IESE Insight
Mastering GenAI: resources to help you get the most out of your AI buddy
Prompting the AI to get the output you want is a vital skill. Here are some tips for doing it effectively.
By now, most people have discovered and applied GenAI to automate routine tasks — generating emails and reports, doing basic analyses — or to brainstorm with tools like ChatGPT. In theory, we’re freeing up our time so we can focus on higher value work.
Yet, to reap such benefits, we have to understand how to make the most of these new technologies. Rather than using whatever tool seems handy, we should first know the answers to some basic questions:
- Are your objectives clear and aligned with organizational objectives?
- Are your chosen tools likewise aligned and do you know how best to use them?
- Most important, do you know their limits?
If you can’t answer these basic questions, how will you ever know when the AI is giving you wrong answers?
Dedicate time to developing some basic skills and mastering the nuances of the GenAI you intend to use. Below are some resources to help.
Crafting the perfect prompt for my AI buddy
The Personalization Revolution is upon us. The business trend is for individualized experiences, tailoring product recommendations, content curation and marketing messages to each customer, with real-time adjustments, all thanks to advances in GenAI. To help us in this effort, we can now count on the ultimate personalization assistant — what I affectionately refer to as my AI buddy. But getting the most out of GenAI for appropriate personalization hinges on a new skill: prompting.
Do you know how to prompt effectively? Or do you simply ask and hope for the best? Unlike the straightforward nature of keyword-based internet searches, prompting demands a structured approach for desired outcomes. Mastering the art of prompting — asking for help from your AI buddy — is a vital skill.
Before diving in, you have to realize that not all tasks are suitable for every GenAI platform. Each excels at something different. If your company provides you with ChatGPT, Microsoft Copilot or Google’s Gemini Advanced, for instance, be aware of the existence of competing alternatives and decide which GenAI is best for your needs.
Once chosen, give your AI buddy some context: who you are, your aims, the task at hand. The more it knows about you and your frames of reference, the better your AI buddy will be able to adjust and tailor its responses closer to your needs. When prompting, ensure the following:
- Context & background: Provide details about the setting, the audience and the purpose behind your request.
- Task definition: Clearly state what you need your AI buddy to accomplish, avoiding ambiguous statements.
- Clear instructions: Specify the format, structure, style and tone desired for the output. Are you looking for persuasive, informative, playful or technical language? Do you need a tagline, a social media post, a blog article or a product description? The more specific you are, the better the outputs.
- Boundaries: Tell your AI buddy what not to do. What limitations must be adhered to, in line with organizational standards?
In this process, appreciate the value of iteration and refinement. Prompting is like having a conversation — a dynamic process where you see what does and doesn’t work. Were you too vague? Was the tone off? Then, make adjustments accordingly, refining your prompts — and learning — until expectations are met. This feedback loop is crucial for continuous improvement.
Like any skill, practice makes perfect. By mastering the basics, you’ll pave the way for GenAI to streamline tasks and elevate your impact in the Personalization Revolution.
One Useful Thing/More Useful Things
Ethan Mollick, co-director of the Generative AI Lab at Wharton, offers research-based views on the implications of AI in his blog/newsletter, One Useful Thing. A companion site, More Useful Things, provides practical resources and a library of prompts that users are free to distribute, remix, adapt and build upon, even for commercial purposes, so long as they give proper credit. Read more about the ins, outs and ethics of GenAI in his new book, Co-intelligence: living and working with AI (Penguin Random House, 2024).
Barcelona Deep Tech Summit
IESE is a core partner with Aldea Ventures, Barcelona Supercomputing Center (BSC), Mobile World Capital and Tech Barcelona to discuss building and scaling deep-tech startups, access to investors and corporate venturing, as part of the Smart City Expo World Congress in Barcelona, Nov. 5-7, 2024.
IESE regularly leads visits to the Barcelona Supercomputing Center (BSC) to highlight how supercomputing capabilities can help executives resolve real-world business challenges, with practical tips for integrating these technologies into their strategic decision-making.
BSC Director Mateo Valero will be speaking on “Computing & Innovation” at the Global Alumni Reunion in Madrid on Nov. 14, 2024.
DeepLearning.AI
Machine learning pioneer Andrew Ng founded DeepLearning.AI to be a platform for world-class AI education. The site offers a variety of short courses, ranging from non-technical ones (how to apply AI tools to problems in your organization) to specialized ones (how to use tools to translate languages, summarize text and even build your own chatbot) with specific courses on “Prompt compression and query optimization” and “ChatGPT prompt engineering for developers,” among others. For educators, there are downloadable course slides and guides. Subscribe to The Batch newsletter to receive weekly AI news, insights and commentaries.
Women4Ethical AI
ChatGPT recommends men, not women, update their wills to bequeath their assets. It gives tips for women to help them with meal planning and grocery shopping. The prompt “nurse” shows pictures of women, while the prompt “lawyer” shows men.
All the stereotypes that exist in the world have been trained into the AI models, so it’s important for users to be alert to them when their prompts return biased results.
To help maintain ethical standards in the use of AI, UNESCO launched the Women4Ethical AI platform to ensure AI tools don’t end up reproducing or magnifying gender inequalities. This network of women in tech seeks to promote non-discriminatory algorithms and data sources, to avoid propagating unhelpful content.
This article is published as Smart Picks in IESE Business School Insight magazine #168 (Sept.-Dec. 2024).