Build, operate, and govern business applications
Define your data, design the screens, automate the process, and add AI agents, all on a single platform with permissions and auditing. Build with Studio, the visual editor, or describe what you need and the AI Builders, AI agents that build, put it together for you. What you build runs on the same platform, standalone or integrated with your ERP.
What a solution is made of
It isn't a development environment: it's the system where your process runs every day.
Data
Entities, records, and business rules.
Processes
Workflows, states, and rules.
Screens
Portal, views, and forms by role, on web, desktop, and mobile.
AI agents
Agents with the user's permissions.
Operations
Reports, dashboards, evidence, and control.
01 · Build
Build
Model your domain
Define your business entities and their rules without reinstalling the platform. The model reflects how your company works, not the other way around.
- More than 30 field types: text, numbers, dates, selection, 1:N and N:N relations, references, attachments, signature, and image
- Formulas and computed fields, also in Python when needed
- Validations and default values
- Dynamic conditions: show, hide, lock, or require fields based on the record's status
- Sequences and automatic numbering
- Model templates to reuse structures, and custom fields on existing models
The model runs on a Python ORM. Changes are applied live, without reinstalling the platform core.
Build the experience
The same data can be viewed and worked with in several ways. Choose the view type that best represents each role's work, without coding.
Views computed from your data and dashboards that the AI Builders put together for you. And on the plant floor or in the field: a desktop app with local printing, and a mobile app.
Automate with workflows
Model the lifecycle of each process in a visual editor: states, transitions, and the actions that happen at each step.
- States and transitions, with buttons right in the view
- Triggers: a button, a schedule, an API call or webhook, or another workflow
- Ready-made actions: email, Gmail, Outlook, Google Drive, HTTP calls, and Python logic
- Steps that ask the user for data partway through the workflow
- AI agent nodes: the workflow waits while the agent works and resumes when it finishes
- Execution in segments, with rollback on errors and a log of every run
A node's logic is written in Python when needed. The rest is assembled in the visual editor.
Reports and documents
Every process ends in a document: certificates, batch reports, attestations. Design them once and generate them from any record.
- Template editor, with shared headers, footers, and styles
- PDF and Excel output
- Reports in multiple languages
- QR codes to verify or trace each document
- Email delivery from a workflow and local printing from the desktop app
Certificate of analysis
Sample M-2611-0409 · Batch 24-112
Extend without limits
Whatever the platform doesn't include out of the box can be built: custom interfaces, your own logic, and an integrated code environment.
- Custom Views: embedded React components, with a preview before they're published
- Python logic you can reuse as a workflow step, an agent tool, or a view's data source
- Reusable packages of code, styles, and assets
- Code editor with change history: compare, commit, and revert
- Integrated development environment for model and package code
React on the front end, Python on the back end, without setting up or maintaining your own infrastructure.
Build with AI
Describe what you need and the AI Builders, AI agents that build, propose a plan. You review and approve it, and only then do they build.
- From need to plan, and from plan to models, views, workflows, dashboards, reports, code, and agents
- Specialized builders (modeling, workflows, analytics, reports, code) that coordinate with each other
- A wizard to create a new solution and start building it with AI
- End-to-end scenario tests to verify what was built before publishing it
- Everything they generate is subject to permissions, audit, and your approval
It's not “generate and pray”: the plan is approved before anything is built, and the code and the solution are versioned.
We need to record supplier nonconformances and see a monthly dashboard by supplier.
02 · Operate
Operate
AI agents that work within the process, alongside your team and with your company's context.
Jump to: Agents · Workspace and artifacts · Context · Collaboration
AI agents that operate your process
AI agents are programs that work inside the platform, with the user's permissions, defined limits, and a record of every action. Fixed rules stay in the workflows. Agents take care of what a rule can't: reading, connecting, and drafting.
What they do
- Set up and configure the system
- Search and query information in your data
- Summarize records, cases, and history
- Automate repetitive process tasks
Customize them to operate within your workflows: define their task, their tools, and their limits.
Where they work
- On every record: an AI panel in any view. The form locks while the agent works and refreshes when it finishes
- In workflows, as one more step in the process
- In the background: tasks that notify you in the inbox when they finish
- With your documents: they read images and PDFs, even scanned ones
- In their workspace: each conversation has its own, with files, drafts, and deliverables (learn more)
Two types of agents work side by side, with different roles:
| AI Builders (extend) | Agents that operate | |
|---|---|---|
| What they do | Build or adapt the solution | Carry out business process tasks |
| Who uses them | The implementation team | The people in the process, supervised by their manager |
| Outcome | Models, views, workflows, reports, and agents | Tasks completed within the process |
AI agents expand operational capacity for repetitive tasks. Process owners define their limits and which actions require approval.
From request to deliverable, and from deliverable to solution
Users don't just ask for answers: they ask for an Excel file with subtotals, a report, the analysis of a file. Each conversation with the agents has its own workspace, and whatever is worth keeping is saved as an artifact.
One workspace per conversation
- Files, drafts, and intermediate data, shared by the agents and tasks in that conversation
- Working notes that the agent keeps from one turn to the next
- Upload your files (Excel, PDF, images) and the agent analyzes them, compares them, or loads them into the platform
- A panel next to the chat to view, download, or receive each file, with previews of PDFs, spreadsheets, HTML, and images
Artifacts: versioned deliverables
- Documents, reports, spreadsheets, and charts with version history and comparison between versions
- Export to PDF, and back into a conversation to keep improving them
- Shared within the organization, with users or groups, with an audit log
Put together an Excel file with this quarter's nonconformances by supplier, with subtotals and a summary.
Done. I matched the nonconformances against the supplier file you uploaded and left the Excel file in the workspace.
Workspace
Artifact
You decide what is kept and who it's shared with. The agent can suggest saving a result as an artifact, but it never saves or shares it on its own.
From deliverable to solution
Conversation
You ask for an analysis, a report, or a file.
Workspace
The agent works with your data and files and delivers the result.
Artifact
Whatever is worth keeping is saved, versioned, and shared.
Solution
If it's needed on an ongoing basis, the AI Builders turn it into a view with live data or a recurring report. It can also be attached to a record.
With your company's context
A useful agent knows the case and the organization. Kalenis gives agents that context within the platform, always with the user's permissions.
Memory
Agents remember what they've learned, with a defined scope and periodic consolidation of what is stored.
Per-record dossier
Each record brings together its background, findings, and decisions, which the agent consults before acting.
Knowledge panel
Documents, comments, attachments, activity, and versions of each record, in one place.
Business documentation
Your procedures and definitions, written once and available in context to people and agents.
People and agents, in the same conversation
Collaboration happens next to the data, not in another tool.
Head of quality: @Summaries what happened with this batch before release?
The batch had a temperature deviation at stage 3, which Quality assessed and closed. The 5 subsequent tests were compliant. I've attached the details and the certificate.
03 · Govern
Govern
What IT needs to control the platform: who can access what, what agents do, how changes are published, and where it's installed.
Jump to: Roles and access · Governed AI · Lifecycle · Integrations · Deployment
Who sees and does what, in as much detail as you need
Define each role's access and verify it before you publish.
- Roles per solution, with their enabled menu, views, and actions
- Per view: read-only, record filters, visible fields, and permission to create or delete
- Access matrix per solution and preview as another user
- Record features by role: comments, attachments, activity, and chat
- An access agent that explains why a user can or can't see something and proposes the changes for your approval
The operator can't see the “Cost” field because the Batch cards role line hides it. Should I propose the change for your approval?
Agents under the same governance model
The agent is just another user of the system: it inherits the permissions of whoever uses it, operates within the limits you define, and leaves a trace of every action.
The user's permissions
An agent never exceeds the user's access: whatever the role can't see or do, the agent can't either.
Limits and approval
You define what each agent can do and which actions require approval before they run.
Audit of every action
Every action is logged and attributed to the user, the agent, and the session, along with its history.
Escalation
Sensitive decisions are escalated to a person in charge.
Agents by role
Which agents each role uses, and whether they can act or only query.
AI model policy
Define which models each agent uses. Access to AI providers is centralized: your instance doesn't store your keys.
Package, version, and distribute
Build and test in a development or test environment, and promote to production when it's ready.
- Packaged solutions: models, views, workflows, reports, agents, and custom views, in a single package
- App Center: publish versions to a repository and update each instance from there
- Controlled promotion between environments: development and test → production
- Dependencies between solutions, installed automatically
- All-or-nothing installation: if anything fails, the instance stays as it was
- Scenario tests in test environments, with simulated external services
A development → test → production lifecycle, with packaging, promotion, and versioning.
Integrations Optional
Kalenis operates standalone. If your organization already uses SAP, Oracle, or another core system, Kalenis doesn't replace it: it covers the processes that fall outside that system's scope and sends validated data back when needed.
Without an ERP
The entire process is managed in Kalenis: data capture, operations, evidence, and statuses.
With an ERP
Kalenis handles what the ERP doesn't cover and, if required, sends validated data to your core system.
How it integrates
Common systems: SAP, Oracle, and other ERPs or core systems. See use cases
Does your lab use Kalenis LIMS? Kalenis adds on to your LIMS and works on the same data, with no migration: AI agents, dashboards, statistics, and report templates. See Kalenis + Kalenis LIMS (in Spanish)
Cloud or on-premises
The same platform for business and IT, in the deployment model your security and data governance policies require.
Cloud
Kalenis manages the infrastructure, updates, and operations.
On-premises
Installed on your own infrastructure or private cloud.
Web, desktop, and mobile
Web portal, desktop app, and mobile app for Android.
Multi-company
Multiple companies on a single installation.
Everyone builds with AI. The difference is depth
Work platforms, enterprise low-code, and app generators all build with AI now. The difference is what lies underneath and how it's governed. The table compares them criterion by criterion.
| Criterion | Kalenis | AI work platformsmonday, Airtable, Smartsheet | AI app generatorsLovable, Bolt, v0 | Enterprise low-codePower Apps, OutSystems, Mendix |
|---|---|---|---|---|
| Build | ||||
| AI-assisted building | Agents that plan, build, and test on the platform; the plan is approved before building | AI builders inside the platform (e.g., monday vibe, Airtable Omni) | Generate the full application from a description | AI builders inside the suite (e.g., Power Apps Vibe) |
| Relational data model | Models with relationships, constraints, and calculated fields | Linked tables or boards, without the integrity of a relational database | Whatever database each project generates; evolving it falls to the team | The platform's relational database |
| Business logic in code | Python in fields, rules, workflows, and agents' tools | Formulas and automations; limited scripting | It's all code: flexible, but maintaining it falls to the team | Pro-code extensions, depending on the suite |
| Tests before publishing | End-to-end scenario tests, with simulated external services | Mostly manual; some offer a test environment on higher plans | Whatever the team writes | Testing tools and separate environments |
| Promotion between environments and distribution | Development and test → production, with versions; the same solution installs on other instances | Per-base sandbox in some products; no versions or distribution to other accounts | Through the deployment pipeline the team builds | Formal lifecycle across environments (ALM) |
| Operate and govern | ||||
| Governed platform in production | Multi-company, with permissions and continuous operation, and no code left to maintain | Vendor's platform; governance controls depend on the plan | Delivers a standalone application; running and governing it falls to the team | Enterprise platform with a formal lifecycle |
| Workflows that roll back on errors | Execution in segments: on an error, the segment rolls back and is logged | Step-by-step automations: if a step fails, earlier steps stay applied | Depends on how it's coded | Depends on the component; usually has to be designed |
| Permissions, roles, and auditing | Roles, permissions, and traceability included | Adequate controls; auditing depth varies | Must be designed and developed | Platform-level security and compliance |
| AI agents with permissions and escalation | Same permissions as the user, limits, auditing, and escalation to a person in charge | Built-in agents; permissions, limits, and human approval vary by product | Agents embedded in code or running as separate services | RPA or per-module AI; not always with the user's permissions |
| Implement | ||||
| Ready-to-use business solutions | QC, EBR, Batch cards, and Field services, ready to implement and adapt | Templates; the process has to be built | Built from scratch, starting from an idea | Accelerators; implementation is typically custom |
| Cloud or your own infrastructure | The same platform in the Cloud or on-premises | Only in the vendor's cloud | Wherever the generated app is hosted | Depends on the suite: some are cloud-only, others also run on your own infrastructure |
| ERP integration (optional) | Runs on its own; with an ERP, covers what the ERP doesn't | Connectors; the process usually lives in the tool | Custom-built integrations | Designed for enterprise environments |
| Who runs it | Business teams and IT on a shared platform | Business teams, with IT handling administration | Developers or the team that owns the code | IT and centers of excellence |
| No per-user fees | No per-user or per-seat licenses: bring in everyone who takes part in the process | Per-seat licenses: every person you add raises the cost | No end-user fees, but plans per builder or workspace, plus hosting and maintenance | Cost grows with users, often on top of per-app or capacity charges |
About the platform
Have another question? Write to us and we'll answer it based on your case.
What is Kalenis?
Kalenis is a business application platform where people and AI agents work on the same processes, with permissions and a record of every action. It includes ready-to-use solutions and a team that implements them.
Do I need to write code?
Not for the usual work. Data, screens, workflows, and reports are built in Studio, the visual editor, or with the AI Builders. When custom logic is needed, it's written in Python inside the platform.
What is an AI agent in Kalenis?
It is a program that works inside the platform with the user's permissions. It reads documents, connects information, and drafts. Every action is recorded, and sensitive decisions are escalated to a person.
Does it replace the ERP?
No. It covers the processes the ERP doesn't and, when needed, sends validated data back to it. It also works without an ERP.
Where is it installed, and how do people use it?
In the Cloud or on your own infrastructure (on-premises), with the same subscription. People use it from the browser, the desktop app, and the Android mobile app.
How is it different from other platforms that build with AI?
In what lies underneath: a relational data model, Python logic, workflows that roll back on errors, tests before publishing, and controlled promotion to production. The comparison covers it in detail.
See it working with your process
We'll show you the platform based on your case: how it's built, how the agents operate, and how it's governed.