How the response is formed
Read the source: Private AI deployment- Describe the operating environment
- Choose all-in-one or separated services
- Validate storage, identity, readiness and failure behaviour
Guided answer
Are these customer results?
No. This guided demonstration uses approved public Pūnaha content and representative product patterns. It does not present customer case studies, testimonials or invented results.
How the response is formed
Read the source: About Pūnaha- Recognise the question
- Apply the public claim boundary
- Provide a factual response
Have a different question? Bring it to a private demonstration.
What to notice
You can see where each important choice is made
Knowledge choice
A workflow retrieves from the approved knowledge store for that process.
Model choice
Each model step refers to a supported local or external deployment selected for the task.
Human choice
Approval can pause the workflow when a person needs to review or decide.
Access choice
A named resource can have exact actions for people, groups and services, with an explanation of the final decision.
Work structure
Use a repeatable BPMN-based process or a CMMN-based case according to how predictable the work really is.
Deployment choice
The platform runs on customer-controlled infrastructure using an all-in-one or separated service shape.
Before you start the tour
Are these customer case studies?
No. They are product tours based on representative workflow patterns. Pūnaha does not currently present public customer case studies or invented results.
Can a private demo use a different workflow?
Yes. The public tours explain common patterns. A private demonstration can focus on the sources, model choices, review points and outcome you want to evaluate.
Do all workflows need human approval?
No. Approval is available when the process requires human judgement. A workflow can also continue automatically where that is suitable and governed.
Now bring the process you actually care about
Tell us what people need to find, which systems or models are involved and where someone needs to make a decision.