With the Summer 2026 release, the enhanced Embedding control in Bizagi Apps opens the door to these more customized experiences. It combines HTML and JavaScript with access to the Bizagi Apps API, allowing you to create interfaces that are tailored to a particular business need while still working with information and functionality from Bizagi. 

What is the Embedding control? 

The Embedding control existed before Summer 2026 and could already be used to embed content such as websites and other external resources. With the Summer 2026 enhancements, its possibilities expand considerably. 

The control now includes a code editor with search capabilities, light and dark modes, and a live preview where you can see and interact with what you are building. More importantly, the control can access the Bizagi Apps API. It can work with Bizagi information such as cases, views, Persona collections and entity information. It can also perform actions such as creating a case or navigating to another page. 

Creating experiences that look and behave the way you need 

Suppose your users work with a collection containing customer accounts. A standard collection control may already provide everything they need. But perhaps your particular scenario would benefit from something different, such as cards that use your corporate colors, emphasize the most important information, hide sensitive information until requested, and provide buttons for the actions available for each record. 

The Embedding control gives you the flexibility to build that experience. It lets you determine colors, layouts, styles, buttons and other visual behaviors through HTML and JavaScript. This gives you additional flexibility when a particular experience needs to follow an organization's design language more closely. 

You do not necessarily need to start from scratch 

Of course, the idea of creating an experience with HTML and JavaScript might sound intimidating if you are not a developer. 

Custom does not always mean better 

With all these possibilities, there is one important consideration: you do not need the Embedding control for everything. Bizagi Apps already provides controls designed for common application experiences. If an existing control meets the business requirement, using it is usually simpler and requires less custom work. 

The Embedding control becomes particularly valuable when there is a clear reason to go beyond those standard experiences. Perhaps you need a very specific interaction, a distinctive visual presentation, or a combination of information and actions that is difficult to achieve with existing controls. 

So, before building something custom, start with a simple question: What will this experience make easier for the user? 

If there is a clear answer, the Embedding control may be exactly what you need. 

Key takeaway 

The enhanced Embedding control brings a new level of flexibility to Bizagi Apps. You can use it to transform business information into more visual experiences, keep users focused by bringing details into the same interface, combine information with actions, reflect your organization's visual identity and even create completely new ways of interacting with Bizagi data. 

But its real value is not customization for customization's sake. The best use of the Embedding control is when a tailored experience makes work simpler, clearer or more intuitive. 

Start with the user's need, use the Bizagi Apps API to connect the experience with your business information, and take advantage of AI to accelerate development where appropriate. Then test, refine and optimize the experience until the customization has a clear purpose: helping people get their work done more effectively.

In this Bizagi Pulse session, we will take a deep dive into the Embed a Website control in App Designer. We will look at how you can bring custom HTML, CSS and JavaScript into your Bizagi Apps to build richer, more interactive and data-driven user experiences. 

During this session, we will highlight the key capabilities and use cases of the Embed a Website control: 

• Custom Web Content in Your Apps: Learn how to add HTML, CSS and JavaScript directly to your app pages with the built-in code editor. It lets you search your code, expand the workspace with a live preview and add snippets to speed up development. 

• Bizagi Data and Actions through the SDK: Discover how to use the Bizagi SDK to list and retrieve information from Views, Cases and Persona Collections and run their actions. You can also navigate to internal pages, contextualized pages, or external URLs, and contextual autocompletion guides you every step of the way. 

Don't miss this opportunity to learn how to take your Bizagi Apps beyond standard components and deliver tailored experiences that drive user adoption and business value. 

Register now! 

While employees may save time drafting documents, analyzing data, or generating content, business leaders are increasingly asking a different question: Are these AI initiatives reducing costs, accelerating operations, improving customer experiences, or driving measurable business outcomes?

The difference between AI experimentation and AI-driven transformation often comes down to one critical factor: how well AI is connected to business processes.

Adoption Is Not the Same as Impact

Organizations across industries are investing in AI technologies. However, many are still struggling to move beyond isolated pilots and achieve enterprise-wide results.

The reason is simple. AI can optimize individual tasks, but businesses run on processes.

A customer onboarding journey, an insurance claim, a procurement request, or a compliance investigation involves multiple systems, decisions, approvals, documents, and stakeholders. Improving a single activity may create local efficiencies, but lasting value emerges when AI becomes part of the entire workflow.

The question is no longer, "Can AI do this task?" The question is, "How can AI improve the end-to-end process?"

Understanding the Four Levels of AI Maturity

Based on conversations with organizations across industries and regions, Bizagi has observed a common pattern in how companies adopt and scale AI capabilities. While every organization's journey is different, most initiatives tend to evolve through four stages of maturity.

These stages are not a formal framework, but rather a reflection of what we frequently see as customers move from experimentation to enterprise-wide adoption of AI.

Level 1: Assistants and Copilots

At this stage, employees use conversational AI tools to increase personal productivity.

Examples include:

• Drafting emails and reports

• Summarizing documents

• Creating presentations

• Conducting research

These use cases generate immediate productivity gains, but their impact is often limited to individual work.

Level 2: Autonomous Agents

The next step is deploying AI agents that can perform repetitive tasks with minimal human intervention.

Examples include:

• Extracting information from documents

• Classifying requests

• Validating data

• Preparing recommendations or reports

Organizations begin seeing efficiency improvements, but many initiatives remain disconnected from broader operational workflows.

Level 3: End-to-End Process Automation

This is where AI starts to generate measurable business value.

Instead of supporting a single task, AI becomes embedded within a complete business process. Information flows automatically between systems, decisions are supported by business rules, exceptions are routed to the right people, and outcomes can be measured consistently.

Level 4: Orchestration and Governance

As organizations deploy multiple AI agents and automation capabilities, coordination becomes essential.

AI agents, employees, systems, business rules, and data sources must work together within a governed framework.

This includes:

• Security controls

• Audit trails

• Decision traceability

• Compliance requirements

• Human oversight

• Performance monitoring

At this stage, organizations move beyond automation and create truly intelligent operations.

Why Orchestration Changes Everything

A useful way to think about orchestration is through the analogy of an orchestra.

Having talented musicians does not automatically produce great music. A conductor is needed to ensure everyone performs together, follows the same score, and contributes to a shared outcome.

The same principle applies to AI.

Organizations can deploy multiple assistants and agents, but without orchestration, those capabilities often remain fragmented. Information becomes siloed, governance becomes difficult, and scaling becomes increasingly complex.

Process orchestration serves as the conductor, ensuring that people, AI, applications, data, and business rules operate as a coordinated system.

What Successful AI Initiatives Have in Common

Organizations that are achieving significant returns from AI tend to share a common characteristic: they focus on complete business processes rather than isolated tasks.

Examples reported across industries include:

• Significant reductions in invoice processing times through automated document analysis.

• Faster handling of customer requests, petitions, complaints, and claims.

• Improved operational efficiency through intelligent routing and decision support.

• Shorter cycle times achieved by combining AI with workflow automation and business rules.

In each case, success comes not from implementing AI alone, but from embedding AI into a structured operational process.

The goal is not simply to make one activity faster.

The goal is to improve the outcome of the entire business process.

A Practical Checklist for Getting Started

If your organization is looking to expand its AI initiatives, consider starting with the following questions:

1. Select a High-Value Process

Identify a process that has:

• A clear beginning and end

• High transaction volume

• Significant operational cost

• Visible business impact

2. Define Success Metrics Early

Determine what you want to improve before implementation.

Examples include:

• Cycle time

• Cost per case

• Customer satisfaction

• Processing volume

• Error rates

3. Connect AI to the Full Process Context

AI performs best when it can access the information, rules, and context surrounding a business process.

Avoid treating AI as a standalone capability.

4. Establish Governance from Day One

Define:

• Who supervises AI-driven decisions

• How actions are audited

• How data is protected

• How exceptions are managed

5. Think Beyond the Pilot

View AI as a strategic investment rather than a short-term experiment.

The most successful initiatives are designed for scale, measurement, and continuous improvement from the beginning.

Building an AI-Ready Organization

In our conversations with customers, one theme consistently emerges: organizations achieve better outcomes with AI when they first establish visibility, governance, and consistency across their processes.

AI performs best when it operates within structured workflows rather than fragmented systems. By standardizing processes, connecting data sources, and orchestrating work across people, systems, and AI agents, organizations create the foundation needed to scale AI confidently and responsibly.

Becoming AI-ready is not about implementing more AI. It is about creating the operational environment where AI can continuously deliver measurable business value.

The Future Belongs to Orchestrated AI

As AI capabilities continue to evolve, competitive advantage will not come from simply having access to the latest model or the most advanced agent.

Success will belong to organizations that can effectively orchestrate people, processes, systems, and AI into a single operating model.

AI can generate insights, recommendations, and actions. Process orchestration ensures those capabilities deliver measurable business outcomes.

The organizations that move beyond pilots are not necessarily those using the most AI. They are the ones using AI where it matters most: within the business processes that drive operational performance and customer value.

Further Reading

For Spanish-speaking readers interested in exploring the ideas that inspired this article, we recommend:

• ¿Con el uso de la IA aún no ve más ingresos o ahorro de costos?: esta es la clave para ir más allá de los pilotos
  https://www.portafolio.co/negocios/empresas/con-el-uso-de-la-ia-aun-no-ve-mas-ingresos-o-ahorro-de-costos-esta-es-la-clave-para-ir-mas-alla-de-los-pilotos-503129

• La IA sin procesos se queda en un piloto; con procesos se convierte en una ventaja competitiva: Bizagi
  https://www.portafolio.co/tecnologia/la-ia-sin-procesos-se-queda-en-un-piloto-con-procesos-se-convierte-en-una-ventaja-competitiva-bizagi-495333

• ¿Funciona la IA en empresas? El debate sobre el retorno de inversión
  https://www.portafolio.co/tecnologia/funciona-la-ia-en-empresas-el-debate-sobre-el-retorno-de-inversion-495301

At first glance, Ask Ada charts may seem like just another dashboarding capability. In reality, they solve different problems. Rather than replacing traditional dashboards, they complement them. Understanding when to use Ask Ada charts and when a dedicated business intelligence tool remains the better choice will help you get the most value from both approaches. 

What are Ask Ada charts? 

Ask Ada allows users to ask questions about their process data using natural language. Instead of navigating reports, applying filters, or designing charts manually, users can simply ask a question and receive an answer. 

With Ask Ada charts, those answers can now be visualized. The system automatically generates charts based on the information requested, helping users understand patterns and trends more quickly. 

Imagine a manager asking: 

"Which regions generated the highest sales during the last quarter?" 

Instead of receiving only a text response, Ask Ada can generate a chart that immediately highlights the answer. 

This creates a more accessible experience for business users who may not be familiar with reporting tools or dashboard design. 

The biggest advantage: speed 

One of the greatest strengths of Ask Ada charts is how quickly insights can be generated. 

Traditional dashboards often require planning, report design, chart configuration, filter selection, and ongoing maintenance. Before a user can answer a question, someone typically needs to build the dashboard. 

Ask Ada works differently. The user starts with a question and receives an answer immediately. 

This is particularly useful when exploring data or investigating unexpected situations. The user does not need to know whether a report already exists. They simply ask. 

For example, a process owner might wonder why approval times increased during the previous month. Instead of requesting a new report, they can ask Ask Ada directly and visualize the results within seconds. 

When Ask Ada charts are the right choice 

Ask Ada charts are ideal when users need quick answers to changing business questions. 

Many organizations face situations where decision makers continuously ask new questions that were not anticipated when dashboards were designed. Building a new dashboard for every question is inefficient and often slows down analysis. 

This is where Ask Ada shines. 

It is particularly valuable for managers, team leaders, and business users who want to investigate process performance without depending on technical teams or reporting specialists. 

Ask Ada charts are also a great fit for exploratory analysis. Users can start with a question, review the answer, ask a follow-up question, and continue refining their understanding of the data. The experience feels much more conversational than traditional reporting. 

Another advantage is accessibility. Employees who are unfamiliar with complex reporting tools can still obtain meaningful insights simply by describing what they want to know. 

When traditional dashboards are still the better option 

Although Ask Ada charts provide tremendous flexibility, they are not designed to replace every reporting scenario. 

Traditional dashboards remain the preferred choice when organizations need highly structured, standardized reporting. 

For example, executive dashboards that are reviewed every week or month often require carefully designed layouts, consistent metrics, and precise formatting. These dashboards are intended to provide the same view every time so that performance can be monitored consistently. 

Business intelligence platforms also continue to offer advantages when working with very large datasets, advanced calculations, complex data models, or cross-system reporting. Organizations frequently use these tools to combine information from multiple applications and create enterprise-wide analytics solutions. 

If a dashboard is considered a critical business asset and must be shared across large audiences with strict governance controls, a dedicated BI solution may still be the best approach. 

The best approach is often both 

The most effective organizations will not choose between Ask Ada charts and traditional dashboards. They will use both. 

Think of traditional dashboards as your organization's official scorecards. They provide standardized metrics that everyone trusts and reviews on a regular basis. 

Think of Ask Ada charts as your exploration tool. They help users investigate, ask follow-up questions, and discover insights that may not yet exist in a formal dashboard. 

A manager might begin by reviewing a monthly operations dashboard. After noticing an unexpected increase in processing times, they could use Ask Ada to investigate further, generating additional charts and asking questions that were never included in the original report design. 

Together, these capabilities create a more complete analytics experience. 

A practical recommendation 

If your users frequently request new reports, ask ad hoc questions, or depend on analysts to build simple visualizations, Ask Ada charts can dramatically reduce the time required to obtain insights. 

At the same time, continue using traditional dashboards for executive reporting, governance, regulatory requirements, and highly structured analytics. 

The goal is not to replace existing reporting investments. Instead, it is to empower users with a faster and more intuitive way to explore their data. 

Ask Ada charts and traditional dashboards serve different purposes. Ask Ada excels at answering questions quickly, supporting data exploration, and helping business users uncover insights through natural language. Traditional dashboards remain the best choice for standardized reporting, enterprise analytics, and long-term performance monitoring. 

By understanding the strengths of each approach, organizations can create a reporting strategy that combines the flexibility of conversational AI with the reliability of well-established business intelligence practices.

At first glance, selecting a model may seem like a technical detail. In practice, however, the model you choose can influence how quickly an agent responds, how deeply it analyzes information, and how many AI resources it consumes. Understanding the difference between reasoning and non-reasoning models will help you design agents that are both effective and cost-efficient. 

Understanding reasoning and non-reasoning models 

Not all AI models work in the same way. Some are designed to provide answers quickly based on patterns, while others spend more time analyzing information before responding. 

Non-reasoning models are optimized for direct tasks. They shine at pattern recognition, summarization, content generation, classification, and other scenarios where the answer can be produced without a deep analytical process. These models are typically faster and consume fewer resources, making them a great choice for routine tasks. 

Reasoning models, on the other hand, are designed to spend additional effort evaluating information before generating a response. Rather than immediately producing an answer, they perform a more deliberate analysis of the request. This makes them useful for situations that involve multiple variables, complex decision-making, comparisons, or detailed analysis. 

Think of it this way. If you ask an agent to summarize a document, a non-reasoning model will likely be the most efficient option. If you ask an agent to evaluate several supplier proposals, compare them against multiple criteria, and justify a recommendation, a reasoning model will typically deliver better results. 

Understanding reasoning levels 

For reasoning models, Bizagi introduces reasoning levels that allow you to control how much analytical effort the model should apply to a task. 

The available options are None, Low, Medium, and High. 

Each level influences how deeply the model evaluates information before responding. 

When to use None 

The None level is appropriate when you want the reasoning model to provide a direct answer without engaging in significant analysis. This setting works well for straightforward tasks such as formatting content, rewriting text, extracting information from a structured source, or producing basic summaries. 

For example, if an agent receives a customer comment and needs to classify it into a category, a None reasoning level may be perfectly sufficient. Because the model performs minimal reasoning, responses are generally faster and less resource intensive. 

When to use Low 

The Low level introduces a moderate amount of analysis while maintaining good response times. This option is useful when the agent needs to evaluate simple business rules or perform light comparisons. 

Imagine an agent that reviews several support tickets and suggests priority levels based on predefined criteria. The task requires some thinking, but not extensive analysis. In this case, Low reasoning often provides the right balance between performance and quality. 

When to use Medium 

Medium is likely to become the default choice for many business scenarios. It provides a balanced approach that combines analytical reasoning with reasonable performance. 

Consider an agent that reviews an employee request, analyzes supporting information, checks policy requirements, and recommends an approval decision. The task involves multiple variables and requires contextual understanding. Medium reasoning gives the model enough analytical capability without introducing unnecessary complexity. 

When to use High 

High reasoning should be reserved for situations where deeper analysis creates measurable value. 

Examples include complex financial evaluations, multi-criteria supplier assessments, technical investigations, risk analysis, or mathematical calculations. 

Imagine an agent responsible for evaluating several vendor proposals. The agent must compare pricing, technical capabilities, service quality, implementation timelines, and previous performance. In this situation, a High reasoning level allows the model to spend more effort analyzing the information and producing a more robust recommendation. 

Start simple and test 

A common mistake is assuming that more reasoning automatically produces better results. In reality, the best configuration depends on the specific use case. 

The best approach is to test different configurations using the Test prompt feature in Bizagi Studio. Try the same prompt with different models and reasoning levels. Compare the responses, review their quality, and evaluate whether the additional reasoning improves the result. 

You may discover that a Medium reasoning level provides the same business value as High reasoning while consuming fewer resources. Or you may find that a non-reasoning model is perfectly adequate for a task that originally seemed more complex. 

During this session, we will highlight key innovations introduced in this release: 

• Next-Generation AI Agents: Take advantage of new AI models, advanced reasoning capabilities, and expanded execution options that make AI Agents smarter, faster, and better equipped to handle complex business scenarios with greater precision and flexibility. 

• AI Monitoring and Consumption Analytics: Gain complete visibility into AI adoption, performance, dependencies, and consumption across your organization. With centralized monitoring, AI metrics, and consumption reporting, leaders can confidently scale AI while maintaining control of usage and investment. 

• Improvements in Real-Time Event Streaming with Kafka Integration: Connect Bizagi to your enterprise ecosystem through Apache Kafka, enabling real-time event-driven architectures. Automatically react to business events, trigger processes, create cases or records, and streamline integrations across systems with greater flexibility, scalability, and agility. 

Don’t miss this opportunity to discover how Bizagi Summer 2026 empowers organizations to move beyond experimentation and operationalize AI at scale—focusing intelligence where it creates the greatest business value. 

Register now!