Advertising research firm MediaScience has introduced a new AI-driven approach that could fundamentally change how marketers test and optimize ad creative.

The company announced Creative Twin™, a new methodology that uses artificial intelligence to recreate advertisements with near-perfect accuracy and then test the impact of individual creative elements within them. The breakthrough will be formally presented at the Audience x Science Conference hosted by the Advertising Research Foundation.

Developed using proprietary technology within MediaPET.ai—a spinoff initiative from MediaScience—the system enables researchers to generate AI replicas of advertisements that audiences reportedly cannot distinguish from the originals.

For marketers increasingly focused on creative effectiveness and personalization, the technology promises something long considered nearly impossible: isolating and measuring the precise impact of individual creative decisions within an advertisement.

From Creative Guesswork to Scientific Measurement

Creative optimization has traditionally relied on A/B testing or multiple ad versions, both of which can be expensive and time-consuming. Testing individual variables—such as a celebrity endorsement, visual style, or casting decision—often requires entirely new ad production.

Creative Twin aims to eliminate that barrier.

Once an advertisement is recreated as an AI-generated replica, researchers can systematically modify individual elements within the ad—such as talent, visuals, messaging, or background details—while keeping the rest of the creative identical.

The result is a controlled testing environment where marketers can measure exactly how each element influences audience response.

“This represents a fundamental shift in how advertising creative can be evaluated and optimized,” said Duane Varan. “For the first time, researchers can isolate and measure the contribution of individual creative elements within an advertisement.”

AI Replicas That Audiences Can’t Distinguish

To validate the methodology, MediaScience conducted controlled testing with 812 U.S. respondents in collaboration with the Ehrenberg-Bass Institute, one of the world’s most respected academic marketing research institutions.

Participants were shown both original advertisements and AI-generated replicas created using the Creative Twin technology.

According to the study, respondents were unable to distinguish between the original ads and the AI-generated versions, confirming that the replicas maintained the full production quality and realism of the original creative.

That fidelity is critical, because it allows researchers to modify ad components without introducing unintended differences that could skew results.

Testing the True Impact of Creative Elements

With the AI-generated “twin” in place, advertisers can test a wide range of creative variables.

For example, marketers can explore:

• The impact of different celebrities or influencers
• Alternative casting choices
• Visual design variations
• Messaging changes
• Background imagery or brand elements

Because each version of the ad remains visually identical except for the specific variable being tested, researchers can determine the incremental impact of each creative choice.

This level of control has historically been difficult to achieve without producing multiple expensive ad variations.

Personalized Advertising Without New Production

Beyond research applications, the technology could also reshape how brands personalize advertising at scale.

Addressable advertising—where ad creative is tailored to specific audience segments—often requires producing multiple versions of the same ad. Creative Twin enables marketers to generate these variations digitally without reshooting the ad.

One example highlighted by MediaScience involved a shampoo commercial.

In the original advertisement, the featured model had straight hair. Using Creative Twin, researchers generated an AI-modified version of the same ad where the model appeared with curly hair.

When shown to audiences with curly hair, the modified ad delivered significantly stronger results across several key marketing metrics, including:

• Brand recognition
• Brand attitude
• Brand preference (a proxy for purchase intent)

The results suggest that subtle creative adjustments—when matched to the right audience—can meaningfully improve advertising effectiveness.

Expanding Creative Testing Across Categories

The methodology also opens the door to highly targeted creative variations across different industries.

MediaScience outlined several potential applications:

• Pet food advertising: A commercial featuring a Labrador could be digitally adapted to show a poodle or French bulldog for owners of those breeds.
• Celebrity endorsements: Marketers could test whether replacing a celebrity with another talent—or removing the celebrity altogether—changes ad impact.
• Product personalization: Visuals and messaging could be adapted to align with specific audience demographics or preferences.

In each case, the modified ad retains the same production quality as the original version, ensuring that the test focuses solely on the element being evaluated.

Measuring the Economic Value of Creative Decisions

One of the most intriguing aspects of Creative Twin is its potential to quantify the financial impact of creative decisions.

Advertising production often involves major investments in talent, filming, and design, but marketers rarely have precise data on which elements deliver the greatest return.

Creative Twin allows researchers to test these elements individually, helping brands determine:

• Whether a celebrity endorsement justifies its cost
• Which visuals resonate most strongly with target audiences
• How casting or styling decisions influence brand perception
• Which creative components drive purchase intent

For marketing teams under pressure to prove ROI on advertising budgets, that level of insight could be transformative.

A New Era of AI-Driven Advertising Research

The introduction of Creative Twin reflects a broader trend toward AI-driven experimentation in marketing and advertising.

As generative AI tools become more sophisticated, marketers are gaining the ability to simulate and test creative variations without the traditional production costs associated with advertising development.

For MediaScience, the technology represents the next phase in advertising research—moving from observation to controlled experimentation.

If widely adopted, AI-powered creative replication could reshape how brands design, test, and personalize advertising campaigns in the years ahead.

Get in touch with our MarTech Experts.

Turning business intelligence into measurable outcomes remains one of the biggest challenges organizations face when adopting AI. Now, research and market intelligence firm Keypoint Intelligence is attempting to close that gap with the launch of Keypoint Engage, a new AI-powered execution services suite designed to transform industry insights into real operational results.

The new offering combines automation, AI-driven tools, and managed services to help companies accelerate execution across sales, marketing, and customer service operations—areas where organizations often struggle to translate strategic insights into day-to-day performance improvements.

Rather than offering standalone software or advisory services, Keypoint Engage is positioned as a managed, outcome-focused solution that delivers operational execution based on Keypoint Intelligence’s decades of industry data and research.

Turning Industry Intelligence Into Business Action

For years, Keypoint Intelligence has been known for its research and testing in the print, imaging, and document technology industries. With Keypoint Engage, the company is shifting from a purely advisory role into a more operational one—helping organizations implement AI-driven strategies directly within their business processes.

According to Randy Dazo, the move reflects a common challenge companies face when adopting AI technologies.

“Businesses want the benefits of AI, but most don’t have the time, resources, or expertise to implement it effectively,” Dazo said. “Keypoint Engage closes the gap between insight and execution by delivering real action and measurable results.”

Instead of requiring companies to master complex AI tools or develop in-house expertise, Keypoint Engage provides end-to-end execution services, including automation, campaign support, and AI-driven operational tools.

AI-Powered Execution Across Core Business Functions

The Keypoint Engage platform focuses on three primary operational areas where AI can deliver immediate business impact: sales execution, marketing execution, and service automation.

Each module within the suite is designed to streamline workflows and improve performance through a combination of AI agents, automation tools, and strategic guidance.

Sales Execution

The sales-focused component of Keypoint Engage aims to help organizations generate and convert leads more efficiently.

Capabilities include:

• AI-assisted prospecting and lead generation
• Automated inbound and outbound calling agents
• Improved sales communication and messaging
• Training and enablement tools for sales teams

By automating routine outreach and supporting more personalized engagement strategies, the platform aims to strengthen pipeline development and improve customer interactions.

Marketing Execution

Marketing teams often face pressure to launch campaigns faster while producing more content across multiple channels. Keypoint Engage addresses that challenge with AI-powered campaign planning and content development tools.

The platform supports:

• Campaign strategy and execution planning
• AI-generated thought leadership content
• AEO-optimized websites and blogs
• Video and multimedia marketing assets

AEO, or Answer Engine Optimization, focuses on optimizing content for AI-powered search and conversational platforms rather than traditional search engines alone.

By automating content development and campaign workflows, Keypoint Engage aims to reduce the time and manual effort required to deploy marketing initiatives.

Service and Support Automation

Customer service operations represent another area where AI-driven automation is rapidly gaining traction.

Keypoint Engage introduces tools designed to streamline support interactions and improve service consistency, including:

• AI-powered chatbots for customer interactions
• Automated outbound service calls
• Faster response and issue resolution workflows
• Scalable support infrastructure

The goal is to help organizations provide faster, more consistent customer experiences while reducing the operational costs associated with high-volume support environments.

Expanding Beyond Research and Advisory Services

The launch of Keypoint Engage marks a significant evolution for Keypoint Intelligence as the company moves deeper into operational services.

Traditionally, firms like Keypoint have focused on market analysis, benchmarking, and strategic insights. But as AI tools become more accessible, businesses increasingly want partners that can not only provide insights but also help execute strategies.

Keypoint Engage reflects that shift by blending industry expertise, automation technologies, and managed service delivery into a single offering.

The company says the suite will also help organizations better evaluate and optimize the role of print and document technologies within modern business workflows—areas where Keypoint Intelligence has long-standing expertise.

AI Execution as a Service

The broader market trend behind Keypoint Engage is the emergence of AI execution services—solutions that go beyond software to deliver managed outcomes.

Many organizations are enthusiastic about AI’s potential but lack the internal resources needed to integrate it into existing processes. That has created growing demand for vendors capable of delivering AI-powered business execution without requiring deep technical expertise from customers.

By packaging AI automation with industry intelligence and operational services, Keypoint Intelligence hopes to provide companies with a more accessible path to measurable results.

Availability

Keypoint Engage is available now as a managed service suite for organizations looking to accelerate sales, marketing, and service performance using AI-driven automation.

The company says additional resources—including a dedicated landing page and overview materials—are available to help organizations explore how the platform can support their operational strategies.

For businesses seeking to move beyond experimentation and toward real AI-driven execution, Keypoint Engage represents a new model: turning insight into action—and action into measurable growth.

Get in touch with our MarTech Experts.

Enterprise resource planning (ERP) platforms hold some of the most valuable operational data inside organizations—but extracting meaningful insights from that data often requires complex reporting tools and technical expertise. A new partnership aims to change that for the NetSuite ecosystem.

Detroit-based NetSuite consulting firm Snapshot and enterprise AI platform provider MindsDB have announced a strategic collaboration to deliver AI-powered conversational analytics for businesses running NetSuite.

The partnership combines Snapshot’s operational and ERP expertise with the AI capabilities of MindsDB Enterprise AI Platform. The goal: allow companies to interact with their NetSuite data—and data from other enterprise systems—through AI-driven insights, automation, and natural language queries.

For organizations that rely on Oracle NetSuite to manage financials, supply chains, and operations, the integration could unlock faster decision-making across business units.

Turning ERP Data Into AI-Ready Intelligence

ERP systems like NetSuite serve as the operational backbone for many organizations, managing everything from accounting and inventory to procurement and logistics. But turning that data into actionable insights typically involves manual reporting workflows or external analytics platforms.

Snapshot and MindsDB aim to streamline that process by creating an AI layer capable of connecting directly to enterprise datasets.

At the center of the collaboration is the Minds Enterprise platform, which acts as the AI backbone for deploying models, agents, and conversational analytics across enterprise data sources.

Snapshot will build on that foundation by applying its expertise in NetSuite data modeling and ERP implementation. The firm plans to translate complex ERP structures into AI-ready data frameworks that can power predictive analytics and automated insights.

“Our partnership with MindsDB allows us to deliver AI capabilities, autonomous agents, and statistical analytics that are deeply connected to how NetSuite customers actually operate,” said Tania Sottrel.

AI Analytics Across Multiple Business Systems

While the collaboration focuses on NetSuite environments, the platform is designed to extend beyond a single ERP system.

Companies will be able to combine NetSuite data with information from other operational platforms, including:

• E-commerce platforms
• Supply chain and logistics systems
• Customer relationship management tools
• Warehouse management systems
• External market datasets

This multi-source approach enables organizations to build a unified intelligence layer across their operations rather than relying on isolated reporting tools.

According to Brad Gyger, the goal is to bring AI directly to where enterprise data already lives.

“MindsDB was built to bring AI directly to enterprise data,” Gyger said. “By partnering with Snapshot, we’re enabling NetSuite customers to unlock AI insights across their ERP systems and the broader platforms that power their businesses.”

Moving Beyond Traditional ERP Reporting

Traditional ERP reporting often focuses on historical metrics—financial statements, inventory counts, or operational summaries. While useful, those reports rarely provide predictive insights or automated recommendations.

The Snapshot–MindsDB platform aims to expand those capabilities with AI-driven features such as:

• Predictive demand forecasting
• Inventory and supply chain optimization
• Pricing and margin analysis
• Operational anomaly detection
• AI-assisted reporting and decision support
• Cross-system operational intelligence

Instead of static dashboards, companies will be able to ask questions in natural language and receive AI-generated insights derived from ERP and operational data.

This conversational approach reflects a broader shift toward AI-driven business intelligence, where machine learning models continuously analyze operational data to identify patterns and recommend actions.

Targeting Distribution and Supply Chain Industries

The initial focus of the partnership will be industries where NetSuite plays a major role in operational and supply chain management.

These sectors include:

• Fastener and industrial component distribution
• Automotive parts and aftermarket supply chains
• Food manufacturing and distribution
• Agriculture and landscaping supply
• HVAC distribution
• Plumbing and electrical supply

Companies in these industries often manage large product catalogs, complex supplier networks, and fluctuating demand conditions—making them ideal candidates for predictive analytics and automated operational insights.

By combining ERP data with external datasets, the companies hope to enable businesses to anticipate demand shifts, optimize inventory, and identify operational risks earlier.

Building an AI Layer for the NetSuite Community

Snapshot will lead development of NetSuite-specific AI solutions, including models, AI agents, ERP knowledge layers, and connectors that translate NetSuite data into AI-ready frameworks.

The broader goal is to create an intelligence layer that sits on top of NetSuite and integrates with other enterprise systems.

That approach could help companies modernize their analytics infrastructure without replacing existing ERP systems—an attractive option for organizations heavily invested in NetSuite environments.

“Our goal is to empower the NetSuite community with AI that understands how their businesses actually operate,” Sottrel said.

Availability and Next Steps

The NetSuite connector for MindsDB is available now, enabling organizations to begin integrating ERP data with the AI platform.

Additional solutions, pilot programs, and industry-focused deployments are expected to roll out throughout 2026 for NetSuite customers and ecosystem partners.

As enterprise AI adoption accelerates, partnerships like this one highlight a growing trend: embedding AI directly into core operational systems rather than layering analytics tools on top.

For NetSuite users seeking deeper insights from their operational data, that shift could transform ERP systems from record-keeping platforms into real-time intelligence engines for business decision-making.

Get in touch with our MarTech Experts.

Agentic AI is quickly becoming a strategic priority for governments seeking faster, data-driven decision-making. Now, AI platform provider Seekr and government technology contractor General Dynamics Information Technology are joining forces to bring those capabilities to federal missions.

The two companies announced a collaboration to develop agentic AI solutions tailored for government agencies, combining Seekr’s secure AI platform with GDIT’s integration expertise and deep experience supporting federal operations.

At the center of the partnership is SeekrFlow, Seekr’s end-to-end AI operating system designed to build and deploy AI agents in highly secure environments. The platform enables agencies to develop AI-powered workflows intended to improve operational efficiency, accelerate decision-making, and reduce costs across government missions.

Building AI for High-Stakes Government Environments

Unlike many commercial AI platforms built primarily for cloud environments, SeekrFlow is designed to operate in mission-critical conditions often required by defense and government agencies.

The system unifies several core AI capabilities into a single platform, including:

• Model hosting and lifecycle management
• Model fine-tuning for specialized workloads
• Agent orchestration and automation
• Full observability and monitoring for AI agents

This integrated approach is intended to eliminate the need for agencies to assemble multiple AI tools into a single workflow—an often complex and resource-intensive process.

SeekrFlow also supports deployment in air-gapped networks, disconnected systems, and tactical edge environments, making it suitable for military and classified operations where internet connectivity may be restricted.

According to the company, the platform is already deployed across branches such as the U.S. Army and U.S. Navy, as well as other defense agencies.

From AI Pilots to Operational Systems

One of the biggest challenges facing government agencies today is moving AI initiatives from experimental pilots into operational deployments.

Through the collaboration, Seekr and GDIT aim to accelerate that transition by delivering production-ready AI agents capable of supporting real-world missions.

“By combining Seekr’s agentic AI with GDIT’s leadership in federal mission delivery, we’re enabling agencies to move faster, operate smarter, and achieve outcomes once thought impossible,” said Rob Clark.

For its part, GDIT brings decades of experience implementing technology systems for federal civilian agencies, defense organizations, and intelligence communities.

Ben Gianni said government organizations increasingly need advanced technologies capable of keeping pace with evolving mission demands.

“Our collaboration with Seekr will enable us to deliver differentiated agentic AI solutions that help customers advance missions faster, smarter, and more securely,” Gianni said.

Key Government Use Cases for Agentic AI

The partnership is focused on developing AI agents designed to handle complex operational tasks across government agencies.

Early use cases include:

Case management automation
AI agents can streamline administrative workflows and process large volumes of government records more efficiently.

Risk and fraud detection
Agentic systems can identify suspicious patterns across financial, operational, or procurement datasets.

Cross-database intelligence analysis
AI agents can analyze data across multiple disconnected systems, helping agencies identify policy-aligned actions and operational priorities.

These applications are particularly valuable in government environments where data often resides in siloed systems spread across multiple departments and networks.

The Rise of Agentic AI in Government

The collaboration reflects a broader shift toward agentic AI systems capable of autonomous task execution, rather than simple AI assistants.

Federal agencies are increasingly exploring AI agents that can analyze data, execute tasks, and generate recommendations with minimal human intervention.

Seekr’s platform is also part of the Chief Digital and Artificial Intelligence Office Tradewinds Solutions Marketplace, a program designed to accelerate the adoption of AI capabilities across the Department of Defense.

Being available through that marketplace allows agencies to evaluate and procure AI technologies more quickly, bypassing some of the traditional procurement hurdles that often slow technology adoption.

AI-Powered Security Operations Centers

Another area of collaboration focuses on the future of cybersecurity operations.

Seekr is participating in GDIT’s ecosystem of Digital Accelerators and Centers of Excellence, working with technologists and mission teams to develop scalable AI-powered solutions.

One example is the integration of autonomous AI capabilities into next-generation Security Operations Centers (SOCs) using GDIT’s internal innovation platforms:

• Eclipse
• Luna

These initiatives aim to build more adaptive security systems capable of identifying threats, prioritizing risks, and responding to cyber incidents in near real time.

A Growing Market for Government AI

Government spending on AI technologies continues to expand as agencies seek ways to improve operational efficiency, strengthen national security, and modernize public services.

At the same time, concerns around data security, transparency, and operational resilience mean that many agencies prefer AI platforms designed specifically for secure and classified environments.

That trend has created a growing market for vendors capable of delivering AI systems that operate both on-premises and in restricted networks—a capability SeekrFlow emphasizes.

By pairing Seekr’s AI platform with GDIT’s integration and mission expertise, the companies are positioning themselves to address that demand.

Looking Ahead

As AI adoption across government accelerates, partnerships between AI platform developers and large federal integrators are becoming increasingly common.

For agencies tasked with managing complex missions and vast amounts of data, agentic AI platforms could offer a path toward faster insights, smarter automation, and improved operational resilience.

The Seekr–GDIT collaboration aims to deliver exactly that: mission-ready AI systems capable of operating in the most demanding government environments.

Get in touch with our MarTech Experts.

The telecom industry’s shift toward AI-driven operations is accelerating, and two technology providers are joining forces to help operators keep pace.

Digital telecom software firm Circles has signed a strategic collaboration agreement with Huawei to explore the joint delivery of AI-native, next-generation telecom platforms for operators worldwide.

The partnership aims to combine Huawei’s network and cloud infrastructure with Circles’ digital business support system (BSS) software delivered through a vertical SaaS platform. Together, the companies plan to help telecom providers modernize legacy systems, introduce AI-powered services, and unlock new monetization models.

For telecom operators grappling with growing data consumption, complex pricing models, and increasingly demanding customer expectations, the collaboration signals a push toward AI-first telecom architecture.

A Push Toward AI-Native Telecom Infrastructure

Telecommunications companies are under pressure to evolve beyond traditional network operators into digital service providers. That transformation often requires modernizing billing systems, automating customer engagement, and introducing dynamic pricing models tied to real-time network conditions.

Under the new agreement, Circles and Huawei will explore integrating policy control, charging systems, cloud infrastructure, and intelligent automation into a unified platform.

At the center of the initiative is the potential integration between Huawei’s charging and policy management capabilities and Circles’ digital BSS SaaS platform.

The companies say the combined system could enable:

• Real-time monetization through advanced charging and policy orchestration
• AI-driven policy optimization to dynamically manage network performance and resources
• Automated customer lifecycle management powered by data-driven personalization

The goal is to allow telecom operators to launch new digital services faster while simultaneously improving revenue generation and customer experiences.

“Telecom operators are at an inflection point where AI is no longer optional—it is foundational,” said Sanjay Kaul. “By combining Huawei’s network expertise with our AI-native digital BSS platform, operators can accelerate monetization and deploy intelligent services at scale.”

The Role of Digital BSS in Telecom Transformation

Business support systems (BSS) handle critical telecom operations such as billing, customer management, and product catalog management. Historically, many telecom operators rely on legacy BSS platforms that can be expensive to maintain and difficult to modernize.

Circles has positioned its platform as a cloud-native, AI-powered BSS alternative, designed to help operators transition toward digital-first business models.

Integrating that software layer with Huawei’s telecom infrastructure stack could offer operators a more cohesive network-to-digital architecture—linking infrastructure management with customer-facing digital services.

For telecom providers launching 5G services and exploring AI-driven network optimization, such integrations could help streamline operations across multiple layers of the technology stack.

Huawei Cloud as the Deployment Backbone

The partnership also includes plans to explore deploying Circles’ SaaS platform on Huawei Cloud infrastructure.

Running the platform on Huawei Cloud could allow telecom operators to deploy AI-powered systems in environments designed to meet regulatory compliance, data residency requirements, and performance demands across different global markets.

This “sovereign-ready” architecture is increasingly important as governments introduce stricter data governance rules and telecom companies expand into new regions.

According to Alex Kang, Huawei Cloud’s long-standing work with telecom operators positions the company well to support such deployments.

“Huawei Cloud has been deeply engaged in supporting telecom operators’ digital transformation worldwide,” Kang said. “We look forward to working with Circles to develop joint solutions and bring Circles’ products onto the Huawei Cloud Marketplace.”

Beyond Technology: Joint Market Expansion

The collaboration extends beyond technical integration.

Circles and Huawei are also exploring joint go-to-market initiatives, which could include co-selling integrated telecom solutions to global operators. These offerings would target telecom providers seeking to replace legacy operational systems with modern digital platforms powered by AI and automation.

Such initiatives could position the combined stack as an alternative to traditional telecom vendors that provide monolithic infrastructure and operational software.

By pairing Huawei’s large global telecom footprint with Circles’ specialized SaaS capabilities, the companies hope to reach operators transitioning toward software-driven telecom operating models.

AI Becomes the Telecom Industry’s Next Platform

The broader telecom sector is increasingly embracing AI across network operations, customer service, and revenue management.

Operators are exploring AI-driven network optimization, predictive maintenance, automated customer support, and personalized service offerings. At the same time, software-defined networking and cloud-native architectures are reshaping how telecom systems are built and deployed.

Partnerships like this reflect a growing industry consensus: the next phase of telecom innovation will rely heavily on AI-native infrastructure integrated across network and business layers.

For telecom providers navigating the shift to 5G, edge computing, and AI-enabled services, the ability to integrate infrastructure with digital monetization platforms may become a competitive advantage.

Looking Ahead

While the collaboration remains exploratory, the companies say their shared goal is to develop an integrated architecture capable of supporting rapid service innovation, automation at scale, and operational efficiency.

If successful, the partnership could give telecom operators new tools to modernize legacy systems and build intelligent service platforms designed for the AI era.

As the telecom industry moves toward software-defined operations, alliances between infrastructure providers and digital platform companies may become an increasingly common strategy for delivering next-generation telecom services.

Get in touch with our MarTech Experts.

AI agents are rapidly evolving from passive assistants into active collaborators—and website publishing platform WordPress.com wants them managing your site.

The platform, operated by Automattic, has launched new write capabilities for its Model Context Protocol (MCP) server, allowing AI agents to create, edit, and manage content directly on WordPress.com websites. The update enables conversational control over site publishing through AI tools such as ChatGPT, Claude, and Cursor.

The feature marks a significant step toward what many developers call the “agentic web”—an emerging model where AI agents don’t just generate text but actively interact with software platforms to complete tasks.

Given WordPress.com’s scale—70 million posts published each month—the platform offers one of the largest real-world environments for AI-powered site management.

Turning AI From Assistant Into Site Manager

Until now, AI tools integrated with WordPress largely focused on content generation. The new MCP capabilities push things further by letting AI agents execute actions inside a WordPress site through conversation.

In practice, that means users can instruct an AI agent to publish a blog post, update a page, or manage content without logging into the WordPress dashboard.

“WordPress.com is where millions of people build and manage their sites every day, and more and more of them are using AI tools to get work done,” said Ronnie Burt. “Now those tools can actually take action—draft a post, build a page, manage comments—directly on your site through conversation.”

For marketers, bloggers, and content teams, the workflow shift could be substantial. Instead of toggling between AI writing tools and a CMS interface, publishing tasks can now happen within a single AI-driven conversation.

What AI Agents Can Now Do on WordPress

The MCP write update gives compatible AI agents direct operational access to WordPress.com sites through a structured API.

Users can instruct their AI assistant to:

• Draft and publish blog posts or pages
• Edit and update existing content
• Create new pages and manage site structures
• Handle various content management tasks through natural language

The result is a conversational CMS workflow, where AI acts as a publishing operator rather than just a writing assistant.

That capability could prove particularly appealing for marketing teams managing high-volume content strategies or multi-site publishing operations.

Built on the Model Context Protocol

The feature is powered by the Model Context Protocol (MCP), an emerging open standard designed to let AI agents securely connect to external services.

Through the MCP server, AI agents can interact with WordPress.com sites using a structured interface secured with OAuth 2.1 authentication. The protocol allows agents to read site data, retrieve analytics, and now—thanks to the latest update—write and manage content.

In simple terms, MCP acts as the bridge between AI models and real-world tools.

The update builds on the initial MCP server release in October 2025, which allowed AI agents to access site content and analytics but not modify them. The new write capabilities close that loop, enabling agents to act on user instructions.

A Year of AI Expansion on WordPress.com

The new capabilities are part of a broader AI strategy for WordPress.com that has steadily expanded over the past year.

In April 2025, the platform introduced an AI-powered website builder, allowing users to generate fully designed websites from simple prompts. The system automatically creates layouts, pages, and starter content.

Later, the company launched the WordPress AI Assistant, embedded directly into the site editor and media library. The assistant helps users generate, edit, and refine content without leaving the editing interface.

Together, these features signal WordPress.com’s ambition to position itself as a central hub for AI-driven website creation and management.

Safeguards to Keep Humans in Control

While AI agents can now perform publishing actions, WordPress.com emphasizes that users remain firmly in control.

Several safeguards are built into the MCP system:

• Explicit confirmation required before actions are executed
• Immediate warnings when edits affect published content
• Opt-in activation for MCP functionality
• Site-by-site permissions for AI capabilities

This layered approach reflects growing concerns around autonomous AI systems making changes to live digital properties.

For enterprises and professional publishers, the safeguards are likely essential for maintaining editorial oversight and brand consistency.

Why WordPress Is a Natural Fit for AI Agents

WordPress.com’s scale makes it a particularly attractive target for AI-powered automation.

The platform runs on the open-source WordPress software, which powers more than 40% of all websites globally. That massive footprint gives AI developers a familiar and widely supported environment for integration.

Automattic’s broader ecosystem also handles hundreds of billions of page views annually, further reinforcing WordPress’s role as one of the web’s largest publishing infrastructures.

For AI developers, integrating with a platform operating at that scale provides immediate real-world relevance.

The Rise of the Agentic Web

WordPress.com’s move reflects a broader industry shift toward AI agents capable of operating software tools directly.

Tech companies are increasingly designing APIs and protocols specifically for agent-based interactions. Instead of simply generating outputs, AI models are expected to perform tasks across software ecosystems—from writing code to managing websites and executing marketing workflows.

For digital marketers and content teams, this could reshape how publishing pipelines work. A single AI agent could eventually research topics, generate drafts, optimize SEO, publish content, and track analytics—all within a conversational interface.

WordPress.com’s MCP update brings that vision closer to reality.

Availability

The MCP write capabilities are available immediately for all paid WordPress.com plans. The feature works with any AI agent that supports the MCP standard, including ChatGPT, Claude, and Cursor.

The MCP server is included at no additional cost for paid users and can be enabled directly within WordPress.com settings.

For a platform that already processes 70 million new posts every month, the introduction of AI-driven site management could mark the beginning of a new publishing era—one where websites are managed as much through conversation as through dashboards.

Get in touch with our MarTech Experts.

As enterprises accelerate AI adoption, data security is becoming just as critical as model performance. Vector database provider Zilliz is betting that stronger encryption controls will be a deciding factor for organizations deploying AI at scale.

The company announced the general availability of Customer-Managed Encryption Keys (CMEK) for Zilliz Cloud, giving enterprises the ability to retain full ownership and control of their encryption keys when running AI workloads. The move is aimed squarely at organizations in heavily regulated industries such as healthcare, financial services, and government—sectors where strict data protection rules often slow or block AI deployments.

Zilliz is best known as the company behind Milvus, widely used for similarity search and AI applications such as recommendation engines, semantic search, and large language model retrieval pipelines.

Security Becomes the Next AI Battleground

Vector databases have emerged as a core component of modern AI stacks. They store embeddings—numerical representations of text, images, or other data—that power applications like semantic search and generative AI retrieval systems.

But those embeddings are often derived from highly sensitive data sources, including customer records, medical scans, and financial transaction histories. That creates new security and compliance challenges.

Standard encryption-at-rest is typically not enough for enterprises operating under regulations such as GDPR, HIPAA, PCI-DSS, or SOC 2. Increasingly, regulators and auditors require proof that companies—not their vendors—maintain exclusive control over encryption keys.

With CMEK support, Zilliz aims to close that gap.

“Security teams in regulated industries don’t just want encryption—they want proof that no one else, including their database vendor, can access their data,” said Charles Xie in the announcement. “Customer-managed keys provide the strongest form of data sovereignty available in a managed service.”

What CMEK Brings to Zilliz Cloud

The new feature separates encryption key ownership from the infrastructure running the database. In practical terms, this means customers maintain full authority over their keys while Zilliz continues to manage the underlying vector database infrastructure.

That architecture introduces several security advantages for enterprise deployments.

True separation of duties
Organizations keep exclusive ownership of encryption keys while Zilliz handles the compute and data operations. This clear separation is often required for compliance audits.

Immediate access revocation
If a company disables its key in AWS Key Management Service, any associated cluster data instantly becomes cryptographically inaccessible—without needing coordination from the vendor.

Centralized audit logging
All key access events are logged in AWS CloudTrail, enabling enterprises to integrate encryption activity into their existing security monitoring systems.

From an operational standpoint, the company says setup takes only a few minutes through the Zilliz Cloud console. The platform automatically generates required IAM policies and supports zero-downtime key rotation—a key requirement for large production environments.

Why Vector Databases Need Stronger Security

The timing of the release reflects a broader shift in the AI infrastructure market. As organizations move from experimental AI pilots to production systems, security requirements are tightening.

Vector databases have rapidly become a cornerstone technology for AI applications, especially retrieval-augmented generation (RAG). Competitors such as Pinecone, Weaviate, and Qdrant are also racing to build enterprise-grade security and compliance features into their managed offerings.

Industry analysts note that encryption control is often a dealbreaker in sectors where data privacy laws are strict. Financial institutions and healthcare providers, for example, may be legally required to demonstrate that encryption keys are fully under their control—even when infrastructure is hosted in the cloud.

In that context, CMEK has become a baseline capability across many enterprise cloud services. Bringing it to vector databases signals that the AI infrastructure market is maturing quickly.

Removing a Barrier to Enterprise AI

For organizations deploying large-scale AI systems, the biggest obstacles are rarely model accuracy or compute capacity. Instead, they’re governance and risk management.

Features like customer-managed encryption keys address those concerns directly by allowing enterprises to enforce internal security policies while still benefiting from managed cloud infrastructure.

Zilliz clearly sees this as a strategic unlock for enterprise AI adoption.

By allowing customers to control encryption keys externally—while still running fully managed vector database clusters—the company hopes to remove one of the last barriers preventing regulated organizations from deploying AI applications at scale.

Availability

Customer-Managed Encryption Keys are generally available now for Dedicated clusters on the Zilliz Cloud Business-Critical plan. The initial rollout supports deployments running on AWS, with expansion to other cloud providers expected over time.

Enterprises can enable the feature directly through the Zilliz Cloud console or work with the company to configure production deployments.

For organizations navigating the complex intersection of AI innovation and regulatory compliance, the message from Zilliz is clear: data sovereignty may soon be just as important as AI capability itself.

Get in touch with our MarTech Experts.

Reply and Mistral AI are joining forces to accelerate enterprise adoption of secure, locally deployed, and fully customizable generative AI solutions. The partnership aims to help organizations in regulated sectors harness AI while ensuring data control, privacy, and compliance.

The collaboration leverages Mistral AI’s high-performance models alongside Reply’s expertise in designing Large Language Models (LLMs) trained on proprietary and domain-specific datasets, allowing AI deployments to integrate seamlessly into operational workflows across industries like finance, healthcare, public administration, defence, telecommunications, and energy & utilities.

Custom AI for Regulated Industries

A central goal of the partnership is enabling organizations to deploy generative AI solutions that meet stringent regulatory and operational requirements. By combining model performance with operational governance, organizations can:

• Maintain strict control over sensitive data

• Ensure compliance with local and European regulations

• Deploy AI solutions on sovereign infrastructures

Filippo Rizzante, CTO of Reply, emphasized that the initiative allows enterprises to scale AI deployments while keeping governance, data sovereignty, and security front and center.

Mistral Forge: Tailored Models for Operational Use

Reply will act as a global launch partner for Mistral Forge, enabling the creation of custom LLMs for complex, data-intensive domains. The platform allows teams to design, train, and deploy models on proprietary datasets—turning generic AI into enterprise-grade tools that are both specialized and operationally ready.

This level of customization is particularly important in sectors where standard AI models may not capture domain-specific knowledge or regulatory nuances, such as financial compliance or industrial operations.

Highlight Project: Ancient Greek LLM

The partnership’s capabilities are already being demonstrated through a collaboration with the Austrian Academy of Sciences. Reply and Mistral AI are developing a customized LLM for the Greek language, spanning ancient, medieval, and modern texts.

The model is trained on a highly curated corpus, including:

• Published ancient Greek literature

• Digitized inscriptions and papyri

• Selected modern Greek texts from scholarly and public sources

Designed to assist researchers, it provides advanced text search, completion, and analysis capabilities. This initiative highlights how sovereign AI infrastructure can support highly specialized, data-intensive use cases while maintaining accuracy and reliability.

Why It Matters

Generative AI adoption in enterprise and research environments is often constrained by regulatory, privacy, and operational risks. By combining sovereign infrastructure, model customization, and domain expertise, the Reply–Mistral AI partnership addresses three of the biggest adoption barriers:

1. Control: Organizations retain ownership and oversight of proprietary data and AI models.

2. Compliance: Models operate in alignment with strict privacy and regulatory requirements.

3. Performance: Custom-tailored LLMs deliver relevant outputs for specialized tasks and operational processes.

Marjorie Janievicz, Chief Revenue Officer at Mistral AI, noted that the collaboration will help organizations deploy AI solutions that meet enterprise expectations for control, customization, and performance.

The Bigger Picture

This partnership reflects a growing trend toward “sovereign AI”—locally deployed, regulated, and fully customizable models that allow organizations to unlock AI capabilities without sacrificing compliance or data protection.

By integrating Mistral AI’s high-performance models with Reply’s expertise in domain-specific customization, organizations gain a scalable path to deploy generative AI in operational environments, research settings, or highly regulated industries.

The Bottom Line

Reply and Mistral AI are demonstrating how enterprise-grade generative AI can be both high-performance and compliant. From mortgage and healthcare operations to specialized research like ancient Greek texts, the partnership shows that secure, customized, and scalable AI deployments are now achievable—without compromising data sovereignty or governance.

Get in touch with our MarTech Experts.