241 | 🎛️ the “Precision Prompt”—a simple tool to fine-tune your interactions

In today’s Brainyacts:

1. Clippy 2.0

2. The Precision Prompt Technique

3. UAE launches “virtual lawyer” and other AI model news

4. European Parliament is using Claude AI and more news you can use

   👋 to all subscribers!

To read previous editions, click here.

Lead Memo

📎 🤦 Microsoft’s “Copilot Agents” Are Clippy 2.0 for Enterprises

Microsoft’s announcement of new agentic capabilities for its Copilot platform might sound revolutionary—enabling businesses to create autonomous agents that orchestrate tasks across sales, finance, and supply chains. But before we get caught up in the glossy marketing and success stories featuring tech giants like McKinsey and Thomson Reuters, it’s important to take a critical look at what’s really happening here.

Microsoft Copilot, which has been rolled out across its suite of tools, aims to democratize generative AI by integrating it into everyday workflows. On the surface, that sounds like a noble goal. But for those of us who have been watching this space closely, it’s increasingly clear that Microsoft’s strategy is actually watering down the potential of generative AI, especially for professional users, like those in the legal field.

The Benioff Critique: Not Entirely Wrong

Marc Benioff, CEO of Salesforce and no fan of Microsoft’s Copilot, recently labeled it “Clippy 2.0,” accusing Microsoft of panicking and rebranding a failing product. While it’s easy to write off Benioff’s critique as competitive shade, his underlying argument rings true for many in the AI and enterprise tech space. Copilot, and now its so-called “agents,” do not deliver the horsepower that many expected from a generative AI tool, especially in professional settings where accuracy, security, and depth of knowledge are non-negotiable.

Benioff’s pointed critique that Copilot lacks the “data, metadata, and enterprise security models” to truly transform corporate intelligence underscores an important issue. For all its bells and whistles, Copilot is not the game-changer Microsoft claims. Instead, it risks leading business users to prematurely judge generative AI as limited or inadequate. This is more than a missed opportunity; its a massive disservice to the power and utility of LLM-based AI models.

The McKinsey and Thomson Reuters Cases: A Red Herring?

Microsoft’s blog post touts the success of agents built by McKinsey, Thomson Reuters, and Clifford Chance. But these examples don’t reflect the reality for most organizations, especially law firms that may be considering dipping their toes into AI.

McKinsey, for example, is creating an agent to speed up client onboarding by 90% and reduce administrative work by 30%. This will sounds truly attractive for law firms and their in-house clients who often complain about the onboarding process. But McKinsey is not just any company—it’s a global consulting behemoth with deep pockets and unparalleled technical expertise. It already sells AI transformation services, so it’s no surprise that it can implement cutting-edge tools and extract efficiencies from them. The same goes for Thomson Reuters, which built a legal due diligence agent that speeds up workflows by 50%. Again, this is a massive, resource-rich organization with the capacity to integrate and optimize new technology. Most law firms and in-house teams, large or small, don’t have the luxury of large-scale digital transformation teams or multimillion-dollar tech budgets.

And then there’s Clifford Chance. Microsoft references them in its announcement but curiously provides no details about what they’re actually doing with these agents. It’s telling that the legal use case isn’t highlighted with specific outcomes—perhaps because many law firms lack the internal technical firepower to adopt and scale such solutions without outside help. And therein lies the rub: integrating Microsoft Copilot agents into a law firm’s operations will likely require significant consulting fees and technical know-how, making it prohibitive for most legal teams.

Microsoft’s strategy of integrating Copilot and its agents everywhere risks diluting the broader conversation about AI. By pushing a simplified, accessible version of generative AI, they may be setting the wrong expectations for what these tools can actually do in professional settings.

You should be cautious in buying into the hype. Copilot agents might serve basic operational needs, but when it comes to the high-level work that defines the legal field, more powerful, specialized AI tools are likely needed.

Spotlight

🎛️ Introducing the “Precision Prompt”—A Simple Tool to Fine-Tune Your Generative AI Interactions

It’s been a while since we’ve gone deep on prompting, but I thought it’d be the perfect time to share a technique that I think will level up your interactions—whether you’re a daily user or still dipping your toes in. Let me introduce you to something I’m calling the “Precision Prompt”.

We’ve all learned that it’s not just about what you ask but how you ask. And yet, sometimes you get a response that’s just a bit off—maybe too broad, too generic, or missing that specific detail you were hoping for. Instead of starting over or feeling like the AI just didn’t “get it,” the Precision Prompt gives you a way to refine that interaction on the fly.

What is the Precision Prompt?

It’s like handing you the dials to adjust the AI response. After it gives you an answer, simply use the Precision Prompt to help identify the key concepts, ideas, or words that shaped its reply. It’ll show you exactly what can be tweaked—whether you want more depth, a different tone, or even a shift in focus—and give you specific suggestions on how to adjust.

Why Should You Use It?

• Instant Insight: Instead of guessing what went wrong, this gives you real-time feedback on how to refine its responses.

• Better Targeting: You can zero in on the exact part of the conversation that matters most.

• Learn as You Go: It’s not just about getting a better response; it teaches you how to be a better prompter, which is essential as AI becomes more integrated into our daily workflows.

How Does it Work

After you get your initial reply, copy and paste this follow-up prompt (you might want to save it for future use):

Based on your previous response, please identify the key areas or concepts in your reply that I can adjust or refine to change the outcome. Highlight any specific words, phrases, or ideas that, if clarified or altered, would significantly impact the direction or depth of your next response. Please also suggest potential “dials” or “knobs” I can turn, such as:

• Increasing or decreasing the level of detail or specificity.

• Shifting the tone (e.g., more formal, casual, technical, creative, etc.).

• Focusing on a different aspect of the topic (e.g., a narrower or broader scope).

• Elaborating on particular areas that were mentioned but not fully explored.

• Providing examples, analogies, or further explanations to clarify the response.

By using this prompt, you’ll be able to guide most generative AI tools to deliver answers that are more aligned with your exact needs—without needing to re-prompt endlessly or start from scratch.

Take It for a Spin

Next time you’re interacting with an AI assistant and the response isn’t quite hitting the mark, try this. You’ll be surprised at how much more insight, depth, and clarity you can get from a simple adjustment.

Happy prompting! 🚀

AI Model Notables

► 🚨 Hours ago Anthropic unveiled enhanced AI models that look to surpass GPT-4o - an upgraded Claude 3.5 Sonnet, and a new model, Claude 3.5 Haiku.

► 🚨 Anthropic also showing an incredible new feature called: “compute use” where Clause actually drives around your own computer to do certain tasks. This will be available to developers today.

► UAE to launch Virtual Lawyer, the first AI-guided legal professional in the region.

► EU and beyond get the Advanced Voice mode:

► xAI, Elon Musk’s AI startup, launches an API. An API feature is important news because it means that other companies, developers, or apps can now easily connect to and use Grok, the AI model provided by xAI. Essentially, the API (Application Programming Interface) is like a bridge that allows external software to communicate with and use Grok, which uses the Twitter/X feed to inform its responses.

Before the API, only X (formerly Twitter) could use Grok for its own features. But now, with the API available, anyone can integrate Grok into their own applications, websites, or services. This opens up many new possibilities for using the AI in different industries and products, making the technology more widely accessible. I cannot wait to see what people build, using the X/Twitter feed! 😆

► Anthropic has started evaluating AI models for sabotage risks—like whether a model can hide capabilities or deceive human supervisors.

► Also from Anthropic:

► A recent study revealed that Google's AI-overviews are misleading or incorrect in 43% of finance-related queries, especially around tax and financial aid topics.

► Google released updates to its viral AI note-taking assistant NotebookLM, adding new features that let users guide AI-generated audio summaries and announcing the upcoming launch of a new business-focused version.   

• Users can now customize the AI podcast Audio Overviews feature by providing instructions to focus on specific topics or adjusting the expertise level.

• A new Background Listening feature allows users to listen to Audio Interviews while multitasking within NotebookLM.

• A pilot program for NotebookLM Business is coming, offering enhanced features for organizations like higher usage limits and team collaboration tools.

News You Can Use:

➭ SEC to increase scrutiny of AI tools used by financial brokers, advisers.

➭ Dow Jones and New York Post sue AI startup Perplexity, alleging ‘massive’ copyright infringement.

➭ Alcon Entertainment filed a lawsuit against Elon Musk’s Tesla and Warner Bros. Discovery, alleging the unauthorized use of AI-generated images from the film Blade Runner 2049 in Tesla’s Robotaxi promotion.

➭ Publisher Penguin Random House revised its global copyright notices to include a statement explicitly preventing the use of texts for AI training purposes, which will now appear on all titles.

➭ European Parliament expands access to their archives with Claude.

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.