The Gilbane Advisor

Curated for content, computing, data, information, and digital experience professionals

Gilbane Advisor 5-6-26 — Language games, agent reality gap

May 6, 2026 /

Featured articles from: Stephanie Shen, and Cobus Greyling.

Additional reading from: Annalise Sumpon, Daniel Tunkelang, Rachel Draelos, and Michael Andrews.

News from: Microsoft, Deepgram, Acquia, Otter. ai.

Our next issue arrives 5/20/26.

All previous issues are available at https://gilbane.com/gilbane-advisor-index

Opinion / Analysis

The illusion of understanding

Wittgenstein, neuroscience, and why language is not thought in humans or AI

Stephanie Shen’s intriguing essay provides novel way to look at how humans and LLMs each “think”. Shen wrote this piece for a general audience; no philosophical, neuroscience, or AI expertise required. Every reader will learn something. Highly recommended. (12 min)

https://medium.com/illumination/the-illusion-of-understanding-8b083cc1e11e

The AI agent reality gap

Researching AI agents in production and the gap between demo and deploy

Cobus Greyling summarizes a study, Measuring Agents in Production, from UC Berkeley, Intesa Sanpaolo, UIUC, Stanford University, and IBM Research, that surveyed 306 practitioners and conducted 20 in-depth case studies across 26 domains. Greyling’s article design is a bit loud, but a way to quickly get the take-aways, some of which are surprising. Link to the paper below. (8 min)

Article: https://medium.com/@cobusgreyling/the-ai-agent-reality-gap-143c04136b5b
Paper (PDF): https://arxiv.org/pdf/2512.04123

More Reading

All Gilbane Advisor issues

Content Technology News

Microsoft 365 E7 and Agent 365 now generally available

Agent 365 is the control plane for agents, extending the visibility, control, and trust that organizations rely on for users to the agents they deploy.
https://www.microsoft.com/en-us/security/blog/2026/05/01/microsoft-agent-365-now-generally-available-expands-capabilities-and-integrations/

Deepgram launches Flux Multilingual

Flux Multilingual supports 10 languages in a single conversational model, enabling teams to build and deploy voice agents globally with one integration.
https://deepgram.com/learn/introducing-flux-multilingual

Acquia enhances Acquia Source

The unified intelligent workspace with the new Acquia AI allows digital teams manage content, applications, AI agents, and analytics in one place.
https://www.acquia.com/newsroom/press-releases/acquia-announces-significant-updates-acquia-source-and-launches-acquia-ai

Otter. ai launches enterprise Conversational Knowledge Engine

Knowledge Engine turns every meeting, call, and discussion into data that is searchable, cross-referenced, connected to your existing tools, and available to your AI systems.
https://otter.ai/blog/otter-ai-evolves-from-ai-notetaker-to-create-100b-enterprise-conversational-knowledge-engine-market

All content technology news

The Gilbane Advisor is authored by Frank Gilbane and is ad-free, cost-free, and curated for content, computing, data, web, and digital experience technology and information professionals. We publish recommended articles and content technology news. We do not sell or share personal data.

Subscribe | View online | Editorial policy | Privacy policy | Contact

Microsoft 365 E7 and Agent 365 are now generally available

May 1, 2026 /

From the Microsoft 365 Community blog…

Today, Microsoft 365 E7 – the Frontier Suite – is generally available, offering a single, integrated solution that unifies intelligence and trust, so organizations can move from experimentation to enterprise-wide AI adoption. Microsoft 365 E7 brings these capabilities together in one suite: 

  • Microsoft 365 Copilot: AI built for work
  • Microsoft 365 E5: enterprise‑grade productivity, security, identity, and compliance
  • Microsoft Entra Suite: identity and network access controls that secure employee use of all apps and AI
  • Microsoft Agent 365: the control plane for governing, observing, and securing AI agents at scale

Today also marks the general availability of Microsoft Agent 365. Agent 365 provides a consistent way to:

  • Observe: Gain visibility into agents in your environment, understand how they’re used, and act quickly on performance, behavior, and risk signals before they impact the business.
  • Govern: Establish guardrails for agents and people, onboard agents with IT oversight, and govern agent access to resources and data. Be audit ready with built-in compliance and data retention.
  • Secure: Secure agent identities, control access to resources, prevent data oversharing and leaks, and defend against threats and vulnerabilities with enterprise-grade security solutions.

Microsoft 365 E7 and Agent 365 are now generally available | Microsoft Community Hub

Deepgram launches Flux Multilingual

April 29, 2026 /

Deepgram, a real-time AI infrastructure company in the Voice AI economy, announced the general availability (GA) of Flux Multilingual, expanding its conversational speech recognition model beyond English to support: English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian, and Dutch, with the ability to automatically detect, understand, and switch languages dynamically within a single conversation in real time. Developers, enterprises, and product teams building voice agents now have access to a real-time conversational speech recognition model, delivering accurate turn-taking, interruption handling, low latency, and natural human-like conversations at global scale. 

Traditional automatic speech recognition (ASR) is designed for transcription. Flux introduced a new approach, conversational speech recognition (CSR), built from the ground up to understand dialogue flow and enable real-time interaction. With native support for turn-taking, interruptions, and code-switching within a single interaction, voice applications remain fluid, responsive, and natural regardless of language or region. Flux Multilingual delivers monolingual-grade accuracy across languages. Developers can guide the model with language hints or let it auto-detect, adapting in real time even mid-conversation.

Flux Multilingual is available via Deepgram’s Cloud API or as a self-hosted deployment. Developers can get started today at deepgram.com or try Flux Multilingual directly in the Deepgram Playground.

https://deepgram.com/learn/introducing-flux-multilingual

Acquia enhances Acquia Source

April 28, 2026 /

Acquia announced the next evolution of Acquia Source as a digital command center, a unified intelligent workspace where digital teams manage content, applications, AI agents, and analytics in one place. Acquia also announced Acquia AI, built into Acquia Source, giving builders the tools to create, manage, and deploy AI agents that optimize content, automate governance, and drive AI discoverability.

Acquia Source unifies content management, digital asset management, and web governance into a single agentic command center, aggregating insights across every layer of the content supply chain, surfacing what needs attention, and deploying AI to resolve it. The result reduces operational overhead that disconnected systems impose on enterprise marketing and technical teams.

Acquia Source addresses discoverability at the architectural level. Content is structured for both human engagement and LLM readabilit, ensuring organizations remain the authoritative answer in ChatGPT, Perplexity, and Google AI Overviews. Answer Engine Optimization (AEO) is built directly into the content creation workflow.

Acquia AI enables organizations to build a portfolio of purpose-built agents or deploy Acquia’s ready-made agents, which come with built-in knowledge and domain-specific skills. These capabilities deliver cross-channel ROI intelligence and insight-to-action workflows that help organizations to make faster, more informed decisions.

https://www.acquia.com/newsroom/press-releases/acquia-announces-significant-updates-acquia-source-and-launches-acquia-ai

Otter. ai launches enterprise Conversational Knowledge Engine

April 28, 2026 /

Otter.ai announced the launch of a new Conversational Knowledge Engine, a platform to connect conversations across teams and time, turning what’s said in meetings into structured, searchable knowledge that can drive agentic actions automatically. The launch is supported by three new product capabilities, an expanded MCP server, AI Chat and Otter for Desktop.

The Conversational Knowledge Engine solves a long-standing gap in enterprise software where the decisions, context, and intent of conversational data are lost the moment a meeting ends. Otter.ai’s proprietary Conversational Knowledge Engine can be broken down into three connected ideas: 

  • Conversational. Mapped to meetings and voice, where the majority of enterprise knowledge is actually created.
  • Knowledge. A longitudinal knowledge graph (a continuously growing record of conversations over time), connecting thousands of conversations across thousands of people over time, linking decisions to who made them, the context around them, and the workflows that depend on them.
  • Engine. Puts that graph to work through conversational agentic workflows, where conversations become the trigger for automated action across the enterprise.

All new features are available today.

https://otter.ai/blog/otter-ai-evolves-from-ai-notetaker-to-create-100b-enterprise-conversational-knowledge-engine-market

Gilbane Advisor 4-22-26 — Flight simulator, neurosymbolic AI

April 22, 2026 /

Featured articles from:  Rohit Krishnan, and Gary Marcus.

Additional reading from: Michael Andrews, Kurt Cagle & Chloe Shannon, Stanford HAI, Ship X/ TechX.

News from: Adobe, Canva, Tiny Technologies, Pantheon.

Our next issue arrives 5/6/26.

All previous issues are available at https://gilbane.com/gilbane-advisor-index

Opinion / Analysis

Can we build a management flight simulator?

Turns out, yes

This piece is a follow-up from Rohit Krishnan’s article last month on world models, in particular enterprise world models. He has now built an experimental world model flight simulator for testing practical applications for large (and small) companies, and illustrates its use with real public data from the Enron scandal. Really interesting. (8 min)

https://www.strangeloopcanon.com/p/can-we-build-a-management-flight

Even more good news for the future of neurosymbolic AI

And vindication for Apple’s unfairly maligned 2025 reasoning paper

Gary Marcus celebrates a new paper from Tufts that backs up Apple’s earlier research on LLM reasoning, and a showed a neurosymbolic hybrid out-performed LLMs, including VLAs (Vision-Language-Action models) (4 min)

https://garymarcus.substack.com/p/even-more-good-news-for-the-future

More Reading

All Gilbane Advisor issues

Content Technology News

Tiny Technologies debuts TinyMCE AI

TinyMCE AI gives content teams everything they need to write, refine and review conversational AI, text transformations without leaving the editor.
https://www.tiny.cloud/tinymce/

Adobe unveils Firefly AI Assistant with new creative agent

Creators direct outcomes as the assistant orchestrates and executes multi-step workflows across Creative Cloud apps in a single conversational interface.
https://news.adobe.com/news/2026/04/adobe-new-creative-agent

Canva introduces Canva AI 2.0

New architecture layer spans conversational design, iterative agentic editing, layered object intelligence, and living memory, alongside six new workflows.
https://www.canva.com/newsroom/news/canva-create-2026-ai/

Pantheon adds Next.js to WebOps platform

The platform for WordPress and Drupal now runs Next.js, reducing multi-vendor complexity, unifying IT, developers, and marketers around a single system.
https://pantheon.io/platform/nextjs

All content technology news

The Gilbane Advisor is authored by Frank Gilbane and is ad-free, cost-free, and curated for content, computing, data, web, and digital experience technology and information professionals. We publish recommended articles and content technology news. We do not sell or share personal data.

Subscribe | View online | Editorial policy | Privacy policy | Contact

Tiny Technologies debuts TinyMCE AI

April 16, 2026 /

Tiny Technologies announced the release and general availability of TinyMCE AI, a fully integrated AI writing environment built into the editor.

TinyMCE AI gives content teams everything they need to write, refine and review conversational AI, instant text transformations and automated quality checks without leaving the editor. This addition to the platform means that developers can meet user demand for robust AI content capabilities without adding a technical burden to application teams or imposing model lock-in.

With the new functionalities, developers can enable enterprise-grade AI features via a simple drop-in module. Content teams gain a collaborative writing partner that helps with research, understands the context of a full document, suggests changes through familiar editing markups, adapts to brand guidelines through custom prompts, and automatically catches quality issues.

Starting with version 8.4, key features include:

  • Conversational AI Chat: Enables multi-turn, natural-language conversations with awareness of the active document, and the ability to add additional context via web searches and file attachments.
  • Context-Aware Quick Actions: Applies rewriting, expansion, shortening and tone adjustments.
  • AI Review: Runs automated quality checks; delivers inline suggestions to improve clarity, consistency and accuracy.
  • Custom Prompts: Organizations can define and enforce brand voice, style guidelines, and content standards.

https://www.tiny.cloud/tinymce

Canva introduces Canva AI 2.0

April 15, 2026 /

Canva introduced Canva AI 2.0, transforming Canva into a conversational, agentic platform where teams can go from the spark of an idea to complete, published work in one place.

Powered by the Canva Design Model – a foundation model built to understand the structure, hierarchy, and complexity of real-world design – Canva AI 2.0 generates layered, editable output from a single prompt, and gets smarter the more you use it.

Available as a research preview, Canva AI 2.0 introduces a new architecture layer spanning conversational design, iterative agentic editing, layered object intelligence, and living memory, alongside six new intelligent workflows: connectors, scheduling, web research, brand intelligence, Sheets AI, and Canva Code 2.0.

Conversational design Describe an idea, goal, or rough structure, and Canva AI generates an editable design with layout, hierarchy, and brand, then stays with you throughout the process.

Agentic orchestration gives Canva AI access to Canva’s complete design engine, calling the right tools at the right time.

Layered object intelligence everything Canva AI generates is built from scratch using individual, editable objects.

Memory Library turns interactions into building blocks. With persistent memory, Canva AI 2.0 understands how you work, keeps every project on brand, and applies your style automatically.

https://www.canva.com/newsroom/news/canva-create-2026-ai

« Older posts

Policies

Contact

twitter   rss   Mastadon   Bluesky   Medium

Contact • customerservice@gilbane.com

© 2026 Bluebill Advisors (except for contributed posts which are licensed under creative commons)

Banner photos: Sakonnet Harbor, Little Compton, RI, Fort Point Channel, Summer St bridge, and Northern Ave bridge, Boston

Bluebill Advisors Inc logo

© 2026 The Gilbane Advisor

Theme by Anders Noren — Up ↑