EF News
Earlier this week, Forum members came together for a practical, peer-led conversation about how they are actually using AI in their businesses. Rather than chasing hype or theoretical use cases, the focus was on real experience: what’s working, what isn’t, and where AI is genuinely improving productivity.
This kind of peer-to-peer sharing is powerful. It allows entrepreneurs to cut through the noise and, in this case, explore and adopt tools that solve real problems, rather than losing time (and money) experimenting with an ever-expanding universe of apps and agents.
The session was expertly hosted by Imran Anwar, founder of Alt Labs, and Hannah Brennan, co-founder of Seer and researcher at the Turing Institute, who brought both commercial and academic perspectives to the discussion.
There is no “Top 10” AI stack
One thing became clear very quickly: there is no single list of must-have AI tools.
There are hundreds of strong applications available to entrepreneurial businesses, but the right one depends entirely on the problem you are trying to solve. And that insight matters. Too many teams start with the tool, not the challenge.
Without clarity on the problem, it’s easy to fall into “AI playtime”, testing clever tools that may be impressive, but don’t materially improve productivity or decision-making, and ultimately cost a business money.
Core AI assistants: similar foundations, different strengths
When discussing general AI assistants, members shared candid comparisons between ChatGPT, Claude, Copilot, and Gemini.
Within the group, ChatGPT and Claude were generally preferred over Copilot, with Claude often cited as particularly strong for research accuracy and longer-form reasoning. Gemini stood out for businesses already embedded in Google Workspace, where its ability to access Google Docs, email and calendars can be genuinely useful.
However, this raised an important governance point. Tools like ChatGPT and Claude do not access internal files or emails unless documents are explicitly uploaded. Gemini’s deeper integration can be a strength but also a risk if not properly governed. Understanding the security and data-sharing implications of any AI tool is essential, and good company governance remains critical to mitigating those risks.
Prompting is now a core business skill
Another strong theme was the growing importance of good prompting. Members shared how dramatically the quality of outputs improves when teams invest time in learning how to ask better questions.
A practical recommendation emerged: build a shared company prompt library. This allows prompts to be stored, refined and reused across teams embedding learning and avoiding everyone starting from scratch. Some members are even using AI itself to critique and improve their prompts, accelerating capability across the business.
Making meetings more effective with AI
AI-powered meeting tools were another area of active use. Many members are now using transcription apps to support clearer follow-up and accountability. However, there was near-universal agreement that no one ever goes back to read full transcripts.
The real value lies in tools that summarise discussions, surface key decisions, and clearly assign actions. Granola was the most commonly mentioned transcription tool, valued for being free, simple and relatively lightweight. More structured approaches are emerging through tools like Notion, which can move from simple to-do lists to collaborative action tracking, and Monday.com for larger teams needing more robust workflow management.
A clear warning accompanied this discussion: transcription alone isn’t enough. Without a deliberate system for storing, tracking, and closing actions, AI simply creates more information, not better execution.
AI for developers
For those already coding, members shared positive experiences using Cursor to integrate AI directly into their development workflow, speeding up iteration while keeping human oversight firmly in place.
The real takeaway
Every entrepreneur in the room recognised the potential for significant productivity gains from AI when it is adopted well. The tools are already here.
The hard work lies in clearly understanding the problem you’re trying to solve, putting the right governance in place, and embedding AI into day-to-day workflows rather than treating it as a novelty.
Get that right, and AI stops being overwhelming and starts becoming a genuine competitive advantage.
