INTEGRATE 2025 Day 1 Highlights

#1: Keynote

In a latest keynote handle, Microsoft executives Balan and Slava shared their insights on the transformative energy of integration companies and generative AI. The session highlighted the essential position of integration in digital transformation and the way AI is reshaping the operational panorama for organizations. Under is a complete abstract of the details coated by each audio system.

Balan’s Key Factors

Introduction

• Balan opened the keynote by expressing pleasure concerning the developments in integration companies since his final look in 2022.
• He famous that whereas the technological panorama has advanced considerably, significantly with the rise of generative AI.

The Actuality of Generative AI

• Generative AI is not a hype; it’s a actuality that’s actively reshaping industries.
• AI is making beforehand unimaginable duties possible, from pure language processing to visible understanding.
• Organizations are more and more viewing AI as a strategic differentiator reasonably than a value middle, resulting in a surge in progressive use circumstances throughout varied sectors.

New Use Instances and Purposes

Balan highlighted a number of thrilling functions of AI, together with:

• Buyer Expertise Personalization: Retailers are leveraging AI to boost buyer interactions.
• Drug Discovery: AI is being utilized to streamline processes within the pharmaceutical business.
• Insurance coverage Trade Improvements: AI brokers have gotten extra prevalent in managing insurance coverage claims and buyer inquiries.

The Significance of Integration

• Integration is changing into the spine of recent functions.
• Simply as organizations as soon as wanted a web site or a cell app, they now require AI capabilities to stay aggressive.
• The mixing of AI into present functions is crucial for maximizing effectivity and effectiveness.

Traits of Clever Purposes

Intelligence is the brand new baseline for Trendy apps and Balan outlined two important traits of clever functions:

• Ease of Data Retrieval: Clever functions ought to simplify the method of discovering related info.
• Information Surfacing: They need to leverage AI to current the appropriate information on the proper time, enhancing decision-making processes.

The Function of AI in Enterprise Processes

• AI isn’t just for customer-facing functions; it is usually getting used internally to enhance worker productiveness.
• Balan shared examples of how Microsoft staff are using generative AI for market analysis, content material creation, and product utilization evaluation.

What’s Coming Subsequent: The Rise of Brokers in Integration

As we glance forward, the introduction of brokers is changing into more and more vital for organizations, many organizations are exploring how brokers could be launched in each customer-facing processes and inside operations. This pattern emphasizes the significance for integration builders to contemplate how brokers will impression integrations and the way integrations will assist the creation of recent brokers.

Challenges and Alternatives

• Regardless of developments, many generative AI functions are nonetheless not shifting into manufacturing.
• Agentic AI growth challenges resembling mannequin choice, advanced workflows, content material security, observability & governance, and power sprawl stay vital hurdles.
• Nonetheless, the potential for AI to drive effectivity and innovation is immense.

Understanding the Distinction: Brokers vs. Instruments

A key distinction between an agent and a software lies of their use of huge language fashions (LLMs). Whereas instruments can make the most of LLMs for duties like summarization, they don’t interact in planning, which is a elementary attribute of brokers.

This distinction could appear daunting, however the considerate integration of brokers into enterprise processes is crucial. LLMs can now rework prolonged workflows into pure language directions for planning and execution. Brokers execute these plans utilizing varied instruments, and the Mannequin Context Protocol (MCP) performs a job on this course of.

Moreover, brokers can obtain inputs from finish customers in conversational contexts and may talk with one another via a two-way protocol supported by Azure. This implies you don’t have to fret about managing these protocols, as Azure handles the complexities, whether or not you’re creating conversational chatbots or autonomous programs.

Azure AI Foundry

Azure AI Foundry is a platform that gives a versatile vary of AI fashions — from OpenAI to light-weight fashions — which could be hosted serverlessly or on devoted compute.

It consists of an agent service that allows sturdy communication between AI brokers, helps testing, and helps deploy them into manufacturing.

Foundry additionally affords specialised companies like speech translation (used efficiently in real-world situations), and robust observability, giving full visibility into each agentic workflows and broader integration workflows (like these constructed with Logic Apps).

The platform is deeply built-in with Azure Integration Companies, making it ideally suited for constructing clever, end-to-end automated options

Highlights from Microsoft Construct

• Logic Apps + Azure AI Search: Native integration now permits real-time, automated updates to vector retailer indices as supply information modifications.
• Software Calling from AI Foundry: Brokers in Azure AI Foundry can now invoke Logic Apps and Azure Capabilities — enabling richer AI-driven workflows.
• Serverless GPU Internet hosting: Experiment with generative AI with out ready for GPU capability, utilizing ACA to host fashions serverlessly for POCs.
• Superior API Administration Use Instances:
  1. Load balancing throughout fashions (throughout areas or clouds)
  2. Utilization monitoring throughout apps and brokers
  3. Actual-world instance: A Korean telecom constructed a ChatGPT-like B2B platform utilizing APIM to watch and management utilization.
• Semantic Caching in APIM: Reduces pointless mannequin calls utilizing fuzzy logic, optimizing mannequin value and response time.
• Content material Security by way of APIM: Implement AI content material moderation throughout apps with centralized filtering insurance policies.
• Steady Suggestions Loop for AI Tuning: Log prompts and responses by way of APIM to feed again into Foundry and fine-tune fashions.
• MCP (Mannequin Context Protocol): Permits brokers and instruments to change metadata/context. Supported natively in APIM and Azure Capabilities.
• Serverless APIM for Generative AI: Light-weight, cost-efficient setup for managing APIs and MCP servers.
• Credential Administration Enhancements: On-behalf-of auth and fine-grained operation-level entry management for MCP.
• Organizational Registries in API Middle: Centralized MCP administration for large-scale enterprise adoption.
• MCP + Dev Instruments: MCP servers are actually simply usable in GitHub Copilot, VS Code, and Copilot Studio

The Way forward for Integration Companies

Balan emphasised the necessity for organizations to spend money on three key areas:

1. Understanding Use Instances: Clearly defining the enterprise metrics that AI will impression.
2. Modernizing Software Infrastructure: Addressing information silos and disconnected workflows to facilitate AI integration.
3. Establishing Sturdy Operations: Making a suggestions loop to constantly enhance AI functions primarily based on consumer interactions.

Conclusion

• Balan concluded with a name to motion for organizations to embrace the combination of AI into their workflows.
• He expressed confidence that the way forward for integration companies, powered by AI, will allow companies to function extra effectively and successfully.

Slava’s Key Factors

Introduction and Context

• Slava started by acknowledging the thrill surrounding the developments in AI and integration companies.
• He emphasised that enterprise automation isn’t just about responding to messages however coordinating a number of steps and programs to realize particular outcomes.

Agentic Workflows

• Slava launched the idea of agentic workflows, which can contain one or a number of brokers working sequentially or in parallel.
• He defined that some workflows could contain human interplay, whereas others run absolutely autonomously.

Logic Apps and Agent Loop

• Slava mentioned how Logic Apps present the management, integration, and governance wanted to construct real-life workflows that incorporate agentic capabilities.
• He launched the Agent Loop, a brand new characteristic in Logic Apps that enables customers to construct autonomous or conversational brokers immediately inside the Logic Apps customary.
• Customers can carry their fashions, system prompts, and instruments, enabling brokers to run natively as a part of the workflow.

Conversational Brokers

• Slava highlighted that brokers cannot solely assume but additionally talk, because of the versatile idea of channels that enable for real-time, interactive conversations throughout platforms like Groups and Slack.
• The chat testing capabilities embedded in Logic Apps Designer allow customers to simulate and debug conversations with out leaving the workflow context.

Multi-Agent Orchestration

• The Agent Loop helps multi-agent orchestration, permitting for varied execution patterns, together with sequential and concurrent execution, loops, branching, and handoffs between brokers.
• This functionality allows organizations to combine agentic steps with deterministic steps in the identical workflow, offering a complete orchestration resolution.

Integration with AI Foundry

• Slava emphasised the deep integration between Agent Loop and AI Foundry, permitting customers to construct brokers that run inside Logic Apps whereas leveraging the capabilities of AI Foundry.
• Brokers registered with AI Foundry can entry fashions and benchmarking for ongoing high quality checks, combine with accountable AI companies, and make the most of shared data and grounding instruments.

Adoption and Use Instances

• Slava shared that for the reason that public preview of Agent Loop, a whole bunch of consumers throughout varied industries have already constructed and run over 1,500 brokers, producing vital utilization of AI companies.
• He famous the number of use circumstances, together with challenge estimation, information mapping, ticket dealing with, and even game-playing brokers.

Code Full Workflows

Slava introduced the introduction of Code Full Workflows in Logic Apps, which mixes the ability of conventional coding with the simplicity of Logic Apps. This new authoring mannequin permits builders to put in writing workflows utilizing C# whereas nonetheless benefiting from the operational reliability of the Logic Apps platform.

Way forward for Integration

Slava emphasised that because the demand for integration grows, Azure Integration Companies should evolve to fulfill the wants of organizations. With the projected improve in using brokers, a robust integration spine will probably be important for fulfillment.

Conclusion

• Slava concluded by encouraging attendees to discover the capabilities of Agent Loop and Logic Apps, emphasizing that these instruments present a quick and simple method to construct brokers which can be grounded in real-time information and enterprise functions.
• He invited individuals to affix devoted classes to be taught extra about these new capabilities.

Remaining Ideas

The keynote by Balan and Slava offered a complete overview of the present state and way forward for integration companies and generative AI at Microsoft. Their insights highlighted the significance of embracing AI and integration to drive effectivity, innovation, and aggressive benefit in as we speak’s quickly evolving digital panorama. Organizations that prioritize these areas will probably be well-positioned to thrive sooner or later.

#2: What’s new in Azure Logic Apps

The session “What’s the Way forward for Logic Apps?” on the INTEGRATE occasion featured key insights from Microsoft integration consultants, together with Kent Weare, Parth Shah, and Divya Swankar. Right here’s a abstract of the details mentioned:

1. Funding Focus: The way forward for Azure Logic Apps will concentrate on three important areas:
   • Reshaping Enterprise Processes: Leveraging AI to boost enterprise processes.
   • Optimizing Operations: Bettering operational effectivity via integration environments and enterprise course of monitoring.
   • Enhancing Developer Experiences: Offering higher instruments and experiences for builders.
2. New Capabilities:
   • Agent Loop: A brand new characteristic permitting customers to construct brokers in Azure Logic Apps, integrating AI fashions to boost enterprise course of orchestration.
   • Agent Parameters: This characteristic simplifies workflow creation by permitting dynamic inhabitants of parameters utilizing giant language fashions.
   • SRE Agent: A proactive agent that identifies points and assists in triaging and resolving operational duties.
   • Code for Workflows: A brand new functionality that enables builders to put in writing code inside Logic Apps, offering extra management over advanced logic.
3. Information Mapper Enhancements: The info mapper has been improved to facilitate simpler information transformations, with a concentrate on efficiency and reliability. It now helps customized XPath and XSLT features.
4. Hybrid Logic Apps: The hybrid mannequin permits for native processing and storage, catering to prospects who require native management over their information and processes.
5. Organizational Templates: A brand new characteristic enabling organizations to create and handle customized templates for workflows, enhancing growth productiveness and standardization.
6. AI Integration: The session emphasised the combination of AI capabilities, together with immediate templates for Azure OpenAI, permitting for extra dynamic and context-aware interactions.
7. Suggestions and Future Plans: The workforce inspired suggestions from customers relating to the consumption mannequin of Logic Apps, in addition to the longer term route of the platform, significantly in regards to the steadiness between consumption and customary fashions.
8. Q&A Session: The session concluded with a Q&A section the place attendees raised questions on particular options, together with the information mapper’s schema necessities and the singleton design sample with out service bus dependencies.

General, the session highlighted Microsoft’s dedication to enhancing Azure Logic Apps via AI integration, improved developer experiences, and strong operational capabilities, whereas additionally in search of consumer suggestions to information future developments.

#3: How are we going to assist all of this AI and Integration stuff?

In his session at INTEGRATE 2025, Michael Stephenson(aka Mike) – The product proprietor for Turbo360 and an Azure MVP emphasised the pivotal position of Turbo360 in enhancing AI and integration functions. He addressed the rising demand for efficient assist mechanisms as organizations quickly develop these options. The software simplifies the assist expertise, enabling builders to concentrate on innovation reasonably than upkeep.

Mike showcased key options of Turbo360, together with Enterprise Exercise Monitoring (BAM) and Price analyzer, which empower organizations to handle their integration options and optimize Azure prices extra successfully. BAM permits non-technical stakeholders to trace and analyze integration efficiency, fostering collaboration between technical and enterprise groups.

He additionally highlighted the significance of simplifying the cloud working mannequin by offering a unified view of varied cloud companies, which reduces complexity for builders and designers. This single-pane strategy streamlines monitoring and administration of integrations, making it simpler for groups to navigate the complexities of recent expertise.

Finally, Mike positioned Turbo360 as a necessary software for organizations dealing with integration challenges, emphasizing its potential to drive effectivity and ship enterprise worth. His insights underscored the significance of Turbo360 in serving to organizations adapt to the fast-changing technological panorama whereas maximizing their integration capabilities.

#4: Reworking BizTalk: A journey to fashionable integration

At INTEGRATE 2025, Harold Campos, Principal Program Supervisor for Azure Integration Companies, offered a compelling session on modernizing BizTalk Server options utilizing Azure Logic Apps.

The Evolution of Integration

Harold started by discussing the numerous modifications in integration over the previous few many years. He outlined the development from file-based interfaces to Service-Oriented Structure (SOA), after which to Enterprise Service Bus (ESB) and Integration Platform as a Service (iPaaS) options. This evolution displays the rising complexity and calls for of recent enterprise environments, necessitating extra agile and scalable integration options.

As a long-standing resolution, BizTalk Server has successfully addressed many conventional integration wants. Nonetheless, with the introduction of Azure Logic Apps roughly eight years in the past, Microsoft has positioned Logic Apps because the successor to BizTalk. Harold emphasised that Logic Apps is designed to be the quickest and most cost-effective choice for purchasers migrating from BizTalk Server.

Migration Assist and Timeline

Harold highlighted that BizTalk Server 2020 will proceed to obtain mainstream assist via 2028, giving prospects an extended runway for migration. He described a phased migration technique—rehost, replatform, and refactor—tailor-made to organizational readiness.

Key Options of Azure Logic Apps

He emphasised the ability of Azure Logic Apps as a contemporary platform, with assist for wealthy XML integration, customized .NET code, and a hybrid internet hosting mannequin that runs throughout on-prem and cloud environments. He additionally showcased its deep integration with Azure Monitor and Software Insights, enabling higher observability and diagnostics.

Harold closed by encouraging the neighborhood to share suggestions and assist form the way forward for Logic Apps, reinforcing Microsoft’s dedication to supporting fashionable integration wants.

#5: Actual-time occasion streaming ingestion and processing with Material

Introduction

Kevin begins by expressing his enthusiasm for the convention and the neighborhood, sharing his long-standing reference to the occasion. He introduces his position overseeing Azure Messaging, which encompasses varied companies like Service Bus, Occasion Hubs, Occasion Grid, and Stream Analytics.

Challenges in Actual-Time Information Processing

Kevin outlines the challenges organizations face in adapting to the speedy inflow of real-time information and AI. He highlights the restrictions of conventional batch processing programs, which are sometimes sluggish and cumbersome. The necessity for organizations to react rapidly to incoming information from varied sources—resembling manufacturing gadgets, functions, and automobiles—is emphasised. He notes the problem to find related insights amidst huge quantities of information and the fragmentation of present tech stacks, which require experience in a number of applied sciences and result in information silos.

Microsoft Material Overview

To handle these challenges, Kevin introduces Microsoft Material, a unified platform designed to streamline information acquisition and analytics. He emphasizes the purpose of shifting from a fragmented tech stack to a single, cohesive mannequin that democratizes entry to information analytics throughout organizations. Key options of Material embody:

• Actual-Time Streaming: Transitioning from quicker batch processing to real-time information ingestion and response.
• Unified Information State: Eliminating information silos and making a single supply of fact for information.
• AI-Powered Insights: Using AI to establish anomalies and insights inside the information.

Parts of Actual-Time Intelligence

Kevin discusses the elements of real-time intelligence inside Microsoft Material, which embody:

1. Occasion Ingestion: Connecting to varied information sources for real-time information streaming.
2. Actual-Time Analytics: Leveraging Azure Occasion Hubs, Occasion Grid, and Stream Analytics for analytics at scale.
3. Digital Twin Builder: Creating digital representations of bodily belongings for higher monitoring and evaluation.
4. Actual-Time Dashboards: Interactive dashboards for visualizing and analyzing streaming information.
5. Guidelines and Actions: Automating responses primarily based on predefined guidelines triggered by incoming information.

Demonstration of Actual-Time Intelligence

Alicia Li takes the stage to showcase new options in Microsoft Material, specializing in the Occasion Stream capabilities. She introduces a number of key options:

• SQL Operator: Permits customers to precise enterprise logic utilizing acquainted SQL syntax, enhancing productiveness and enabling extra superior situations.
• Schema Inferencing: Robotically infers the schema of incoming information, permitting for a number of occasion sorts inside a single occasion stream.
• Occasion Schema Set: Permits customers to outline and handle schemas explicitly for higher information high quality and governance.

Alicia demonstrates how you can construct a sensible constructing setting monitoring system utilizing real-time information from temperature and energy utilization sensors. The demo highlights:

1. Setting Up Occasion Streams: Connecting to an MQTT dealer to ingest sensor information.
2. Information Transformation: Utilizing filters and managing fields to arrange information for evaluation.
3. Activating Alerts: Establishing guidelines to set off alerts when sure situations are met (e.g., temperature thresholds).
4. Constructing Dashboards: Creating real-time dashboards to visualise information and monitor situations.

Future Roadmap

Kevin wraps up the session by discussing the roadmap for Microsoft Material, which incorporates:

• Enhanced Integration: Improved connections with varied information sources, together with Azure SQL and Microsoft 365 Graph information.
• Customized Enterprise Occasions: Permitting customers to outline their very own occasions primarily based on processing outcomes.
• Anomaly Detection: Using AI to establish anomalies in real-time information streams.
• Material Capabilities: Integrating Azure Capabilities into Material for enhanced event-driven structure.

#6: Constructing agentic workflows utilizing Azure Logic Apps

The session options Kent Weare and Divya Swarnkar from Microsoft, who talk about the evolution of automation via the introduction of AI brokers, significantly specializing in the “agent loop” idea inside Logic Apps.

Introduction

Kent opens the session by emphasizing the transformative potential of AI in automation and integration architectures. He outlines the historic evolution of integration strategies, from file-based interfaces to fashionable iPaaS programs, and highlights the latest emergence of AI and brokers as a brand new paradigm on this house.

Logic Apps Technique

Kent describes the Logic Apps technique, which focuses on three core areas:

1. Reshaping Enterprise Processes: Using brokers to boost and automate workflows.
2. Optimizing Operations: Streamlining processes via automation.
3. Enhancing Developer Experiences: Making it simpler for builders to create and handle integrations.

The Function of AI Brokers

Kent explains that AI brokers characterize an evolution in integration structure, permitting for extra dynamic and adaptable workflows. He emphasizes that brokers should not meant to switch present programs however to enrich and improve them. The structure of an AI agent consists of:

• Instruments: Connectors, workflows, APIs, and customized code.
• Data: Contextual info that helps brokers make knowledgeable choices.
• Autonomy: Brokers can function primarily based on occasions or schedules, reasonably than following a static path.

Buyer Success Tales

Kent shares examples of how firms like AT&T and DocuSign are leveraging AI to enhance productiveness and streamline operations. As an illustration, DocuSign makes use of Azure Logic Apps to orchestrate their clever settlement administration system, leading to vital time financial savings for his or her purchasers.

Demonstration of Agent Loop

Divya takes the stage to show the agent loop in motion, specializing in a recruitment agent designed to help recruiters with duties resembling candidate screening, interview scheduling, and workforce notifications. The demo showcases:

• Creating an Agent Workflow: Utilizing Logic Apps to arrange an agent workflow with triggers and actions.
• Configuring the Agent: Defining the agent’s persona, objectives, and limits via system directions.
• Dynamic Connections: Permitting brokers to make use of user-specific identities for actions like scheduling conferences.

Observability and Transparency

Divya highlights the significance of observability in agent operations, showcasing how customers can monitor agent choices, interactions, and power utilization via run historical past and agent chat panels. This transparency helps customers perceive the agent’s actions and choices.

Mindset Shift

Kent discusses the shift in mindset required for adopting AI brokers, shifting from deterministic workflows to extra versatile, objective-driven approaches. He emphasizes the significance of specializing in outcomes reasonably than predefined paths, permitting for higher adaptability in enterprise processes.

Multi-Agent Eventualities

The session additionally covers multi-agent orchestration patterns, demonstrating how completely different brokers can work collectively to perform advanced duties. Divya presents examples of a journey agent state of affairs and a gross sales report technology state of affairs, illustrating how brokers can hand off duties and collaborate successfully.

Conclusion

Kent concludes by encouraging attendees to discover the potential of AI brokers inside their organizations, emphasizing that Logic Apps already present a strong platform for constructing these options. He invitations individuals to have interaction with the broader Microsoft AI ecosystem and to attend future classes for extra insights.

General, the session highlights the transformative potential of AI brokers in automation, showcasing sensible functions and inspiring a shift in how organizations strategy integration and course of automation.

#7: Messaging for the enterprise

The ultimate session of the day featured Christina Compy, Principal PM Supervisor, and Eldert Grootenboer, Senior Product Supervisor from Microsoft, who mentioned developments in Azure messaging companies. They highlighted 4 important elements of Azure’s messaging ecosystem:

1. Event Grid: A discrete message dealer supporting varied messaging patterns and occasion dealing with throughout Azure.
2. Service Bus: An enterprise queue service that predates Azure, providing strong messaging capabilities.
3. Occasion Hubs: A high-throughput streaming platform used for enormous information masses, together with inside Azure companies.
4. Stream Analytics (ASA): A foundational service for processing information, significantly billing occasions.

The audio system emphasised Azure’s reliability, boasting a five-nines SLA and dealing with 377 trillion requests month-to-month. Additionally they launched Material, a brand new SaaS layer constructed on present Azure companies, aimed toward simplifying occasion stream administration.

Key updates included infrastructure enhancements centered on safety and resilience, resembling zone redundancy for brand spanking new deployments and modernization efforts to boost efficiency and safety.

Eldert outlined the roadmap for upcoming options, together with:

• Geo-application basic availability (GA) for multi-region information dealing with.
• Batch delete capabilities for environment friendly message administration.
• Sturdy terminals for connection restoration.
• Enhanced migration choices between service tiers and areas.

Additionally they mentioned geo-replication, permitting for information and metadata replication throughout areas with minimal shopper modifications required. The session concluded with a concentrate on the significance of normal testing and automation in failover processes, in addition to the introduction of cross-namespace forwarding to boost message routing flexibility.

The audio system inspired viewers engagement and offered sources for additional studying, whereas additionally inviting attendees to achieve out with questions.