AI 模型和平台
Digitate 新增 Governed IT Agents和MCP 与ign蜂鸟融合
在 Google 上的首选来源中添加 Unite.AI
一个AI助手可以解释为什么一个商业系统失败而不解决失败。 更困难的一步是将这一解释与授权的行动联系起来，加以执行，并检查其是否奏效。 Digitate正在解决与ignio Humingbird的这一差距，该版本结合了专门的IT操作agents，定制的agent创建以及已经使用的AI工具企业的连接。
宣布 10月27日并计划于2026年11月提供，蜂鸟推出内置的模型背景协议(MCP)和Agent2Agent(A2A)服务器。 这些将ignio的操作能力暴露于其他的agents和接口，包括Microsoft Teams,Claude和Gemini，同时保留企业审批，变更管理和审计工作流程。
该版还连接了技术事件与业务流程的影响。 ignio不是将延迟的发票和中间软件故障作为单独的工作排队处理，而是旨在帮助团队了解关系，并决定首先处理什么问题。
三个具有不同级别权威的Agents
蜂鸟最初的agents涵盖SAP IDoc管理，服务台支持和补丁规划。 它们的作用不同，对企业评估它们下放了多少业务权力来说，区别很重要。
SAP IDoc agent诊断SAP数据交换中使用的失败中间文档，提出解析选项，经批准后可以重新处理。 然后核查结果。 这规定了从调查到核定行动和结果检查的顺序。
服务台agent诊断设备问题并提出解析选项，包括其优点和权衡。 补丁规划agent建议计及脆弱性风险，停工时间，变化范围和可用维护的调度windows. 这些描述以诊断和建议为中心；不应将其理解为保证每个新的agent独立执行固定。
“在信息技术业务中，AI的下一阶段不是提出更多的建议。 这份公告中，业务问题是，这项工作如何通过已经用于管理变化的控制。
MCP和A2A 将操作能力引入现有工具
内建的MCP和A2A服务器为gnio提供了一条通向更广泛的agent生态系统的路线。 MCP为AI应用程序提供工具和上下文的标准接口；A2A支持agents之间的交互。 在蜂鸟(Humingbird)中，目的是让ignio的专业知识和业务功能超越自己的接口。
对于已经在Microsoft Teams中合作或使用Claude或Gemini的团队，这可以减少在开始调查前将一个问题转移到另一个控制台的需要。 Digitate 称在团队中的合作分门别类可以将相关专家围绕共享的业务环境聚集在一起。
仅靠连通性不能确定一项行动是否得到授权。 发布称执行仍需经过批准，变更管理和审计要求。 这种分离很重要：外部接口可能要求一种能力，而不会成为改变基础设施的不受限制的路线。
让自己的Agent 将同样的想法延伸至特定组织的工作。 企业可以建设适合其系统，程序和治理要求的agents，而不是完全依赖于预建的集。 该发布没有指定通用部署时间，也没有披露每个环境所需的配置。
商业观察新增
技术警报不会自动显示其商业后果。 蜂鸟增强业务的可观察性，将订单到现金、采购到付款和记录到报告等流程与支持这些流程的应用程序、中间软件和基础设施联系起来。
进程观察为这些依赖性提供了动态视图。 目的是把有违反服务级协议风险的卡住订单、延迟发票或活动与可能的技术原因联系起来。 反复出现的失败模式可以告知早期的调查。
这基于更广泛的igno AIOps平台，该平台将监测、信息技术服务管理、发现、配置管理以及混合和多云产业区的业务信息结合起来。 该平台的现有方法将可观察性和事件相关性与自动化联系起来，而不是将生成的解释作为事件工作流程的结束。
对于业务团队来说，业务环境的价值是优先的。 影响关键过程的组件故障可能需要注意，然后才出现技术类似的问题，对业务影响不大。 排名的有用性取决于依赖性信息的准确性和新性。
补救延伸到补丁、硬化和证书
蜂鸟还拥有扩展操作风险和安全能力。 其业务风险管理解决方案整合了安全扫描仪的发现，并按业务影响对结果进行优先排序。 自动补丁和3 300多件现成的硬化解决办法旨在将这些结果与纠正工作、核查和审计文件联系起来。
证书生命周期管理新增增强发现和使用ACME协议自动更新。 过期的证书可能干扰本来可以运作的服务，使更新成为例行工作的具体例子，从监测的执行循环中受益。
该版进一步描述了采矿历史分辨率指出和企业知识文章，使用语言模型填补空白，并生成具有验证和回收保障的执行代码。 专家的反馈为这一进程提供了信息。 这些是Digitate描述的产品能力，而不是每个拟议补救办法都将取得成功的独立证据。
11月可用性将焦点放在管理下的执行上
Humingbird针对的是信息技术业务领导者、可靠性工程师、服务台以及其流程依赖于这些系统的企业所有者。 Digitate 称发布时间将于2026年11月提供；公告没有提供特定的一天或全发布性能基准测试。
发行版的中心命题是一个更相互关联的业务工作流程：找出问题，了解其业务背景，提出或授权答复，在允许的情况下执行，并核实结果。 读者可以探索Digitate的ignio平台，以了解该方法如何适应其现有的可观察性和自动化能力。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
AI Models & Platforms
Digitate Adds Governed IT Agents and MCP Integration With ignio Hummingbird
Add Unite.AI to your preferred sources on Google
An AI assistant can explain why a business system has failed without resolving the failure. The harder step is connecting that explanation to an authorised action, carrying it out and checking that it worked. Digitate is addressing that gap with ignio Hummingbird, a release that combines specialised IT operations agents, custom agent creation and connections to the AI tools enterprises already use.
Announced October 27 and planned for availability in November 2026, Hummingbird introduces built-in Model Context Protocol (MCP) and Agent2Agent (A2A) servers. These expose ignio’s operational capabilities to other agents and interfaces, including Microsoft Teams, Claude and Gemini, while retaining enterprise approval, change-management and audit workflows.
The release also connects technology events with business-process impact. Rather than treating a delayed invoice and a middleware failure as separate queues of work, ignio is intended to help teams understand the relationship and decide what to address first.
Three Agents With Different Levels of Authority
Hummingbird’s initial agents cover SAP IDoc management, service-desk support and patch planning. Their roles differ, and the distinction matters for enterprises assessing how much operational authority they are delegating.
The SAP IDoc agent diagnoses failed intermediate documents used in SAP data exchange, proposes resolution options and can reprocess them following approval. It then verifies the result. That provides a defined sequence from investigation to an approved action and outcome checking.
The service-desk agent diagnoses device issues and presents resolution options, including their advantages and trade-offs. The patch-planning agent recommends schedules that account for vulnerability risk, downtime, the scope of a change and available maintenance windows. Those descriptions centre on diagnosis and recommendations; they should not be read as a promise that every new agent independently executes a fix.
“The next phase of AI in IT operations is not about generating more recommendations. It is about getting work done,” said Ritu Dubey, Digitate’s Chief Executive, in the announcement. The operational question is how that work moves through the controls already used to manage changes.
MCP and A2A Bring Operational Capabilities Into Existing Tools
Built-in MCP and A2A servers give ignio a route into a broader agent ecosystem. MCP provides a standard interface for making tools and context available to AI applications; A2A supports interaction between agents. In Hummingbird, the purpose is to make ignio’s expertise and operational functions accessible beyond its own interface.
For a team already collaborating in Microsoft Teams or using Claude or Gemini, this could reduce the need to move a problem into another console before beginning an investigation. Digitate says collaborative triage in Teams can bring the relevant specialists together around shared operational context.
Connectivity alone does not determine whether an action is authorised. The release says execution remains subject to approval, change-management and audit requirements. That separation is important: an external interface may request a capability without becoming an unrestricted route to change infrastructure.
Make Your Own Agent extends the same idea to organisation-specific work. Enterprises can build agents suited to their systems, procedures and governance requirements rather than relying entirely on the prebuilt set. The release does not specify a universal deployment time or disclose the configuration needed for each environment.
Business Observability Adds a Reason to Prioritise
A technical alert does not automatically reveal its commercial consequence. Hummingbird enhances Business Observability to connect processes such as order-to-cash, procure-to-pay and record-to-report with the applications, middleware and infrastructure supporting them.
Process Watch provides a live view with drill-down into those dependencies. The aim is to connect stuck orders, delayed invoices or activity at risk of breaching a service-level agreement with likely technical causes. Recurring failure patterns can then inform earlier investigation.
This builds on the broader ignio AIOps platform, which combines monitoring, IT service-management, discovery, configuration-management and business information across hybrid and multi-cloud estates. The platform’s existing approach links observability and event correlation with automation, rather than treating a generated explanation as the end of an incident workflow.
For operations teams, the value of the business context is prioritisation. A component failure affecting a critical process may need attention before a technically similar issue with little immediate business impact. The usefulness of that ranking depends on the accuracy and freshness of the dependency information.
Remediation Extends to Patches, Hardening and Certificates
Hummingbird also expands operational risk and security capabilities. Its Operational Risk Management Solution consolidates findings from security scanners and prioritises them by business impact. Automated patching and more than 3,300 ready-to-use hardening fixes are intended to connect those findings to corrective work, with verification and audit documentation.
Certificate lifecycle management adds enhanced discovery and automated renewal using the ACME protocol. Expired certificates can disrupt otherwise functioning services, making renewal a concrete example of routine work that benefits from a monitored execution loop.
The release further describes mining historical resolution notes and enterprise knowledge articles, using language models to fill gaps and generating implementation code with validation and recovery safeguards. Expert feedback informs that process. These are product capabilities described by Digitate, rather than independent evidence that every proposed remedy will succeed.
November Availability Puts the Focus on Governed Execution
Hummingbird is aimed at IT operations leaders, reliability engineers, service desks and the business owners whose processes depend on those systems. Digitate says the release will be available in November 2026; the announcement does not provide a specific day or release-wide performance benchmark.
The release’s central proposition is a more connected operational workflow: identify a problem, understand its business context, propose or authorise a response, execute where permitted and verify the outcome. Readers can explore Digitate’s ignio platform for more context on how that approach fits its existing observability and automation capabilities.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.