News You Can Use

The article reports that as legal and corporate teams pour money into AI tools, they are now being asked to prove those tools are worth the cost. Legal departments are trying to quantify time savings, output increases, and reduced headcount to make the case to finance teams - but many benefits, such as better strategy or risk avoidance, are harder to measure. Some departments track hours saved or faster task completion, while others focus on "return on experience" like reduced busywork and higher engagement.

This highlights a growing challenge: AI being bought must be justified. For AG, it means we should be proactive about measuring and articulating impact (e.g., time saved, consistency improved, lawyer satisfaction). Anecdotes are useful, but building data-backed narratives will make internal approvals and client conversations more persuasive. For example, showing how tools like AGPT have helped with daily legal exercises through data reporting on time, prompt categories, etc. It also shows we are not alone in figuring out new pricing dynamics and the value of legal work in this market.

The author argues that OpenAI is under intense competitive pressure and has declared an internal "code red" - an emergency directive from CEO Sam Altman to prioritise improving ChatGPT in the face of rising rivals like Google's Gemini, Anthropic's Claude and other emerging AI players. The shift comes after earlier success with ChatGPT (once seen as the defining AI breakthrough) and reflects a broader sense that OpenAI's product momentum has slowed, its strategic focus is scattered, and competitors have gained ground both in capability and market attention.

This narrative highlights how competitive dynamics are reshaping the AI landscape, and why organisations investing in AI must stay alert to potential shifts among major providers. When we benchmark models for internal or external solutions, vendor direction and strategy matter as much as raw capability. This means we need clear criteria for selecting tech, focusing on reliability, support, integration and scale. In client conversations, it reinforces the idea that the competition in innovation, whilst beneficial for advancement, also introduces risk, changeability or discontinuities in tooling - we can help clients plan for both innovation and resilience in their AI strategies.

Based on 37.5 million anonymised interactions, the report reveals that Copilot is heavily used in both work and everyday life. On desktop, the AI is mainly used for work and career tasks during office hours. On mobile, users repeatedly turn to it for health, wellness and personal guidance, with spikes around events like Valentine's Day. The data shows users rely on Copilot for advice, not just search or productivity.

AI assistants like Copilot are no longer just tools - they are becoming part of how people think and work. Recognising that Copilot supports both work and life tasks suggests we should emphasise workflows and interfaces that feel natural in daily use: e.g., fast context switching, mobile-friendly queries, and quick summaries. This also means that we need to ensure high standards of accuracy, relevance, and safety in how we present outputs, and align our design and training with real use patterns.