Blog posts by Opti Chronicles

Blog posts by Opti Chronicles2026-04-14T19:50:12.0000000Zhttps://world.optimizely.com/blogs/opti-chronicles/ Optimizely World Optimizely CMS 13: A Strategic Reset for Content, AI, and Composable MarketingOptimizely CMS 13 is not just another version upgrade—it represents a deliberate shift toward a connected, AI-enabled, and API-driven content platform. Released March 31, 2026, CMS 13 introduces: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l20 level1 lfo1; tab-stops: list .5in;">A Graph-first architecture for content delivery</li> <li class="MsoNormal" style="mso-list: l20 level1 lfo1; tab-stops: list .5in;">Embedded AI workflows via Opal</li> <li class="MsoNormal" style="mso-list: l20 level1 lfo1; tab-stops: list .5in;">A Visual Builder for marketer-led content creation</li> <li class="MsoNormal" style="mso-list: l20 level1 lfo1; tab-stops: list .5in;">A standardized approach to integrating external systems</li> </ul> This isn’t about improving the CMS editor, it is about redefining how content operates across your entire digital ecosystem. <h2 class="MsoNormal">Market Context: The CMS Is No Longer the Center</h2> Most enterprise CMS platforms were built for: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l10 level1 lfo2; tab-stops: list .5in;">Page-based publishing</li> <li class="MsoNormal" style="mso-list: l10 level1 lfo2; tab-stops: list .5in;">Centralized content storage</li> <li class="MsoNormal" style="mso-list: l10 level1 lfo2; tab-stops: list .5in;">Custom integrations for everything else</li> </ul> That model is increasingly misaligned with reality. Today: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l17 level1 lfo3; tab-stops: list .5in;">Content lives across CMPs, DAMs, commerce platforms, and PIMs</li> <li class="MsoNormal" style="mso-list: l17 level1 lfo3; tab-stops: list .5in;">Experiences span websites, apps, marketplaces, and AI interfaces</li> <li class="MsoNormal" style="mso-list: l17 level1 lfo3; tab-stops: list .5in;">Marketing teams need speed without engineering dependency</li> <li class="MsoNormal" style="mso-list: l17 level1 lfo3; tab-stops: list .5in;">AI requires structured, accessible content across systems</li> </ul> The implication is clear: The CMS is no longer the system of record—it’s the orchestrator of distributed content. CMS 13 is Optimizely’s answer to that shift. <h2 class="MsoNormal">What Is Optimizely Graph?</h2> Optimizely Graph is the central layer that connects, indexes, and delivers content across your ecosystem. Think of it as a unified content engine that sits between your systems (CMS, DAM, commerce, etc.) and your channels (web, apps, AI). <h3 class="MsoNormal">Why this matters?</h3> 1. Content is no longer siloed Graph enables content from multiple systems to be accessed and used together in a single experience. 2. It replaces legacy search and delivery models CMS 13 deprecates Search & Navigation and positions Graph as the primary content delivery mechanism. 3. It powers AI across the platform Opal’s AI capabilities—including content creation and optimization—depend on Graph indexing and understanding your content. 4. It improves performance and scalability Graph acts as a high-performance delivery layer, enabling faster and more consistent content retrieval. <h3 class="MsoNormal">The Strategic Shift</h3> Before: CMS → APIs → Channels Now: Graph → unified content → channels + AI + integrations This is not a technical distinction—it’s a platform-level change in how content is accessed and activated. <h2 class="MsoNormal">The Three Strategic Shifts in CMS 13</h2> Optimizely frames CMS 13 around three outcomes: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l16 level1 lfo4; tab-stops: list .5in;">Faster content creation</li> <li class="MsoNormal" style="mso-list: l16 level1 lfo4; tab-stops: list .5in;">Better delivery capabilities</li> <li class="MsoNormal" style="mso-list: l16 level1 lfo4; tab-stops: list .5in;">AI-powered workflows</li> </ul> Let’s break it down to what that means. 1. Visual Builder: Marketing Speed Without Developer Dependency CMS 13 introduces a Visual Builder that enables marketers to create experiences directly. <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l0 level1 lfo5; tab-stops: list .5in;">Grid-based layout system (rows, columns, sections)</li> <li class="MsoNormal" style="mso-list: l0 level1 lfo5; tab-stops: list .5in;">Real-time preview across breakpoints</li> <li class="MsoNormal" style="mso-list: l0 level1 lfo5; tab-stops: list .5in;">Reusable templates (“blueprints”)</li> <li class="MsoNormal" style="mso-list: l0 level1 lfo5; tab-stops: list .5in;">Governance through predefined styling controls</li> </ul> From a business perspective: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l21 level1 lfo6; tab-stops: list .5in;">Campaigns launch faster</li> <li class="MsoNormal" style="mso-list: l21 level1 lfo6; tab-stops: list .5in;">Less reliance on development teams</li> <li class="MsoNormal" style="mso-list: l21 level1 lfo6; tab-stops: list .5in;">Brand consistency is maintained through guardrails</li> </ul> Existing CMS 12 content still works, teams can adopt this incrementally. (With some required updates and changes, e.g. Framework updated to .NET10.0, check DB compatibility, Optimizely Graph is needed to use CMS 13, among others) 2. Graph-First Delivery: The Foundation for Everything Else Graph is no longer optional, it’s the backbone of CMS 13. <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l23 level1 lfo7; tab-stops: list .5in;">Powers search, APIs, and content delivery</li> <li class="MsoNormal" style="mso-list: l23 level1 lfo7; tab-stops: list .5in;">Enables semantic search and AI use cases</li> <li class="MsoNormal" style="mso-list: l23 level1 lfo7; tab-stops: list .5in;">Replaces legacy search infrastructure</li> <li class="MsoNormal" style="mso-list: l23 level1 lfo7; tab-stops: list .5in;">Standardizes how content is accessed across channels</li> </ul> This aligns CMS 13 with modern architectures where: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l6 level1 lfo8; tab-stops: list .5in;">Content is API-driven</li> <li class="MsoNormal" style="mso-list: l6 level1 lfo8; tab-stops: list .5in;">Frontends are decoupled</li> <li class="MsoNormal" style="mso-list: l6 level1 lfo8; tab-stops: list .5in;">Experiences are assembled dynamically</li> </ul> 3. External Content: Composable CMS in Practice CMS 13 introduces a new external content model, enabling content from other systems to be used directly. <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l14 level1 lfo9; tab-stops: list .5in;">Integrations with commerce, PIM, DAM, and more</li> <li class="MsoNormal" style="mso-list: l14 level1 lfo9; tab-stops: list .5in;">Content remains in the source system (no duplication)</li> <li class="MsoNormal" style="mso-list: l14 level1 lfo9; tab-stops: list .5in;">Real-time updates reflected in experiences</li> <li class="MsoNormal" style="mso-list: l14 level1 lfo9; tab-stops: list .5in;">Standardized integration via Optimizely Connect Platform</li> </ul> This eliminates a long-standing problem: Custom integrations that become outdated and expensive to maintain. 4. Embedded DAM: From Asset Storage to Asset Strategy CMS 13 includes an enterprise-grade DAM directly in the CMS experience. <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l2 level1 lfo10; tab-stops: list .5in;">Centralized asset management</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo10; tab-stops: list .5in;">Metadata, tagging, and renditions</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo10; tab-stops: list .5in;">Unified asset access across workflows</li> </ul> This reduces operational friction: fewer tools, fewer handoffs, and better governance. 5. Opal AI: Embedded, Not Bolted On CMS 13 integrates Opal AI directly into editorial workflows. Opal enables: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l13 level1 lfo11; tab-stops: list .5in;">Content creation and publishing directly within CMS</li> <li class="MsoNormal" style="mso-list: l13 level1 lfo11; tab-stops: list .5in;">SEO & AEO optimization and automated recommendations</li> <li class="MsoNormal" style="mso-list: l13 level1 lfo11; tab-stops: list .5in;">Content management operations (create, update, manage variations)</li> </ul> Additional capabilities include metadata and schema generation, content audits and refresh recommendations, and AI-assisted content modeling and page creation. The key shift: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l4 level1 lfo12; tab-stops: list .5in;">AI is no longer a side tool</li> <li class="MsoNormal" style="mso-list: l4 level1 lfo12; tab-stops: list .5in;">It becomes part of how content gets created and optimized daily</li> </ul> <h2 class="MsoNormal">How It All Fits Together?</h2> CMS 13 operates across three coordinated layers: 1. Creation Layer <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l3 level1 lfo13; tab-stops: list .5in;">Visual Builder</li> <li class="MsoNormal" style="mso-list: l3 level1 lfo13; tab-stops: list .5in;">CMP integration</li> <li class="MsoNormal" style="mso-list: l3 level1 lfo13; tab-stops: list .5in;">Embedded DAM</li> </ul> 2. Content Layer <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l19 level1 lfo14; tab-stops: list .5in;">Graph as the unified index</li> <li class="MsoNormal" style="mso-list: l19 level1 lfo14; tab-stops: list .5in;">External content integration</li> <li class="MsoNormal" style="mso-list: l19 level1 lfo14; tab-stops: list .5in;">Structured content models</li> </ul> 3. Intelligence Layer <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l5 level1 lfo15; tab-stops: list .5in;">Opal AI agents</li> <li class="MsoNormal" style="mso-list: l5 level1 lfo15; tab-stops: list .5in;">Semantic search and analytics</li> <li class="MsoNormal" style="mso-list: l5 level1 lfo15; tab-stops: list .5in;">AI-driven optimization workflows</li> </ul> The implication is that content becomes structured, connected, and AI-readable by design. <h2 class="MsoNormal">What This Means for Different Stakeholders?</h2> Marketing Leaders <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l11 level1 lfo16; tab-stops: list .5in;">Faster campaign execution</li> <li class="MsoNormal" style="mso-list: l11 level1 lfo16; tab-stops: list .5in;">AI-assisted content production</li> <li class="MsoNormal" style="mso-list: l11 level1 lfo16; tab-stops: list .5in;">Consistent brand governance</li> </ul> Digital Experience Managers <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l22 level1 lfo17; tab-stops: list .5in;">Unified content access across systems</li> <li class="MsoNormal" style="mso-list: l22 level1 lfo17; tab-stops: list .5in;">Improved discoverability and reuse</li> <li class="MsoNormal" style="mso-list: l22 level1 lfo17; tab-stops: list .5in;">Reduced operational complexity</li> </ul> Engineering Leaders <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l8 level1 lfo18; tab-stops: list .5in;">Less custom integration work</li> <li class="MsoNormal" style="mso-list: l8 level1 lfo18; tab-stops: list .5in;">Standardized APIs via Graph</li> <li class="MsoNormal" style="mso-list: l8 level1 lfo18; tab-stops: list .5in;">More focus on architecture, less on maintenance</li> </ul> <h2 class="MsoNormal">Implementation Considerations</h2> 1. This Is a Replatforming Decision While CMS 12 → 13 is described as an evolution, the reality is: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l18 level1 lfo19; tab-stops: list .5in;">Graph adoption is required</li> <li class="MsoNormal" style="mso-list: l18 level1 lfo19; tab-stops: list .5in;">Search must be migrated</li> <li class="MsoNormal" style="mso-list: l18 level1 lfo19; tab-stops: list .5in;">Architecture changes impact delivery models</li> </ul> 2. Content Modeling Must Evolve To unlock value: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l12 level1 lfo20; tab-stops: list .5in;">Move from page-centric to component-based design</li> <li class="MsoNormal" style="mso-list: l12 level1 lfo20; tab-stops: list .5in;">Structure content for reuse and AI consumption</li> </ul> 3. Cloud Becomes the Default <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l9 level1 lfo21; tab-stops: list .5in;">Graph is a SaaS service</li> <li class="MsoNormal" style="mso-list: l9 level1 lfo21; tab-stops: list .5in;">Opal and DAM depend on cloud capabilities</li> <li class="MsoNormal" style="mso-list: l9 level1 lfo21; tab-stops: list .5in;">Opti ID is required for key features</li> </ul> 4. AI Governance Is Critical Organizations must define: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l15 level1 lfo22; tab-stops: list .5in;">Approval workflows for AI-generated content</li> <li class="MsoNormal" style="mso-list: l15 level1 lfo22; tab-stops: list .5in;">Brand and compliance guardrails</li> <li class="MsoNormal" style="mso-list: l15 level1 lfo22; tab-stops: list .5in;">Data access policies for AI agents</li> </ul> <h2 class="MsoNormal">Risks and Trade-offs</h2> Transition Complexity: New architecture, new workflows, new mindset. Dependency on Graph: Graph is foundational, there is no fallback to legacy models. Organizational Readiness: Success depends on people and process, not just technology. Cloud Constraints: On-premise environments face limitations for full feature adoption. <h2 class="MsoNormal">Key Decisions</h2> CMS 13 reflects a broader industry shift: From CMS as a publishing tool → to CMS as a content orchestration platform The real question for decision-makers is not: “Should we upgrade?” It’s: “Are we ready to operate in a connected, AI-driven content ecosystem?” <h2 class="MsoNormal">Final Takeaway</h2> CMS 13 is not incremental, it’s directional. It aligns with where MarTech is going: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l1 level1 lfo23; tab-stops: list .5in;">API-first architectures</li> <li class="MsoNormal" style="mso-list: l1 level1 lfo23; tab-stops: list .5in;">AI embedded in workflows</li> <li class="MsoNormal" style="mso-list: l1 level1 lfo23; tab-stops: list .5in;">Composable ecosystems over monolithic platforms</li> </ul> The organizations that win won’t just adopt CMS 13. They’ll use it to rethink how content flows across their entire business. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">References:</h2> <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/cms-13-overview">https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/cms-13-overview</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://support.optimizely.com/hc/en-us/articles/44937048830221-2026-Optimizely-CMS-13-release-notes">https://support.optimizely.com/hc/en-us/articles/44937048830221-2026-Optimizely-CMS-13-release-notes</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/get-started-with-cms-for-users">https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/get-started-with-cms-for-users</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/overview-of-c-sharp-sdk">https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/overview-of-c-sharp-sdk</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/optimizely-opal-in-cms-13">https://docs.developers.optimizely.com/content-management-system/v13.0.0-CMS/docs/optimizely-opal-in-cms-13</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://www.msqdx.com/en/insights/optimizely-cms-13-what-it-actually-means-for-your-marketing-team">https://www.msqdx.com/en/insights/optimizely-cms-13-what-it-actually-means-for-your-marketing-team</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://www.matthewdamon.com/optimizely-cms-13-what-partners-need-to-know-before-the-upgrade/">https://www.matthewdamon.com/optimizely-cms-13-what-partners-need-to-know-before-the-upgrade/</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a href="https://wseweryn.dev/blog/2026-04-07-optimizely-cms-13-graph-migration/">https://wseweryn.dev/blog/2026-04-07-optimizely-cms-13-graph-migration/</a></li> <li class="MsoNormal" style="mso-list: l7 level1 lfo24; tab-stops: list .5in;"><a style="font-family: Aptos, sans-serif; font-size: 12pt;" href="https://www.rollin.ca/resources/optimizely-visual-builder-complete-implementation-guide/">https://www.rollin.ca/resources/optimizely-visual-builder-complete-implementation-guide/</a></li> </ul>2026-04-14T19:50:12.0000000Z

Blog post

My Takeaway from Optimizely Opal Agents in Action 2026 - What Agentic AI Means for the Future of Digital MarketingI would like to share with you what stayed in my head after this amazing virtual event organized by Optimizely. Agents in Action 2026, a live demonstration of Optimizely's AI assistant platform, Opal. The session showcased how AI agents embedded across the Optimizely ecosystem can conduct research, analyze documents, orchestrate tools, and automate operational workflows for marketing teams. Artificial intelligence in marketing is entering a new phase. For the past two years, the focus has largely been on generative AI—tools that produce copy, images, summaries, and marketing ideas. But as organizations begin integrating these technologies into their workflows, a more powerful model is emerging: agentic AI. Agentic systems go beyond answering questions. They can plan tasks, call tools, retrieve data, analyze information, and execute multi-step workflows. Optimizely's demonstration highlighted a shift that is beginning to reshape digital marketing operations: the move from isolated AI features to AI agents that perform work inside enterprise platforms. Industry research suggests the timing is not accidental. Organizations are rapidly experimenting with AI but still struggle to integrate it deeply into operational workflows. According to McKinsey’s latest global survey on AI adoption, many companies have begun using AI widely but have not yet embedded it deeply enough into processes to generate full enterprise value (<a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai">McKinsey</a>). What Optimizely showed with Opal agents points toward the next stage of AI maturity—where marketing teams work alongside AI agents that automate research, analysis, and operational tasks. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">The Rise of Agentic AI in Marketing</h2> Most early generative AI tools behaved like conversational assistants. You ask a question, yhe system produces an answer. That model has proven useful for content generation, brainstorming, and summarization, but enterprise environments require more than output generation. They require systems that can execute processes. Agentic AI systems are designed for that purpose. Instead of generating a single response, they can: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Execute multi-step research</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Retrieve external information</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Call APIs or tools</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Analyze documents</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Trigger workflows</li> <li class="MsoNormal" style="mso-list: l2 level1 lfo1; tab-stops: list .5in;">Coordinate multiple systems</li> </ul> This shift reflects a broader trend in enterprise AI adoption. McKinsey research shows that AI is increasingly used across multiple business functions—including marketing and sales—where generative AI can support activities such as knowledge synthesis, content generation, and strategic analysis (<a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024">McKinsey</a><a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024)">)</a>. But many organizations still struggle to scale these capabilities across their workflows. The opportunity lies not in adding more AI tools, but in redesigning operational processes so AI becomes embedded within them. This is where agentic platforms begin to matter, instead of interacting with AI occasionally, teams interact with systems that continuously assist, automate, and coordinate operational tasks. The Optimizely Opal demonstration illustrated how such systems might work in a marketing technology environment. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">What Optimizely Demonstrated at Opal Agents in Action 2026</h2> The event focused on Opal, Optimizely’s AI assistant platform integrated across its product suite, including CMS, experimentation tools, analytics capabilities, and content marketing workflows. At its core, Opal operates as a conversational AI interface embedded directly inside the Optimizely ecosystem. Users interact with the system through a chat interface that maintains persistent conversation history and contextual awareness. Before each interaction with the underlying language model, the platform executes a multi-step enrichment process. This pipeline loads contextual information such as: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l0 level1 lfo2; tab-stops: list .5in;">organizational instructions and brand guidelines</li> <li class="MsoNormal" style="mso-list: l0 level1 lfo2; tab-stops: list .5in;">available agents and tools</li> <li class="MsoNormal" style="mso-list: l0 level1 lfo2; tab-stops: list .5in;">operational prompts or system rules</li> </ul> Organizations can configure these instructions at multiple levels. Instance-level instructions apply across the entire organization—for example, defining brand voice or communication guidelines. Personal instructions allow individual users to tailor how the system responds to their requests. Role-based permissions control who can modify these instructions, which helps maintain governance in enterprise environments. The demonstration showed how this contextual layer ensures AI interactions remain aligned with organizational policies and brand requirements. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">Key Capabilities of Opal Agents</h2> The most interesting aspect of the event was not the conversational interface. It was the demonstration of specialized AI agents capable of executing complex operations. <h3 class="MsoNormal">Specialized agents</h3> Agents in Opal are created with defined roles, instructions, and tool access. When an agent runs, it operates in an isolated context. This prevents large multi-step processes from overwhelming the main conversation. During the demonstration, a research agent was configured to analyze companies. The agent accepted a topic parameter and automatically executed multiple research passes.  In another example, the agent performed several research queries about Toyota, gathering information about company background, news, and financial performance. The system automatically executed a chain of operations: <ul> <li class="MsoNormal">search → scrape → analyze → summarize</li> <li class="MsoNormal">The process generated a detailed report after approximately 234 seconds and involved roughly 44,000 tokens across multiple tool calls.</li> <li class="MsoNormal">Execution logs displayed the reasoning cycle in real time, including predicted tool calls and returned outputs.</li> </ul> The demonstration highlighted how AI agents can handle research tasks that typically require multiple manual steps. <h3 class="MsoNormal">Tools and extensibility</h3> A second important capability is Opal’s tool framework. Developers can create custom tools using the Opal Tools SDK, which supports Python and FastAPI. Functions can be exposed as tools through simple annotations, allowing the AI system to invoke them during task execution. The demonstration included a simple tool that generated random numbers based on user parameters. While trivial in itself, the example illustrated how tools return structured responses that can render as interactive user interface elements. More advanced tools are provided directly by Optimizely product teams. Examples demonstrated during the session included: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l1 level1 lfo3; tab-stops: list .5in;">Web search integration that performs Google searches and scrapes relevant websites</li> <li class="MsoNormal" style="mso-list: l1 level1 lfo3; tab-stops: list .5in;">PowerPoint processing that converts slides into analyzable formats</li> <li class="MsoNormal" style="mso-list: l1 level1 lfo3; tab-stops: list .5in;">Product-specific tools for CMS and experimentation features</li> </ul> The system determines when to call these tools based on their descriptions and parameter schemas. The result is an AI system capable of chaining together multiple operations without manual intervention. <h3 class="MsoNormal">Enterprise document analysis</h3> In the event it was also demonstrated document processing capabilities. <ul> <li class="MsoNormal">PowerPoint files can be uploaded and converted into a structured Canvas format, allowing the system to analyze slides individually.</li> <li class="MsoNormal">A code execution tool—similar to those used in other advanced AI systems—will enable analysis of spreadsheets and other Office documents through generated Python scripts.</li> </ul> This allows agents to read, manipulate, and extract information from enterprise files, so organizations can also attach reference documents or knowledge sources to agents. These materials become searchable through retrieval-augmented generation workflows using Optimizely’s content graph. <h3 class="MsoNormal">Workflow automation</h3> For me the most operationally significant feature demonstrated during the event was multi-agent workflow orchestration. Agents can be chained together sequentially, passing outputs and parameters between steps. Workflows can be triggered in several ways: <ul style="margin-top: 0in;"> <li class="MsoNormal" style="mso-list: l3 level1 lfo4; tab-stops: list .5in;">Direct chat input</li> <li class="MsoNormal" style="mso-list: l3 level1 lfo4; tab-stops: list .5in;">Email-based triggers</li> <li class="MsoNormal" style="mso-list: l3 level1 lfo4; tab-stops: list .5in;">Webhook integrations</li> <li class="MsoNormal" style="mso-list: l3 level1 lfo4; tab-stops: list .5in;">Scheduled automation</li> </ul> One demonstration showed a research agent analyzing a topic and passing its results to an email agent. The second agent transformed the research findings into a formatted HTML email and automatically delivered it to recipients. Other enterprise use cases discussed during the event included automated market research, competitive intelligence monitoring, and recurring insight reports. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h3 class="MsoNormal">Why Agentic Marketing Matters for Digital Teams</h3> For marketing teams, the operational implications of agentic AI could be substantial. Modern digital marketing involves a constant flow of analysis, reporting, and information gathering. Many tasks follow similar patterns: <ul> <li class="MsoNormal">Collect information → Analyze it → Generate output → Share results </li> </ul> AI agents can automate much of this process instead of manually conducting research or assembling reports, teams can initiate workflows where agents gather information, analyze it, and deliver structured insights. This does not remove human involvement, instead, it shifts where humans spend their time. MIT Sloan research suggests that AI is most effective when it complements human work rather than replacing it, enabling people to focus on higher-value tasks such as interpretation and decision-making (<a href="https://mitsloan.mit.edu/ideas-made-to-matter/when-humans-and-ai-work-best-together-and-when-each-better-alone">MIT Sloan Management Review</a><a href="https://mitsloan.mit.edu/ideas-made-to-matter/when-humans-and-ai-work-best-together-and-when-each-better-alone)">)</a>. For marketing organizations, that means less time spent on operational preparation and more time spent on strategy. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">Industry Validation</h2> Several industry studies reinforce the potential impact of AI in marketing operations. McKinsey estimates that generative AI could generate significant economic value across industries, with marketing and sales among the functions that stand to benefit the most from productivity improvements (<a href="https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/how-generative-ai-can-boost-consumer-marketing">McKinsey</a>). The firm’s research suggests that generative AI could increase marketing productivity up to 15 percent of total marketing spend through improvements in content creation, personalization, and knowledge synthesis. At the same time, organizations are still learning how to operationalize AI effectively. McKinsey’s latest global AI survey notes that many companies have adopted AI tools but have not yet embedded them deeply into their workflows or processes (<a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024">McKinsey</a>). This gap between experimentation and operational integration is precisely where agentic platforms may deliver value. By embedding AI into enterprise systems and allowing agents to execute tasks, organizations can move from isolated experiments to operational transformation. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">Strategic Implications for CMOs and Digital Leaders</h2> For marketing leaders evaluating AI platforms, several strategic considerations emerge from the Opal demonstration. First, governance becomes critical. AI systems embedded in operational workflows must operate within brand, legal, and compliance boundaries. Opal’s instruction framework illustrates how organizations can define these constraints. Second, the ecosystem around AI matters as much as the AI itself. Platforms that support tool integration, APIs, and workflow orchestration will likely deliver more long-term value than standalone assistants. Third, marketing operating models may evolve. As agent-driven workflows mature, teams may increasingly focus on oversight, strategy, and decision-making rather than manual research or analysis. Finally, human supervision remains essential. The demonstration itself acknowledged that large language models can occasionally drift off topic during complex research tasks, requiring prompt refinement and oversight. Agentic systems should therefore be viewed as collaborators, not replacements. <div class="MsoNormal" style="text-align: center;" align="center"><hr /></div> <h2 class="MsoNormal">My Final Takeaways</h2> The Optimizely Opal Agents in Action event offered an early look at how AI agents might reshape marketing operations. Rather than presenting AI simply as a content generator, the demonstrations positioned Opal as a system capable of executing research, analyzing documents, orchestrating tools, and automating workflows across marketing platforms. This approach reflects a broader industry shift. As organizations move beyond isolated AI experiments, the focus is shifting toward operational integration—embedding AI into everyday workflows. For digital marketing teams, this may represent the next stage of platform evolution, not just tools, not just assistants, but ecosystems of AI agents working alongside human teams to execute the operational layer of marketing.  2026-03-06T15:48:32.0000000Z

Blog post