YiENTi

Enterprise Information & Graph-Native Intelligence

Structure systems.Understand relationships.Build real software.

YiENTi represents processes, relationships, dependencies and software structures explicitly, rather than relying only on language-based context.

YiENTi Enterprise
Intelligent Enterprise Information
YiENTi Intelligence
Graph-Native Intelligence Engine

YiENTi combines an Intelligent Enterprise Information platform with a Graph-Native Intelligence Engine. It can use language models where useful, but its core reasoning model is structural, process-aware and independent of any single LLM provider.

One vision. Two core technologies. From model to software
EventContextKnowledgeProcessRuleEntitySystemGatewayDecision

An event triggers a process. Context and rules become relevant. The gateway resolves to a decision.

Since 2010
Evolving from Logical Board to YiENTi
Two pillars
Enterprise Information + Intelligence Engine
LLM-optional
Providers are replaceable components
In use
Core technology in ACCENTSYS engineering

The YiENTi platform

One vision. Two core technologies.

YiENTi is an umbrella technology with two distinct, connected pillars. One represents the operational reality of an organisation. The other reasons over structured context — inside YiENTi Enterprise and beyond it.

YiENTi Enterprise

Intelligent Enterprise Information

Connect people, processes, systems, tasks, knowledge and operational context in a structured enterprise environment.

Explore YiENTi Enterprise

YiENTi Intelligence

Graph-Native Intelligence Engine

Reason across relationships, processes, state, context, software structures and multimodal information.

Explore YiENTi Intelligence

YiENTi at the center

YiENTi is the center. Everything else connects to it.

YiENTi is the structural intelligence and engineering core. Cloud platforms, language models, Playwright, IDEs and local AI tooling are connected systems around YiENTi — each with a distinct role. They are not YiENTi itself.

YiENTi
StructureMeasurementReasoning

  • Structured modeling

    What YiENTi reasons over · → YiENTi

    BPMN · Use cases · Activities · Sequences · Classes · Data / entity models

  • Generative AI

    External · replaceable · ↔ YiENTi

    OpenAI · Anthropic · Perplexity · Gemini · Lovable · Zapier · Pickaxe · Grok · Jasper · Abacus.AI · Relevance AI · Notion AI · And more coming soon · Local / alternative models

  • Developer tooling

    Development environment · ↔ YiENTi

    Visual Studio Code · OpenCode

  • Local / model infrastructure

    Running models locally · ↔ YiENTi

    Ollama · Lemonade · Other compatible local model workflows

  • Testing

    Validation of real behavior · ↔ YiENTi

    Playwright · Abacus.AI · ChatGPT Work · Visual validation · Functional validation

  • Deployment

    Where generated software runs · YiENTi →

    AWS · Microsoft Azure · Google Cloud · Private infrastructure (future)

  • Real software

    Built with YiENTi · YiENTi →

    Every ACCENTSYS product

Feeds into YiENTiExchanges with YiENTiProduced or delivered by YiENTi-supported engineeringExternal & replaceableFuture direction

The structural core

Language is an interface.
The graph is the model.

These structures are not decorative diagrams. They represent machine-understandable system and process state — the model YiENTi reasons over.

  • BPMN
  • Process models
  • Activity diagrams
  • Use cases
  • Collaboration diagrams
  • Sequence diagrams
  • Class models
  • Entity / data models
  • Rules
  • Dependencies
  • State
  • Requirements
  • User stories

Measurement · a core YiENTi function

Structure it. Measure it. Learn from it.

Measurement is not a dashboard feature. YiENTi connects the execution of a process to what that execution actually produces — and turns it into machine-understandable knowledge about how work really performs.

01 · YiENTi Enterprise

Process execution

  1. Task AHigh
  2. Task BNormal
  3. Task CNormal

Priority and weight are part of the structure.

02 · Measurement

What actually happened

Work & effort

  • Priority
  • Task
  • Time / effort
  • Task weight / quantity
  • Time relative to task weight
  • Quantitative performance

Output & outcome

  • Quality
  • Rework / new follow-up work
  • Content reaching the final product
  • Reuse
  • Efficiency

People & capability

  • Roles
  • Competencies
  • Skills
  • Unused potential

03 · YiENTi AI

Reasons over measurable reality

Learn · improve · reuse

YiENTi does not only measure how long somebody worked. It can relate effort to the structure, weight, outcome, quality and reuse of the work.

LEARNING LOOPSTRUCTUREEXECUTEMEASURELEARNIMPROVEREUSE

  1. Structure

    YiENTi Enterprise

    Provides the structured operational and process context: processes, tasks, roles, priorities and the weight of the work.

  2. Reality

    Measurement

    Captures what actually happens inside that structure: effort, outcome, quality, rework and reuse.

  3. Reasoning

    YiENTi AI

    Reasons over that measurable reality — not only over descriptions of how the work is supposed to happen.

YiENTi does not learn only from descriptions of work. It can reason over structured processes and the measurable outcomes produced by them.

This connection was already part of the YiENTi architecture presented in 2016: operational processes linked to measurable output and to YiENTi AI.

Pillar A · Intelligent Enterprise Information

YiENTi Enterprise

YiENTi Enterprise connects people, processes, systems and knowledge in one structured operational environment.

An intelligent enterprise-information platform for modelling and operating the relationships between people, work, processes, systems and information. Integration is one of its capabilities — not its definition.

  • People
  • Roles
  • Competencies
  • Teams
  • Tasks
  • Projects
  • Activities
  • Processes
  • Resources
  • Systems
  • Documents
  • Communication
  • Events
  • Rules
  • Data
  • Knowledge
  • Measurements
  • Decisions
  • Dependencies
  • Enterprise workflows

Diagram A · YiENTi Enterprise Information

ORGANISATIONWORKINFORMATION & SYSTEMSPeopleRolesTeamsProjectsProcessesTasksEventsSystemsDataResourcesDocumentsKnowledge
Organisation, work and information exist as one connected model — not as separate lists, files and tools.

Pillar B · Graph-Native Intelligence Engine

YiENTi Intelligence

YiENTi Intelligence is a graph-native intelligence engine designed to reason across entities, relationships, processes, context, software structures and visual information.

Its role extends beyond the YiENTi Enterprise platform. Core YiENTi intelligence technology is already used internally within ACCENTSYS software engineering workflows; the standalone platform remains in development.

A different path toward general machine intelligence.

The long-term vision of YiENTi Intelligence.

Beyond YiENTi Enterprise

YiENTi Intelligence is not merely the AI module inside YiENTi Enterprise. It is designed as a reusable intelligence layer — and software designed with it is not automatically part of the Enterprise platform:

  • ACCENTSYS software engineeringIn internal use
  • Other ACCENTSYS productsDirection
  • Future external systemsDirection
  • Other structured reasoning use casesDirection

Diagram B · YiENTi Intelligence Graph

Structural context

  • Entities
  • Relations
  • Context
  • State
  • Rules
  • Dependencies
  • Visual Structures
  • Knowledge

YiENTi Intelligence Graph

Connected, persistent and explicit — the reasoning context itself.

Reasoning functions

  • Understand
  • Reason
  • Compare
  • Predict
  • Recommend
  • Plan
  • Orchestrate
Core reasoning functions Development direction

How YiENTi works

From structured reality to reasoning — and back.

YiENTi Enterprise represents what exists, who is involved, what happens, how work flows, what depends on what and what state the organisation is in. YiENTi Intelligence reasons over that structured context.

  1. 01Enterprise

    Represent

    People, systems, information, tasks, processes and context become structured, connected information.

  2. 02Intelligence

    Understand

    YiENTi Intelligence reasons across relationships, dependencies, state and context.

  3. 03Intelligence

    Decide

    Rules, constraints, historical context and reasoning support decisions.

  4. 04Both

    Orchestrate

    Actions, systems and workflows can be coordinated within defined rules and responsibilities.

  5. 05Enterprise

    Learn / Update

    The structured environment changes as work progresses and new information becomes available.

A continuous cycle: each updated state becomes the context for the next round of reasoning.

Structured intelligence · not LLM-first

Language is an interface.
The graph is the model.

Large language models are useful generative language components. YiENTi can use them where they help — but it does not depend on generated prose as its source of truth. Processes, relationships, states, dependencies and structures are modelled explicitly.

YiENTi represents the structure behind the information.

  • Processes are represented as processes.
  • Relationships are represented as relationships.
  • Dependencies are represented as dependencies.
  • State is represented as state.

Language is one way humans interact with that structure.

LLM-first systems derive much of their behaviour from probabilistic model inference. YiENTi's core reasoning context is an explicit structural representation of entities, relationships, processes, state and dependencies.

The question is not only which inputs a system accepts. It is whether its operational truth exists as a probabilistic generated representation — or as explicit structured state.

Structured intelligence vs LLM-first reasoning

Reasoning basis
LLM-first: Probabilistic model inference
YiENTi: Explicit structural representation
Context
LLM-first: Often reconstructed dynamically
YiENTi: Persistent, connected context
Workflow
LLM-first: May exist as description
YiENTi: Processes represented structurally
State & dependencies
LLM-first: Typically carried in prompts and context windows
YiENTi: Represented explicitly in the graph
Language
LLM-first: Excellent generative language
YiENTi: Uses language models where generation is useful
Role of the model
LLM-first: Model / provider forms a major intelligence component
YiENTi: Model-agnostic generative layer — providers are replaceable

Where language models excel

  • Text generation
  • Summarization
  • Translation
  • Natural-language interaction
  • Information extraction
  • Ideation
  • Document assistance

Where structure should stay explicit

  • Exact process state
  • Repeatability
  • Traceability
  • Dependency integrity
  • Workflow correctness
  • Software architecture consistency
  • Persistent structured memory

Probabilistic AI can assist the process. It does not have to define the process.

Generated language can be uncertain. Explicit process and graph state does not need to be reconstructed from generated prose. YiENTi is designed to keep critical process state, dependencies and operational truth outside probabilistic language generation — reducing reliance on it where structural accuracy and repeatability matter.

Model-agnostic language layer

YiENTi can use LLMs. It does not depend on them to know what the system is.

Replaceable providers

  • OpenAI models
  • Anthropic models
  • Local models
  • Open models
  • Future providers

Language layer

  • Text generation
  • Summarization
  • Translation
  • Natural-language interaction
  • Content assistance

YiENTi Intelligence architecture

Graph, state, processes, rules and dependencies — independent of the selected model provider.

Language models are optional, replaceable components that can be plugged into YiENTi where generative capabilities are useful. They are components of the architecture — not the core structural reasoning model.

Process heritage

From process understanding to machine reasoning.

YiENTi was built around the idea that complex systems cannot be understood reliably if their relationships and processes exist only as documents or descriptions.

Processes, activities, roles, events and dependencies should exist as explicit, machine-understandable structures.

  • Activities
  • Tasks
  • Events
  • Gateways
  • Conversations
  • Choreographies
  • Collaboration
  • Swimlanes
  • Data
  • Roles
  • Competencies
  • Resources
  • Projects
  • Process structures
  • Measurement
  • Information flows
ROLESYSTEMyesnoAssessComplete?ApproveRetrieve dataRecord
A modern abstraction of YiENTi's process foundation — lanes, events, activities, gateways and data as explicit structure. Illustrative, not a product interface.

Visual and structural reasoning

YiENTi does not merely describe a process. It represents the process.

Activities, dependencies, states, actors and information flows become explicit parts of the reasoning context. Visual and diagrammatic structures can be represented explicitly — and reasoned over.

  • Activity diagrams
  • Process diagrams
  • Collaboration diagrams
  • System architecture
  • Dependency graphs
  • Roles
  • Events
  • State transitions
  • Information flow

Software engineering

Software development without rebuilding context from prompts.

YiENTi maintains structural context about systems, processes and dependencies so engineering decisions can be reasoned over as part of a connected model.

Typical LLM-assisted development

  1. Prompt
  2. Generated output
  3. Review
  4. New prompt
  5. Recreated context

Context is rebuilt with every iteration.

YiENTi's engineering model

  1. Requirements
  2. Explicit structured model
  3. Systems · processes · dependencies
  4. Reasoning
  5. Implementation
  6. Validation
  7. Updated structural state

Context persists as structure and evolves with the system.

Model-driven software engineering

From model to software.

YiENTi does not guess software from prompts. It uses structured system representations, connects multiple views of the same system and reasons across them — and it can support or perform software generation from those models.

Built with YiENTi.
Not dependent on YiENTi.

YiENTi can design and generate software from structured models. The resulting application runs independently of YiENTi.

YiENTi is part of the engineering process — not a mandatory runtime dependency.

Structured modelsReasoningGeneration & validationRuns independently
  1. 01

    Requirements

  2. 02

    EPICs / User Stories

  3. 03

    Use Cases

  4. 04

    Activity / Process Models

  5. 05

    Collaboration / Sequence Models

  6. 06

    Class / Data Models

  7. 07

    Architecture / Reasoning

  8. 08

    Generated Implementation

  9. 09

    Validation

  10. 10

    Standalone Software

Conceptual chain. Read left to right, row by row.

Model layers YiENTi reasons across

Not isolated drawings, but connected projections of one underlying system model.

  • EPICs
  • User Stories
  • Use Cases
  • Activity Diagrams
  • Collaboration Diagrams
  • Sequence Diagrams
  • Class Diagrams
  • Data / Entity models
  • BPMN / process models
  • Rules
  • Dependencies
  • States

YiENTi during engineering

  • Requirements
  • Models
  • Reasoning
  • Architecture
  • Generation
  • Validation

Resulting application

  • Frontend
  • Backend
  • Database
  • APIs
  • Business logic
  • Integrations

runs independently

A generated application needs no YiENTi runtime, cloud, agent or connection in production — unless a specific product intentionally uses such services. Software created with YiENTi is also not automatically part of the YiENTi Enterprise platform.

Diagrams are not documentation after the fact

In YiENTi, models are not merely pictures that document software after it has been built. They represent the structure from which software can be reasoned about, validated and generated.

A software system can be described through complementary views — epics, user stories, use cases, activity, collaboration, sequence and class diagrams, entity/relationship models, processes, states, rules and dependencies. These are not isolated drawings. They are projections of one underlying system model, and YiENTi is designed to reason across them, so a change in one view can be analysed against the others.

Visual does not mean image recognition. A box, node, edge, gateway or relation is not merely a pixel. It has a defined meaning — and that meaning becomes part of the reasoning model.

YiENTi creates a connected representation of the system so that requirements, processes, software structures and data do not have to exist as disconnected documents.

Multiple views, one system model

  1. 01User story

    What the user needs

    “As a customer, I want to add an article to my cart so that I can order it later.”

  2. 02Use case

    What the actor wants to achieve

    CustomerSearch articleAdd to cartPlace order
  3. 03Activity

    How the process flows

    SearchSelect[in stock]Add to cartConfirm
  4. 04Collaboration

    Which components interact

    :Customer:Article:Cart:Order1.1 show()1.2 add()2 order()
  5. 05Sequence

    In which order interactions occur

    :Customer:Article:Cart:Ordersearchshowaddorder
  6. 06Class / data

    Which objects and relations exist

    Customerid · nameOrderid · dateCartitems · totalArticlesku · price1*11**
  7. 07Implementation

    Generated from the structure

    // generated from the model
class Cart {
  items: Article[]
  add(a: Article): void
  checkout(): Order
}
  8. 08Application

    Runs independently

    Cart (2)Frontend · Backend · Database · APIs

One system, many projections. “Cart” carries the same meaning in every view. Illustrative example.

Code generation is not LLM code generation

YiENTi does not need to guess the application from a prompt. It works from an explicit model of what the application is.

LLMs may assist with copy, documentation, labels, natural-language interpretation and code suggestions where useful. The software architecture does not depend solely on probabilistically generated source code.

Prompt-first generation

YiENTi model-driven

  • PromptExplicit structured model
  • AI writes some codeExplicit system relationships and reasoning
  • Context rebuilt per requestRule-based generation where appropriate
  • Source code as the only truthModel as the machine-readable source of truth

YiENTi is closer to an intelligent, compiler-like engineering environment for complex systems than to an AI dependency embedded in every application.

The flow shown here is conceptual. The exact internal pipeline may differ, and not every stage is presented as fully automated. Traceability, consistency checks, impact analysis and cross-model synchronisation describe the direction of the model-driven approach.

Model & developer tooling

YiENTi uses the right intelligence for the right task.

Generative models accelerate tasks where probabilistic generation is efficient. YiENTi retains the structured system model, process context and measurable state.

Delegated to generative models

  • UI iteration
  • Text generation
  • Wording
  • Documentation
  • Rapid generative work
  • Selected coding assistance where appropriate

Retained by YiENTi

  • Structured system model
  • Process context
  • Measurable state

  • Generative AI

    External · replaceable

    • OpenAI
    • Anthropic
    • Perplexity
    • Gemini
    • Lovable
    • Zapier
    • Pickaxe
    • Grok
    • Jasper
    • Abacus.AI
    • Relevance AI
    • Notion AI
    • And more coming soon
    • Local / alternative models

  • Local / model infrastructure

    Running models locally

    • Ollama
    • Lemonade
    • Other compatible local model workflows

  • Developer tooling

    Development environment

    • Visual Studio Code
    • OpenCode
Providers and tools can change. The YiENTi structural model remains.

These are connected systems with distinct roles — not part of YiENTi itself, and they do not all provide the same capability. Third-party names are trademarks of their respective owners; no partnership or endorsement is implied.

Testing & validation

From model to verified behavior.

Generated code is not the goal — correct behavior is. Playwright can be used to inspect whether the actual application behaves and renders as expected, creating a feedback loop between model, implementation, actual UI and validation.

  1. 01

    YiENTi

    Model & reasoning

  2. 02

    Generated / modified application

    Implementation

  3. 03

    Playwright

    Inspects the running app

  4. 04

    Visual + functional validation

    Renders and behaves as expected?

  5. 05

    Feedback

    Back into engineering iteration

Feedback returns to YiENTi and the engineering iteration until observed behavior matches the model

Playwright is a testing and validation tool — not part of the AI layer.

Automated deployment

Built with YiENTi. Not dependent on YiENTi.

YiENTi-supported workflows can automatically deploy software to AWS, Microsoft Azure and Google Cloud. YiENTi is not a hosting provider, and the deployed application does not run inside YiENTi.

YiENTi is designed to remain infrastructure-agnostic. Future deployment options are intended to include private and customer-controlled infrastructure in addition to public cloud platforms.

YiENTi engineering environment

Where software is modeled, generated, validated and released.

  1. Structured models

    BPMN, use cases, classes, data

  2. YiENTi

    Structural model & reasoning

  3. Implementation / generation

    Source code produced or modified

  4. Playwright validation

    Visual & functional checks of the running app

  5. Deployment pipeline

    Automated build and release

Environment boundary · handover of the deployable application

Application runtime environment

Where the finished application runs.

  • AWS
  • Microsoft Azure
  • Google Cloud
  • Private infrastructureFuture

Standalone application

Runs without YiENTi as a runtime dependency

Shown at provider level only. Private and customer-controlled infrastructure is a future direction, not a production-ready option today.

Product proof

Not a lab demo. Real software.

The entire current ACCENTSYS application portfolio has been built using YiENTi-supported engineering, with YiENTi itself serving as the engineering platform.

The evidence is the entire current ACCENTSYS portfolio — not a selected subset. YiENTi-supported engineering is how ACCENTSYS builds software, across every domain it works in.

YiENTiengineering core

7 proprietary ACCENTSYS applications and 1 client engineering project, all developed using YiENTi-supported engineering

Proprietary ACCENTSYS applications

Client engineering project

Developed by ACCENTSYS AG for its client using YiENTi-supported engineering. Client projects are external engineering references and are not part of the ACCENTSYS AG proprietary SaaS product portfolio.

This list is taken directly from the ACCENTSYS product registry, including its classification of proprietary applications and client projects, so new entries appear here automatically. LOXY Time is examined in detail below because its engineering effort has been measured.

Explore the portfolio

Built with YiENTi. Not dependent on YiENTi.

YiENTi supports how these products are engineered — expert-guided, not autonomous generation of every line of code. The products themselves run independently and do not require YiENTi at runtime.

Client engineering reference · LOXY Time

YiENTi-supported engineering demonstrated through real enterprise software.

LOXY Time is a real-world enterprise engineering project demonstrating YiENTi-supported software development across complex operational workflows, regulatory requirements, data processing and system integration.

LOXY Time is a transport and regulatory compliance platform developed by ACCENTSYS AG for Loxy International AG. It serves as a real-world engineering reference demonstrating the capabilities of YiENTi-supported software development. LOXY Time is not part of ACCENTSYS AG's proprietary SaaS product portfolio.

Implemented in LOXY Time

  • Digital tachograph DDD processing · Gen1, Gen2 V1, Gen2 V2
  • Driver and vehicle data
  • Regulatory checks and compliance
  • Workflows and reports
  • Integrations and bulk data processing
  • Automated quality assurance and regression testing

Internal use is not commercial availability. The standalone YiENTi platform remains in development.

YiENTi compute strategy

YiENTi runs where the compute is.

YiENTi is not tied to a specific cloud or GPU provider. It can operate wherever sufficient CPU, GPU, memory and supporting compute resources are available. Smaller workloads can run locally, while complex structural reasoning can require substantial external or dedicated compute capacity.

More available compute does not only increase speed. It allows YiENTi to work across larger connected process, graph and system contexts simultaneously.

YiENTi workload

Compute requirements

  • CPU
  • GPU
  • Memory
  • Storage / I/O

Supplied by any of

  • Local

    • High-end workstation
    • Local server

  • External compute

    • Alibaba Cloud
    • AWS
    • Microsoft Azure
    • Google Cloud
    • Other suitable providers

  • Private / dedicated

    • Private infrastructure
    • Dedicated compute
Compute sources are interchangeable. They supply capacity — they are not part of YiENTi.

Local execution

Smaller YiENTi workloads can run locally. More complex workloads can consume the available CPU, GPU and memory resources of even high-end engineering systems.

External compute today

ACCENTSYS currently rents external compute from providers including Alibaba Cloud and AWS infrastructure in the United States. Cloud vendors provide capacity. They are not architectural dependencies.

  1. Current

    Today · Flexible compute

    • Rented external compute
    • Alibaba Cloud
    • AWS
    • Other suitable providers
    • Local execution for smaller workloads

    YiENTi runs wherever the required CPU, GPU, memory and supporting resources are available.

  2. Scaling direction

    Scaling · Dedicated compute

    • Reserved / dedicated capacity
    • Private compute environments
    • Dedicated servers or clusters
    • Workload-specific infrastructure
    • Greater control over performance and cost

    Not implemented yet. The direction as YiENTi workloads grow.

  3. Long-term vision

    Long term · ACCENTSYS compute infrastructure

    • ACCENTSYS-controlled compute
    • Dedicated high-performance infrastructure
    • Potential own datacenter infrastructure

    ACCENTSYS does not operate its own datacenter today.

Microsoft Azure and Google Cloud are additional possible compute sources — not later stages. The three cards describe the infrastructure strategy; the providers are interchangeable sources of capacity within it.

Provider-independent

YiENTi is not fundamentally dependent on Alibaba Cloud, AWS, Microsoft Azure or Google Cloud.

Compute-dependent

The practical constraint is sufficient compute for the workload: CPU, GPU, system memory, VRAM, storage / I/O, network throughput and parallel compute capacity.

Commercial strategy · Vision

Long-term dedicated YiENTi infrastructure is intended to follow the commercial scaling of the ACCENTSYS product portfolio, with Merclino as a near-term commercialization priority. Successful commercialization of the portfolio is intended to help finance increasingly dedicated YiENTi compute capacity. Merclino is a commercial priority, not a special technical case: every current ACCENTSYS product is built with YiENTi-supported engineering.

Compute constraint

In current internal workloads, YiENTi's structural reasoning can require roughly an order of magnitude more compute than conventional LLM-first workflows.

YiENTi reasons over graph structures, processes, states, dependencies, measurements, multiple model views and the relationships between them — together with history and validation results. YiENTi trades compute efficiency for richer structural context.

Based on ACCENTSYS internal workloads; not an independent benchmark.

Why more compute matters

Additional compute does more than reduce processing time. It allows YiENTi to reason across larger connected contexts at the same time:

  • Processes
  • Graph relationships
  • Dependencies
  • State
  • History
  • Measurement
  • Software models
  • Multiple model views
  • Validation context

Less structure

  • Lower compute
  • Probabilistic reconstruction of context

Rich structural model

  • Higher compute
  • Explicit state, processes, relationships and measurable context
YiENTi uses substantial compute because it preserves and reasons over more of the structure of the problem.ACCENTSYS uses this architecture because structural consistency and precision matter in software engineering.

Origins

YiENTi has been evolving since 2010.

The origins of YiENTi go back to 2010, when the project began as Logical Board at Grandes Planos S.A. The technology later evolved into YiENTi, combining structured enterprise information with an increasingly sophisticated intelligence layer.

  1. 2010

    Logical Board

    Origins at Grandes Planos S.A.

    The project begins as Logical Board at Grandes Planos S.A., focused on structured enterprise information, processes, tasks and relationships.

  2. 2016

    YiENTi

    First presentation of YiENTi at Swisscom AG

    YiENTi – Intelligent Enterprise Information. The architecture presented already included:

    • Enterprise Information
    • YiENTi AI
    • Structured processes
    • Measurement
    • Graph- / process-based intelligence

  3. Today

    YiENTi Enterprise + YiENTi Intelligence

    Continuation and technical evolution of the 2016 architecture

    Already supports real ACCENTSYS software engineering using:

    • Explicit system models
    • Structured processes
    • Graph reasoning
    • Measurement
    • Model-driven engineering
    • Generation and validation

  4. Next

    Real Intelligence / AGI

    Direction of development

    Further development toward increasingly general intelligence based on structured reasoning, measurable reality and explicit system models.

    Vision / research direction

    Longer-term research direction

    General and potentially superhuman machine intelligence.

“Next” describes a direction of development, not an achieved capability. YiENTi does not claim AGI or superintelligence.

From artificial intelligence toward real intelligence

ACCENTSYS uses the term Real Intelligence for a direction beyond probabilistic language generation: intelligence that can reason over explicit processes, relationships, state, measurements and consequences.

Real Intelligence is ACCENTSYS company terminology, not an established scientific term.

Already in the 2016 architecture

The separation between enterprise information and machine intelligence was already part of the YiENTi architecture presented in 2016.

Processes connected to measurable output, a graph of enterprise information, and an AI layer reasoning over it. YiENTi's architecture was not created as a reaction to the recent rise of large language models.

Why now

The architecture preceded today's AI infrastructure.

The fundamental direction already existed in 2016. What changed is the compute available to execute it at meaningful scale. Today's substantially greater compute capacity enables increasingly complex execution of that architecture. The current implementation is not identical to the 2016 implementation; it is a substantial technical evolution of it.

Greater compute allows YiENTi to work across larger connected process, graph and software contexts simultaneously.

The architectural difference

An LLM can interpret a diagram. YiENTi can make the structure part of the model.

Processes, gateways, activities, classes, sequences, relationships and dependencies can remain explicit objects in the reasoning context.

Image interpretation

Multimodal language models can interpret text, images and diagrams. The structure is inferred from the prompt.

A picture of a process, interpreted on request.

Structural representation

In YiENTi, processes, relationships, states and dependencies can exist as explicit machine-readable structures rather than only as information inferred from prompts.

  • Activity
  • Event
  • Gateway
  • Actor
  • Sequence
  • Class
  • Relationship
  • State
  • Dependency
  • Data entity
  • Process path

YiENTi does not merely look at a diagram. The structure represented by that diagram can become part of the system model.

Multiple model views

Different diagrams become different views of one connected system model.

  • Requirements
  • User Stories
  • Use Cases
  • Activity Diagrams
  • BPMN
  • Collaboration
  • Sequence
  • Class Models
  • Data Models
  • System Dependencies

Software engineering

Not just generated code. A model of the system behind it.

YiENTi takes a fundamentally different approach to software engineering than LLM-first systems. Instead of relying on generated language to reconstruct the system context repeatedly, YiENTi maintains explicit structural representations of the system.

LLM-first loop

  1. Prompt
  2. Generated code
  3. Review
  4. New prompt
  5. Reconstructed context
  6. ↺

The system context is reconstructed from language in each cycle.

YiENTi loop

  1. Structured requirements
  2. Explicit process / system model
  3. Dependencies
  4. Reasoning
  5. Implementation
  6. Validation
  7. Measured result
  8. Updated model
  9. ↺

The system model persists and is updated with each measured result.

Product evidence

  • Arbeitsmarktvollzug
  • Castovia
  • Kicaly
  • Merclino
  • SchoolApp
  • STV Portal
  • xcess

Client engineering project

  • LOXY Time · Loxy International AG

The entire current ACCENTSYS application portfolio has been built using YiENTi-supported engineering, with YiENTi itself serving as the engineering platform. LOXY Time, a client project developed by ACCENTSYS AG for Loxy International AG, is a real-world enterprise engineering project demonstrating YiENTi-supported software development across complex operational workflows, regulatory requirements, data processing and system integration.

Architecture

Two technologies, one connected architecture.

Enterprise information becomes a structured operational graph. YiENTi Intelligence reasons over it, and the results flow back into the enterprise state or out to connected systems. The same intelligence layer also works outside YiENTi Enterprise.

Diagram C · Integration

Working together

YiENTi Enterprise

Enterprise Information Platform

Structured operational graph

What exists · who is involved · what depends on what

YiENTi Intelligence

Graph-Native Intelligence Engine

Reasoning · Decision · Orchestration

Enterprise platform

Updated enterprise state

External systems

Connected interfaces

The updated state becomes the next reasoning context.

Beyond YiENTi Enterprise

YiENTi Intelligence

Reusable intelligence layer

ACCENTSYS software engineering

Core technology in internal use

Other ACCENTSYS products

Development direction

Future external systems

Development direction

Other structured reasoning use cases

Development direction

Current development

Transparent about where YiENTi stands.

YiENTi platform

In Development

The standalone YiENTi platform is in development and is not yet commercially available.

YiENTi Intelligence · core technology

In internal use at ACCENTSYS

YiENTi already supports ACCENTSYS engineering while the standalone platform continues to evolve.

Human-in-the-loop · current maturity

YiENTi is expert-guided today.

The current system delivers substantial engineering leverage when operated by people who understand processes and system structure. Its power comes from combining structured models, process understanding and machine-assisted reasoning.

Knowledge that multiplies YiENTi today

  • BPMN
  • Process modelling
  • System architecture
  • Software engineering
  • Business workflows
  • Interpreting structured models

YiENTi is not yet a finished mass-market product that anyone can use without domain knowledge.

Direction · not a dated promise

  1. Today

    Expert-guided structural engineering

  2. Next

    More automated modelling, reasoning, testing and delivery

  3. Long term

    Increasingly autonomous system engineering

The YiENTi philosophy

From artificial intelligence toward real intelligence.

Intelligence is more than producing plausible language. It requires understanding relationships, context, processes, consequences and how information changes over time. YiENTi is ACCENTSYS's approach to this problem.

“Real intelligence” describes the YiENTi and ACCENTSYS philosophy — it is not an established scientific category.

Strategic growth

Scaling the next phase.

YiENTi is already being used to build real software. The next phase is about scaling productization, compute capacity and commercialization. YiENTi is part of the broader ACCENTSYS technology portfolio.

The limiting factors are capital, scaling and compute — not whether YiENTi works.

ACCENTSYS

Developed by ACCENTSYS AG, Switzerland.

YiENTi plays two roles at ACCENTSYS: an enabling technology layer within ACCENTSYS software engineering, and a future standalone platform.

Interested in the broader platform strategy behind YiENTi? Explore ACCENTSYS, its technology portfolio and the current strategic growth round.

For partnerships, technology discussions, strategic collaboration and investment.

Contact ACCENTSYS Explore ACCENTSYSView Investor Information