The Gilbane Advisor

Curated for content, computing, and digital experience professionals

Gilbane Advisor 5-21-25 — AI user penalty, AI energy math

May 21, 2025 /

This week we feature articles from Jessica A. Reif, Richard P. Larrick, & Jack B. Soll, and James O’Donnell & Casey Crownhart.

Additional reading comes from Ksenia Se, Pavel Samsonov, and Cobus Greyling.

News comes from Perplexity & PayPal, Foxit, Databricks & Neon, and Box.

Our next issue arrives June 4.

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

Opinion / Analysis

Evidence of a social evaluation penalty for using AI

“Through four experiments with over 4,400 participants, we reveal a social penalty for AI use: Individuals who use AI tools face negative judgments about their competence and motivation from others. These judgments manifest as both anticipated and actual social penalties, creating a paradox where productivity-enhancing AI tools can simultaneously improve performance and damage one’s professional reputation.”

Surprise – not surprise, but now we have good data thanks to Jessica A. Reif, Richard P. Larrick, & Jack B. Soll. (28 min)

https://www.pnas.org/doi/full/10.1073/pnas.2426766122

We did the math on AI’s energy footprint. Here’s the story you haven’t heard.

“… new analysis by MIT Technology Review provides an unprecedented and comprehensive look at how much energy the AI industry uses—down to a single query—to trace where its carbon footprint stands now, and where it’s headed, as AI barrels towards billions of daily users.”

This deep dive by James O’Donnell & Casey Crownhart is likely the most thorough and current you will find. (23 min)

https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech

More Reading

All Gilbane Advisor issues

Content technology news

Perplexity and PayPal partner

Consumers can check out instantly with PayPal or Venmo when they ask Perplexity to find products, book travel, or buy tickets.
https://www.perplexity.ai/shopping ■ https://newsroom.paypal-corp.com/2025-05-14-Perplexity-Selects-PayPal-to-Power-Agentic-Commerce

Foxit launches enterprise Smart Redact Server

Enhances redaction technology with batch document processing, scheduled compliance scans, multi-cloud integration, and team collaboration tools.
https://foxit.com/en/

Databricks to acquire Neon

Databricks and Neon will remove the traditional limitations of databases that require compute and storage to scale in tandem.
https://www.databricks.com/company/newsroom/press-releases/databricks-agrees-acquire-neon-help-developers-deliver-ai-systems ■ https://neon.tech

Box announces AI Platform to bring AI agents to enterprise content

Customers will be able to leverage AI agents for Search, Deep Research, and enhanced data extraction to get more value from content in Box.
https://blog.box.com/content-ai-virtual-summit-may-2025

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 most Wednesdays. We do not sell or share personal data.

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

Box announces AI Platform to bring AI Agents to enterprise content

May 15, 2025 /

Box, Inc., an Intelligent Content Management (ICM) platform, announced the next evolution of its Box AI with a new generation of AI Agents designed to transform how organizations work with content. Customers will be able to leverage AI Agents for Search, Deep Research, and enhanced data extraction to uncover more value from their content in Box. Box also introduced a new Box AI Agent for Microsoft 365 Copilot that allows users to search, analyze, and act on Box content directly within Microsoft 365.

Box AI Agents are built with a layered approach, starting with enterprise knowledge found in Box, the Agents then leverage AI models from Amazon, Anthropic, Google, IBM, Meta, OpenAI, and xAI to best match the task’s demands. Every Agent is aligned to an objective, and follows a set of structured instructions that guide its actions. Users will be able to interact with Box AI Agents to extract insights from content and automate enterprise workflows. Box AI Agents can also be customized in Box AI Studio. 

Box AI Agents will be able to analyze large volumes of enterprise content to extract meaning and trends — automatically identifying the most relevant files using Box’s secure, permission-aware, retrieval-augmented generation (RAG) framework. 

https://blog.box.com/content-ai-virtual-summit-may-2025

Perplexity and PayPal partner

May 14, 2025 /

Perplexity today announced that it has partnered with PayPal to power agentic commerce across its Perplexity Pro platform. Starting this summer in the U.S., consumers can check out instantly with PayPal or Venmo when they ask Perplexity to find products, book travel, or buy tickets.

The entire process, including payment, shipping, tracking, and invoicing will be handled behind the scenes with PayPal’s account linking, secure tokenized wallet and emerging passkey checkout flows, which could eliminate the need for passwords and streamline the experience to a single user query or click. Features include:

  • Agentic Commerce: Integration of PayPal’s commerce solutions, enabling users to buy products or services directly in Perplexity’s chat interface.
  • Global Reach: Expanding Perplexity’s commerce tools to PayPal’s 430+ million active accounts across approximately 200 markets.
  • Secure Transactions: Leveraging PayPal’s robust fraud detection and data security protocols.

Whether users are researching a topic, comparing products, or planning a trip, Perplexity turns natural questions into trustworthy, ready-to-use results, streamlining how people learn, decide, and get things done online.

https://www.perplexity.ai/shopping ■ https://newsroom.paypal-corp.com/2025-05-14-Perplexity-Selects-PayPal-to-Power-Agentic-Commerce

Databricks to acquire Neon

May 14, 2025 /

Databricks, a Data and AI company, announced its intent to acquire Neon, a serverless Postgres company. Databricks plans to continue innovating and investing in Neon’s database and developer experience for existing and new Neon customers and partners.

Recent internal telemetry showed that over 80 percent of the databases provisioned on Neon were created automatically by AI agents rather than by humans. These workloads differ from human-driven patterns in three ways:

  1. Speed + flexibility: Neon can spin up a fully isolated Postgres instance in 500 milliseconds or less and supports instant branching and forking of database schema but also data, so experiments never disturb production.
  2. Cost proportionality: Neon’s full separation of compute and storage keeps the total cost of ownership for thousands of ephemeral databases proportional to the queries they actually run.
  3. Open source ecosystem: Neon is Postgres-compatible and works out of the box with popular extensions.

Databricks and Neon will work to remove the traditional limitations of databases that require compute and storage to scale in tandem — an inefficiency that hinders AI workloads. The integration of Neon’s serverless Postgres architecture with the Databricks Data Intelligence Platform will help developers and enterprise teams efficiently build and deploy AI agent systems.

https://www.databricks.com/company/newsroom/press-releases/databricks-agrees-acquire-neon-help-developers-deliver-ai-systems ■ https://neon.tech

Gilbane Advisor 5-14-25 — LLMs prefer analogies, MCP vs CodeAct

May 14, 2025 /

This week we feature articles from Valentin Hofmann, Leonie Weissweiler, David R. Mortensen, & Janet B. Pierrehumbert, and Kelvin Lu.

Additional reading comes from Erica Vendetti, Rachel Dixon, David Leininger & Matt Argomaniz, Gretel Kahn, Scott Brinker, and Thomas Claburn.

News comes from Algolia, OpenSearch, Optimizely, and Bloomreach.

Our next issue arrives May 21.

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

Opinion / Analysis

Like humans, ChatGPT favors examples and ‘memories,’ not rules, to generate language

Really interesting new study…
“by researchers at the University of Oxford, the Allen Institute for AI (Ai2), and others found that large language models (LLMs) generalize language patterns in a surprisingly human-like way: through analogy, rather than strict grammatical rules.”

Summary (3 min): https://techxplore.com/news/2025-05-humans-chatgpt-favors-examples-memories.html

Full paper (37 min):  https://www.pnas.org/doi/10.1073/pnas.2423232122

MCP or not, Manus made a choice

Kelvin Lu discovered that Manus provides more detail on how it works than other AI vendors, and was not using MCP (Model Context Protocol). Instead, Manus is focused on CodeAct. Lu’s comparison of MCP and CodeAct is enlightening. The article is written for developers. (7 min)

https://pub.towardsai.net/mcp-or-not-manus-made-a-choice-40b0a66d2d7c

More Reading

All Gilbane Advisor issues

Content technology news

Algolia unveils AI agents for Salesforce and Adobe

The real-time, context-aware AI agents illustrate how autonomous systems can act on real-time data across enterprise platforms.
https://www.algolia.com/about/news/algolia-unveils-new-real-time-context-aware-ai-agents-across-salesforce-and-adobe

Optimizely updates Opal

Opal unlocks use cases for marketing and digital teams, from task automation, to brand content creation, and experimentation analysis.
https://www.optimizely.com/ai

OpenSearch releases OpenSearch 3.0

To improve efficiency, performance, and accelerate AI application development via data management, AI agent, vector search capabilities.
https://opensearch.org/blog/unveiling-opensearch-3-0

Bloomreach adds marketing and ecommerce features

New features highlight agentic AI personalization and customer engagement across marketing and product discovery.
https://www.bloomreach.com/en/news/2025/bloomreach-unveils-the-features-ushering-in-the-agentic-era-of-marketing-and-ecommerce/

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 most Wednesdays. We do not sell or share personal data.

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

Foxit launches enterprise Smart Redact Server

May 14, 2025 /

Foxit, a provider of PDF and eSignature products and services, announced Smart Redact Server, a new AI-driven platform built to automate the redaction of sensitive data at scale across enterprise environments. Designed for organizations that manage large volumes of regulated content, the solution enables legal, financial, healthcare, and compliance teams to detect and redact classified, personally identifiable (PII), and other sensitive information.

Smart Redact Server equips teams tasked with data privacy and regulatory compliance, including enterprise administrators, legal and compliance departments, financial institutions, and healthcare organizations with a scalable way to identify and redact sensitive content, whether it’s personally identifiable information (PII), financial records, or protected health data. Features: 

  • AI-powered detection and redaction of PII, financial, legal, and medical data 
  • Batch processing and scheduled scans automate compliance workflows, reduce manual effort 
  • Support for 47 file types, including PDF, DOCX, scanned images, HTML, JSON, XML 
  • Cloud-native SaaS platform with integration across OneDrive, SharePoint, AWS S3, and Box 
  • Zero data retention policy with AES-256 and SSL 2048-bit encryption for in-place, secure document processing 
  • Regulatory compliance with GDPR, HIPAA, CCPA, SOX, ISO 27001, and SOC 2 Type II 
  • Role-based team management, audit trails, and reusable redaction templates tailored by industry and region.

https://redact.foxit.com/

Algolia unveils AI Agents for Salesforce and Adobe

May 8, 2025 /

Algolia, a AI-native search and discovery platform, announced a new AI agent experience that illustrates how autonomous systems can act on real-time data across enterprise platforms. Demonstrated using Salesforce’s Agentforce and integrated with Adobe Experience Manager (AEM), Adobe Experience Platform (AEP), and Salesforce Commerce Cloud (SFCC), Algolia shows what’s possible when agents are equipped with live, structured context, without latency or hallucination.

Algolia semantically interprets user intent, retrieves structured content from its index containing data from multiple customer datastores, and assembles context-aware responses in real time—bridging the gap between front-end agent platforms like Agentforce and the backend systems that hold critical content and customer signals.

Algolia handles this orchestration through its AI-native search engine, which retrieves the most relevant information from its index, content that has been ingested and structured from platforms like Adobe Experience Manager (AEM), Salesforce Commerce Cloud (SFCC), and Adobe Experience Platform (AEP). Whether surfacing personalized media, live product availability, or behavioral attributes, Algolia assembles responses that reflect the full customer context and returns them in milliseconds. The result is an agent experience that feels intuitive, precise, and deeply responsive to user needs.

https://www.algolia.com/about/news/algolia-unveils-new-real-time-context-aware-ai-agents-across-salesforce-and-adobe

Optimizely updates OpaI

May 7, 2025 /

Optimizely, a digital experience platform (DXP) provider, introduced the next evolution of Optimizely Opal. With an all-new user interface and a direct embed across the Optimizely One product suite, Opal unlocks use cases for marketing and digital teams—from task automation, to brand content creation, and experimentation analysis.

Opal builds context windows by combining diverse data sources from an organization’s corporate knowledge, including existing assets, campaign performance, and experimentation results. Opal is enhanced by Optimizely’s domain expertise incorporating content and experimentation practices. Opal agents leverage dozens of tools built for marketing that integrate with external systems. Users can fine-tune with custom instructions and orchestrate these agents within workflows. Enhancements include:

  • A redesigned platform experience puts marketers in control with features like Instructions, a dedicated user interface to manage new and pre-built agents, and Chat which is persistent across all Optimizely products.
  • A pre-built tool library helps marketers with analyzing webpages, creating presentation decks, and conducting keyword research.
  • Specialized agents for marketing tasks, each equipped with instructions and tools to deliver results or complete actions on marketers’ behalf.
  • Agent workflows allow admins to combine multiple Specialized Agents to work together—either step-by-step or at the same time.

https://www.optimizely.com/ai

« Older posts

Policies

Contact

twitter   rss   Mastadon   Bluesky   Medium

Contact • customerservice@gilbane.com

© 2025 Bluebill Advisors (except for contributed posts which are licensed under creative commons)

Banner photos: Fort Point Channel, Summer St bridge, and Northern Ave bridge, Boston

Bluebill Advisors Inc logo

© 2025 The Gilbane Advisor

Theme by Anders Noren — Up ↑