As organizations race to move AI from experimentation to production, fragmented infrastructure remains a documented barrier with disconnected tools, siloed teams, and spiraling costs making governance and control of the ecosystem impossible at scale. This integration transforms how developers discover, inspect, debug, and automate API workflows, addressing the three most common sources of friction: API discovery and access, governance and consistency, and the growing need to support AI-driven workflows.

“AI is only as effective as the context it has access to,” said Marco Palladino, CTO and Co-Founder of Kong. “By integrating Insomnia with Kong Konnect, we’re giving both developers and agents instant access to an organization’s entire API ecosystem, resulting in a faster, more seamless experience for developers and a stronger foundation for agentic software development.”

Also Read: CIQ Expands Fuzzball to Full Multi-Cloud so AI and HPC Teams Ship Faster and Spend Less

Core Capabilities for Enterprise API and AI Teams

The integration delivers several key capabilities designed to streamline enterprise workflows:

• Unified Access via Konnect Sync: Developers no longer need to hunt through outdated documentation or manually import API specs. By authenticating with their Kong Konnect Personal Access Token (PAT), developers gain instant access to all available API and AI endpoints directly within Insomnia. This enables developers to seamlessly pull environment configurations from Konnect into Insomnia with a single click.
• A Single Source of Truth: When platform teams update API specifications in Konnect, those changes automatically reflect in Insomnia. This eliminates version mismatches, slashes developer onboarding time, and ensures teams are always testing against approved definitions.
• AI Agent API Access via Insomnia CLI: Coming to Tech Preview – The CLI ensures that large language models (LLMs) and autonomous agents can access the collections, environments, and configurations, directly inside Insomnia without approximation or workarounds. Enterprises can now automate, orchestrate, and extend API testing using AI without workflow bottlenecks. The CLI outputs structured JSON responses specifically optimized for parsing by LLMs, bridging the gap between human-readable interfaces and agentic automation.
• Native Git-Powered Version Control: In tandem with the CLI, collections and configurations can now be managed completely via Git from the command line empowering developers to work from their chosen CLI tool.

Source: PRNewswire

The post Kong Announces Insomnia and Kong Konnect Integration to Unify API and AI Development Workflows appeared first on ITDigest.

AI is everywhere now. Customer support teams use it. Marketing teams use it. Security teams use it. Leadership teams are pushing forward AI initiatives because nobody really wants to be the company that gets left behind, right. The whole rush feels understandable, even if it’s a bit frantic. The part that gets messy is governance, because that’s not moving at the same speed.

Most organizations have spent years saying things about fairness transparency, and accountability. But talking and actually doing, are two totally different animals. The gap is bigger than a lot of leaders are imagining, and it shows up fast. The World Economic Forum says less than 1% of organizations have fully operationalized responsible AI. You’d think that number would make every executive feel pretty uneasy, and not just slightly. AI adoption is scaling. Responsible AI practices are not.

That is why creating responsible AI development frameworks has become a business priority, not a compliance exercise. The goal is simple. Build AI systems that people can trust, regulators can understand, and organizations can manage without creating unnecessary risk.

Ethical AI vs Responsible AI

A lot of people treat ethical AI and responsible AI like they’re the exact same thing. They are connected, sure, but they aren’t identical. Sometimes it feels like they’re just, you know, one concept, but no.

Ethical AI is mostly about principles. It’s about fairness, human rights, privacy, transparency, and also the broader societal impact. Those ideas matter because they kind of set the direction, what organizations should aim for, in the first place.

Responsible AI is more like what follows after the talk ends. It’s the execution part, the practical side, when the conversation turns into decisions.

It turns principles into actions. It asks practical questions. Who owns AI risk? How will bias be tested? What documentation exists? How will decisions be explained? What happens if a model fails?

This distinction is becoming increasingly important as governments and regulators move from discussion to action. UNESCO’s Recommendation on the Ethics of Artificial Intelligence became the first global standard on AI ethics and applies across 194 member states. The message is clear. Ethical AI is no longer a theoretical concept. Organizations are expected to prove that responsibility exists inside their operations.

Pillar 1: Corporate Governance and Oversight

Every responsible AI framework starts with governance. Not technology. Not models. Governance.

One of the biggest mistakes organizations make is treating AI as a technical project owned only by data teams. AI decisions can create legal, operational, security, and reputational consequences. That means governance needs broader representation.

A strong AI governance board should include legal teams, compliance leaders, cybersecurity experts, data scientists, and business stakeholders. Different perspectives matter because AI risks rarely stay inside one department.

However, governance without authority is useless.

If a model shows a pretty major risk, then at least somebody should get the authority to stop the deployment. Governance structures need enforcement mechanisms, escalation routes that make sense, and also clear ownership, not just nice words.

Ownership is where many organizations seem to get stuck. When AI systems fail, a lot of people go ahead and blame the algorithm. That kind of framing avoids taking responsibility, it kind of sidesteps accountability instead of actually creating it. Every stage of the AI lifecycle should have a clearly assigned owner. Somebody owns the data. Somebody owns testing. Somebody owns compliance. Somebody signs off on deployment.

The urgency is obvious. IBM’s 2026 Tech Leader Study found that only 11% of CIOs and CTOs feel fully prepared for the scale of AI agent deployment expected over the next year. Companies are moving fast. Readiness is not.

Pillar 2: Data and Model Lifecycle Methodology

Responsible AI starts long before a model reaches production.

Everything begins with data. Poor data creates poor outcomes. If organizations cannot explain where data came from, whether consent exists, or how bias entered the dataset, they are creating risk from day one.

This is why data lineage matters. Teams should be able to trace data sources, understand transformations, and document ownership throughout the lifecycle. That visibility becomes critical during audits, investigations, and compliance reviews.

The next challenge is transparency.

High-performing models are valuable. Models that nobody understands create a different problem. Organizations increasingly need explainability, especially when AI influences customer experiences, employee decisions, or regulated processes.

Tools like SHAP and LIME help organizations understand why a model reached a specific conclusion. That explanation builds confidence and creates accountability.

Then comes testing.

This is where many companies cut corners. They test for functionality and assume everything else will work itself out. That approach does not survive in modern AI environments.

Responsible AI requires adversarial testing. Teams need to look for prompt injection risks, data leakage, harmful outputs, and unexpected behavior before deployment.

Google offers a useful example of this mindset. Google’s Content Adversarial Red Team completed more than 350 exercises during 2025 to identify vulnerabilities and stress-test systems. Gemini 3 also underwent Google’s most comprehensive safety evaluations to date. The lesson is simple. Strong AI systems are challenged before they are trusted.

Also Read: Best Practices for Cloud Migration and Modernization: A Strategic Roadmap for Enterprise Success

Pillar 3: Regulatory Compliance and International Standards

The compliance landscape is becoming more complicated every year.

Organizations now face overlapping regulations, privacy requirements, and industry standards. A framework that works in one market may not satisfy requirements somewhere else.

The EU AI Act reflects this shift a bit, and honestly it feels like it is saying, ‘not all AI is the same.’ Rather than just treating every AI system identically, it moves toward a risk based approach. In other words, higher-risk applications get tighter duties, while certain uses may even be limited or restricted completely.

At the same time, organizations really should look at the guidance coming from different frameworks like NIST AI RMF, the MeitY recommendations, and also consumer protection authorities.

The biggest mistake companies make is treating compliance as paperwork.

Real compliance is evidence. It is documented testing, risk assessments, governance reviews, monitoring records, and decision logs. When regulators ask questions, organizations need proof that controls exist and actually work.

Standards like ISO/IEC 42001 can help, kind of create that structure. They give you a formal framework for governance and accountability, but also for risk management, and then this whole continuous improvement loop. And more than that, they tend to make things consistent across teams, as well as across business units.

Pillar 4: Operational Monitoring and Continuous Auditing

Many organizations think deployment is the finish line.

It is not.

AI systems change because the world around them changes. Customer behavior evolves. Market conditions shift. New data enters the system. Over time, model performance can drift away from original expectations.

That is why continuous monitoring matters.

Organizations should track performance, review outputs, monitor anomalies, and create alerts when unusual patterns emerge. Waiting for customers to discover problems is not a monitoring strategy.

Continuous auditing is equally important. Governance controls should be reviewed regularly. Risk assessments should be updated. Compliance obligations should be reassessed as regulations evolve.

There should also be a clear response process. High-risk systems need escalation procedures and kill-switch capabilities when necessary. Problems are easier to manage when organizations act early rather than react late.

Conclusion

The real challenge with AI is no longer adoption. Most organizations have already crossed that bridge. The challenge is building systems that remain trustworthy after deployment.

Governance, accountability, transparency, compliance, testing, and monitoring are no longer optional layers. They are becoming core business requirements.

The financial part is kind of coming into view more. McKinsey’s 2026 AI Trust Maturity Survey found that organizations putting $25 million or more into responsible AI are more likely to see EBIT impact above 5% reported. And yeah, that shifts the whole conversation a bit, because it’s not only about lowering risk. Responsible AI is becoming, sort of, a real competitive edge. The firms that catch that early will probably be the ones that end up getting the biggest benefit.

The post Creating Responsible AI Development Frameworks: A Guide to Building Ethical, Transparent and Compliant AI Systems appeared first on ITDigest.

By delivering real-time, context-aware work intelligence across the market’s primary generative ecosystems, Smartsheet eliminates vendor lock-in and allows enterprises to extract deep operational value from the AI tools they have already deployed. Concurrently, the provider rolled out Smart Assist, a native in-product companion that utilizes the exact same continuous data models to surface actionable insights directly inside the Smartsheet interface.

Unifying Distributed Corporate Strategy with Actionable Context

Unlike basic, isolated API integrations that offer simple read-only access without an underlying understanding of corporate workflows, Smartsheet’s MCP architecture links intelligent assistants directly to live data. Grounded in 20 years of rich enterprise performance metrics, the server translates complex, multi-tiered project dependencies, row histories, and cross-functional communications into structured, clear instructions that external algorithms can process accurately. This removes the reliance on generic textual summaries, enabling corporate AI instances to generate real-world, actionable solutions.

“The problem most teams run into isn’t access to AI. It’s that their AI has no idea how their organization actually works,” said Pratima Arora, chief product and technology officer at Smartsheet. “Today’s assistants make one person faster inside one system. But enterprises don’t deliver that way — the build, the launch, the transformation run across teams and systems all at once. That context lives in Smartsheet because that’s where the work happens. When every major AI assistant connects to it, teams stop chasing information and start making decisions.”

Also Read: Cognizant and Snowflake to Drive Enterprise AI Adoption via Cortex-Powered Intelligent Agents

Escalating Enterprise Market Momentum and Tool Utilization

The introduction of the multi-assistant model arrives amid unprecedented customer adoption following the initial rollout of Smartsheet’s MCP server earlier this year. Corporate performance analytics demonstrate that introducing an open, protocol-based layer successfully shifts artificial intelligence from a passive conversational novelty into active operational plumbing.

Key adoption and performance metrics recorded since the platform’s initial compliance launch include:

• Substantial Transaction Volumes: Over 22,000 unique corporate users have executed more than 3 million distinct AI actions within the ecosystem.

• Accelerated User Growth: Weekly active adoption has grown nearly 9x since week one, expanding from fewer than 1,000 active users at launch to more than 9,000, while overall weekly tool call volumes jumped from 42,000 to more than 700,000.

• Consecutive Record Invocations: The first 10 days of June alone yielded more than 860,000 AI actions, achieving consecutive all-time daily usage peaks on June 9 and June 10 with 1,767 and 1,825 active organizations, respectively.

• Measurable Productivity Shifts: Nearly one in three AI-driven queries explicitly creates, updates, or modifies live operational records, illustrating that connected workflows generate concrete business outcomes rather than simple text outputs.

• Rapid Ecosystem Scaling: Approximately 3,000 net-new organizations integrated the technology into their workflows over the last 30 days, averaging close to 700 new corporate teams connecting to the server each week.

“We specialize in complex, technical construction projects—from building large-scale data centers to state-of-the-art healthcare facilities,” said Matthew Feagin, regional operations leader at DPR Construction. “There are thousands of people involved in these projects, and Smartsheet is the backbone for managing all of the most dynamic parts of the process. Now with the Smartsheet MCP Server, our teams can securely connect to their preferred AI tools to quickly build workflows, test ideas and get answers using natural language, all in a fraction of the time. That means our frontline workers can easily create Smartsheet solutions tailored to their unique challenges, helping them solve problems faster and reduce errors.”

Maintaining Security Governance Over Fragmented Tech Stacks

By exposing a single, securely governed data-sharing layer through the Model Context Protocol, Smartsheet provides internal IT departments with a clear blueprint to mitigate the risks associated with “shadow AI” and fragmented corporate software stacks. The system honors all existing, organization-level authorization rules and user permissions, ensuring that automated agents operate strictly within verified, least-privilege guardrails. Additionally, every interactive data request and tool call is fully logged, providing corporate security teams with the transparent audit trails required to maintain regulatory compliance at scale.

The Copilot and ChatGPT connectors are commercially available to United States enterprise accounts immediately, with regional rollouts across Europe, the Middle East, Africa (EMEA), and Asia-Pacific (APJ) scheduled to launch in subsequent product phases. Global corporate networks can access technical documentation, connection maps, and implementation guides via the official Smartsheet developer platform.

The post Smartsheet Expands MCP Server Integration to Support Copilot, ChatGPT, and Gemini Enterprise appeared first on ITDigest.

Also Read: Pega and AWS Partner to Fast-Track Mainframe-to-Cloud Modernization via Generative AI

The hybrid model spans fixed and variable components across subscription, managed-service, and factory delivery frameworks, making it natively compatible with Agentic Engineering, Business AI Transformation, and Application Development & Maintenance (ADM) environments. Underscoring the operational philosophy behind this strategic deployment, Venu Lambu, CEO and Managing Director of LTM, stated: “Enterprises are increasingly looking to align technology investments more closely with measurable business outcomes. BlueVerse Currency brings together talent, platforms, and intelligent agents into a unified commercial model enabling clients to move from input-based constructs to outcome-led value creation, while scaling AI adoption with greater flexibility and transparency.” Ultimately, backed by rigid enterprise security, data governance, and responsible AI guardrails, this formalized solution eliminates wasted technology spend, allowing regulated enterprises to smoothly evolve into scalable, high-performing agentic operations.

Read More: LTM Introduces BlueVerse Currency to Enable Outcome-Based Pricing in the Agentic AI Era

The post LTM Launches BlueVerse Currency to Pioneer Outcome-Based Pricing for Agentic AI appeared first on ITDigest.

Also Read: Hivelocity Launches Bare Metal Cloud Suite to Deliver Predictable Cloud Economics

This automated lifecycle significantly compresses digital transformation timelines, allowing enterprises to rapidly replace decades of technical debt with agile, governed agentic workflows. Emphasizing the operational philosophy driving this collaboration, John Higgins, chief of client and partner success at Pega, stated: “Modernizing core systems isn’t just about moving old code to the cloud – it’s about reimagining how work gets done. By embedding Pega Blueprint AI directly into the time-tested AWS Transform analysis capability, we’re removing friction from the most painful stages of modernization and helping enterprises finally move from legacy COBOL to intelligent, cloud‑ready applications faster and with greater confidence.” Currently available within the AWS Transform platform at no additional cost to existing clients, this joint capability bridges the gap between infrastructure legacy dependencies and the modern requirements of scalable corporate automation.

Read More: Pega Launches AI-Driven Modernization Capability on AWS To Reimagine Legacy Mainframe Apps Into Cloud-Native Systems

The post Pega and AWS Partner to Fast-Track Mainframe-to-Cloud Modernization via Generative AI appeared first on ITDigest.

Also Read: Bitwise and HoneyHive Partner to Advance Governed and Scalable Enterprise AI

This allows organizations to establish a unified platform built for API, AI, event, and agent connectivity while maintaining strict cost and security controls. Explaining the operational necessity of this new offering, Reza Shafii, SVP of Product, Kong, stated: “Every enterprise software vendor is racing to make their platforms accessible to AI agents. The challenge for most organizations is that the APIs powering their business were never designed to be discovered, governed, secured, and consumed by agents. Kong gives customers the platform they need to make their APIs agent-ready, and Ascent dramatically accelerates the journey.” Currently available for migrations from MuleSoft® environments with more platforms to follow, Kong Ascent provides enterprises with the foundational agility needed to confidently scale agentic AI workflows.

Read More: Kong Launches Ascent to Eliminate Legacy API Migration Risk and Help Enterprises Get Agentic AI-Ready

The post Kong Launches Ascent to Accelerate Legacy API Migrations for the Agentic AI Era appeared first on ITDigest.

Enterprise AI and HPC teams pay a compounding price for every cloud (or system) they run on: rebuilt pipelines, rewritten deployment scripts, profiling, testing and validation, before a single workload can run on a new infrastructure. That cost scales directly against the speed the business demands. Fuzzball eliminates it and completely levels the playing field.

A genomics team that validates a sequencing pipeline on AWS moves it to Azure or OCI without modifying a single line in the workflow definition. A model training job that requires H100 density routes to CoreWeave automatically, while a data-sensitive simulation stays on-premises by policy. The workflow definition, container images, data orchestration and job sequencing remain identical across every environment. Teams reach production faster, access better GPU capacity at lower cost and carry no operational overhead for every cloud they add.

Also Read: Beacon.li Launches Implementation Studio, the First AI Platform to Execute Enterprise Software Implementations End to End

“AI teams today are asked to ship faster, control costs and maintain sovereignty over their data, simultaneously, across infrastructure that was never designed to work together,” said Gregory Kurtzer, CEO and founder of CIQ. “We built Fuzzball to solve that problem at the architectural level. When your workflow definition abstracts its requirements properly, you get portable access to every GPU environment the market offers and the freedom to route to wherever the best price, performance and data policy lives. Controlling your infrastructure and workloads is what enterprise AI infrastructure requires for production, and no other platform delivers it.”

One control plane across five clouds and on-premises

Fuzzball’s multi-cloud architecture rests on a provider-agnostic workflow definition. The file that describes compute jobs, data movement, container images and resource requirements carries no cloud-specific logic. Fuzzball’s orchestration layer translates that definition into concrete infrastructure on whichever environment sits underneath, whether that means Google Cloud, Microsoft Azure, Oracle Cloud, AWS or CoreWeave.

Fuzzball federates across all five cloud environments, alongside on-premises clusters. It simultaneously evaluates available environments at runtime and routes each job to its optimal destination. Enterprises gain the GPU density of CoreWeave, the breadth of three major hyperscalers and the sovereignty of on-premises infrastructure from one control plane, with no separate toolchains, deployment scripts or IAM models per provider.

One security model across every environment

Each cloud deployment is provisioned through a two-phase automated process that stands up a complete, production-ready cluster without manual intervention. Fuzzball maintains one IAM model, one set of RBAC policies and one secrets management posture across every cloud it runs on. Static credentials are eliminated at every layer: Workload Identity on GCP, Managed Identities on Azure, Dynamic Groups on OCI and IAM Roles on AWS. Security and compliance posture travel with the workflow, not the cloud.

“Fuzzball turns multi-cloud from a liability into a competitive advantage,” said Bjorn Hovland, president of CIQ. “Five clouds used to mean five IAM models, five deployment pipelines and five sets of operational overhead, with complexity and risk being multiplied.”

Source: PRNewswire

The post CIQ Expands Fuzzball to Full Multi-Cloud so AI and HPC Teams Ship Faster and Spend Less appeared first on ITDigest.

Even though companies are using Artificial Intelligence they still encounter problems like data silos, outdated technologies, and the difficulty of integrating AI into daily activities. Cognizant is planning to solve these problems via its AI Builder approach, which merges state-of-the-art AI technologies, knowledge of different industries, and design centered on the user’s experience to help businesses see tangible results in a short period of time.

“When the data foundation is AI-ready and business context is built in, build cycles collapse,” said Naveen Sharma, Senior Vice President and Practice Head, AI & Analytics at Cognizant. “That is exactly what Cognizant’s AI Builder approach delivers, and with Snowflake CoCo, we are putting production-grade AI into enterprise workflows in hours, not weeks.”

Also Read: Incorta Launches Real-Time Data Foundation for Workday Adaptive Planning

The collaboration has already demonstrated significant momentum within Cognizant’s AI and Analytics practice. The CoCo platform has been adopted by more than 2,250 users across internal teams and client environments, supporting over 30 enterprise use cases. Cognizant has also developed numerous custom CoCo capabilities and AI accelerators that have collectively powered more than 1.3 million AI-driven requests.

A great example of this is A+E Global Media, where Cognizant put in a conversational analytics agent driven by Snowflake CoCo. This tech took care of reporting and data analysis jobs that used to be done manually. Now, business folks can query complex data sets using normal language. As a result, tons of time was saved, and it made ops more efficient and helped with better decision-making throughout important divisions.

Beyond just analytics, Cognizant is now pushing CoCo-powered solutions into other areas like contract smarts, compliance automation, finance ops, and anomaly detection. Companies see huge drops in manual work, major cost savings, and swifter process times thanks to these efforts.

Cognizant and Snowflake keep pouring money into special AI skills for different industries, plus prebuilt agent templates and orchestration abilities. Their goal? To give companies a way to easily deploy agentic AI and turn data into real business insights.

The post Cognizant and Snowflake to Drive Enterprise AI Adoption via Cortex-Powered Intelligent Agents appeared first on ITDigest.

Also Read: Kyndryl and Google Cloud Deepen Alliance to Power Enterprise AI and IT Modernization

The joint offering also delivers real-time observability, governance frameworks for compliance and auditability, federated architecture support for multi-business-unit deployments, and coding agent observability for tools such as Claude Code, Cursor, and Devin. The solution is specially designed for use by various industries that include finance, health care, and insurance industries where transparency, traceability, and adherence to regulations are key. The collaboration between Bitwise and HoneyHive seeks to ensure that businesses undergo successful AI transformation with ease through the management and deployment of their AI systems.

Read More: Bitwise and HoneyHive Announce Strategic Partnership to Enable Scalable, Governed Enterprise AI

The post Bitwise and HoneyHive Partner to Advance Governed and Scalable Enterprise AI appeared first on ITDigest.

The platform connects enterprise systems such as ERP, CRM, HRIS, and operational planning tools directly to Workday Adaptive Planning, with data refreshes available as frequently as every five minutes. According to the company, the offering addresses a common challenge for finance teams that often struggle with outdated actuals, disconnected operational data, and limited visibility into business performance until after financial close processes are completed.

Built on Incorta’s Direct Data Mapping technology, the solution gives finance departments greater control over their own governed data environment. FP&A teams can independently manage data feeds, adjust models, and update planning structures without waiting for IT support or enterprise data platform backlogs.

Also Read: Zendesk Unveils Autonomous Service Workforce to Transform Customer and Employee Support

The company said the platform is intended to improve real-time forecasting and scenario planning by providing access to operational metrics such as revenue trends, inventory changes, headcount movements, and spending activity before those figures formally appear in financial statements. Teams can also drill down from high-level variances to underlying transactions directly within the planning environment.

“Finance has long been forced to work on someone else’s timeline: waiting on IT for data feeds, waiting on the close for actuals, and waiting on operations for the drivers behind the numbers. The Adaptive Data Foundation changes that rhythm. Finance gets a live, governed planning data layer, business-owned and IT governed, giving teams the ability to reforecast earlier, investigate performance faster, and plan forward instead of simply explaining the past.” Mike Nader, VP & Field CTO at Incorta

Incorta showcased the change implementation by YES Communities, one of the largest U.S. operators of manufactured housing communities, that used integrated Workday Adaptive Planning and Incorta solution together to eliminate deadlock systems and delayed batch updates by providing real-time data visualization.

The initiative demonstrates the increasing need for financial systems that can deliver constant operational intelligence and enable faster decisions in the ever more dynamic business environments.

The post Incorta Launches Real-Time Data Foundation for Workday Adaptive Planning appeared first on ITDigest.