Gemini 2.5 Computer Use Model

Hi everyone,

In today's edition, we're diving into Google's impressive release of the Gemini 2.5 Computer Use model, showing off benchmark-crushing performance and robust safety features that have the AI community buzzing.

Our second deep dive takes us into the expanded collaboration between Anthropic and Deloitte, scaling the deployment of Claude across Deloitte's massive employee network and highlighting the push for reliable enterprise AI solutions.

Let's get right into it.

In this issue:

• 🤝 In Partnership: Automate sales follow ups with data

• 🤿Deep Dive: Google introduces Gemini 2.5 Computer Use Model

• 🤝In Partnership: Build your revenue machine in your inbox

• 🖼️AI Art: Examples of great and trending AI art

• 🤿Deep Dive: Deloitte expands Claude alliance

• ⚒️Tool Snapshots: Tools for AI, no-code, and productivity

• 📰Top News: News on AI, no-code, and productivity

🤝IN PARTNERSHIP WITH HUBSPOT

Tired of guessing when to follow up? The improved trumpet app for HubSpot turns buyer engagement into automated workflows—no manual tracking required.

Here's how: Prospect watches your demo video → trumpet syncs it to HubSpot → your workflow auto-triggers a follow-up email.

Also, see AI engagement scores and stakeholder signals right in your CRM. Join customers like Connectd who are already seeing sales cycles shorten by 28% after using the app.

🤿 DEEP DIVE

Developers and Google AI on the Gemini 2.5 Computer Use Model

Intelligence: Google released the Gemini 2.5 Computer Use model in public preview via the Gemini API, enabling agents to operate web and some mobile interfaces with strong benchmark results and built-in safety guardrails.

• The model is accessible through Google AI Studio and Vertex AI and is built on Gemini 2.5 Pro to power agents that click, type, scroll, fill forms, use dropdowns, handle logins, and complete browser-based tasks with lower latency than leading alternatives.

• The operation runs in a loop via the computer use tool, where the model receives a user request, a screenshot, and recent actions, returns an action like click or type, may request user confirmation for sensitive steps, and continues with updated screenshots and URLs until completion or stop.

• Performance is reported as leading on multiple web and mobile control benchmarks, including the Browserbase harness for Online Mind2Web, and the model is optimized for browsers, showing promise on mobile, but is not yet optimized for desktop OS-level control.

• Safety is addressed with model-level protections against misuse, unexpected behavior, and prompt injection, plus developer controls such as a per-step safety service that reviews each proposed action and system instruction requiring refusal or user confirmation for high-stakes tasks like bypassing CAPTCHA or controlling medical devices.

• Early uses include Google teams and external testers applying it to UI testing, Project Mariner, Firebase Testing Agent, and AI Mode in Search, with reports of about 50 percent faster execution, up to 18 percent better performance on hard evaluations, and recovery of over 60 percent of failed runs that had driven about 25 percent of test failures.

• Getting started options include a Browserbase-hosted demo, building locally with Playwright or in a cloud VM with Browserbase, and community support via the Developer Forum.

🤝IN PARTNERSHIP WITH SPACEBAR STUDIOS

👉 What if your inbox became your #1 growth channel?

That’s why we partnered with Spacebar Studios — this is their bread and butter.

They’ve:

• Grown a client newsletter from 0 → 115,000 subscribers in just 14 months in the compliance space

• Scaled another to 35,000+ engaged hospitality operators

• Helped multiple clients turn email into seven-figure of pipeline/ARR

Now, they’re offering our readers something special: a completely revamped newsletter setup + strategy for free.

🖼️ AI ART

Examples of great and trending AI art

🤿 DEEP DIVE

Deloitte and Anthropic Expand Alliance to Deploy Claude at Global Scale

Intelligence: Anthropic and Deloitte expanded their alliance to roll out Claude to more than 470,000 Deloitte employees and co-develop industry-specific solutions, marking Anthropic's largest enterprise AI deployment to date.

• Claude will be available across Deloitte’s global network, with a new Claude Center of Excellence to create implementation frameworks, share leading practices, and provide ongoing technical support, enabling pilots to scale from production.

• Deloitte and Anthropic will run a formal certification program to train and certify 15,000 Deloitte professionals on Claude to support client implementations and Deloitte’s internal AI transformation.

• The partners will build solutions for regulated sectors, including financial services, healthcare and life sciences, and public services, combining Claude safety safety-first design with Deloitte's Trustworthy AI framework to increase transparency and controls.

• Leaders from both firms position Claude as a trusted AI for complex enterprise work across coding, software development, customer engagement, and industry advisory, highlighting alignment on safety, reliability, control, and technology integration.

• Deloitte describes this as a significant investment in Anthropic’s platform, with plans for a Center of Excellence, continued certifications, and industry-specific solutions to help clients through AI transformations.

⚒️ TOOL SNAPSHOTS

Futuristic tools within AI, no-code, and productivity

• 🤖 Orchestra - Streamlines team communication and workflow for efficient product creation. Free to try.

• 👥 Extrovert - AI-powered customer engagement tracking and response tool. Free to try.

• 🖼️ Freshly Squeezed - Simplify your image resizing and optimization on MacOS. Free to try.

• 🔗 Loop MCP by SimpliflowAI - Unify all apps, solves AI integration limits. Free to try.

• 💻 PromptCompose - Effortless AI prompt management and optimization. Free to try.

ℹ️ ABOUT US

The Intelligent Worker helps you to be more productive at work with AI, automation, no-code, and other technologies.

We like real, practical, and tangible use-cases and hate hand-wavy, theoretical, and abstract concepts that don’t drive real-world outcomes.

Our mission is to empower individuals, boost their productivity, and future-proof their careers.

We read all your comments - please provide your feedback!