A few weeks ago, we hosted our biggest Founders Collective yet, bringing 60 companies together at the incredible warehouse HQ of Boden, one of our previous partner brands. This was our fifth event focused on AI, but that is not because we think brands aren’t using it. They are often in smart and creative ways, but use is still too often siloed in teams like customer service (using chatbots) or marketing (testing image generators and writing tools).
The real opportunity? Every team in your business using AI to make everyday tasks quicker and better. However, that shift needs leadership support, as well as the tools, training, and time to help people experiment and build confidence.
Many of you, possibly not all, would have opened this email and immediately thought, ‘there’s a typo in that image. That is not how you spell generative.’ But ChatGPT didn’t realise, even when responding to a prompt that included the correct spelling of ‘generative’. Errors like this can turn people off AI and can be a huge block to experimentation. This is why, alongside encouraging usage, it is essential to make sure your teams are well equipped and AI-savvy.
For the session, we invited Wouter van Haaften from GenAI Strategy to run a masterclass and live demo. As he talked through usage tips, he had AI in the background doing a deep dive on Piper, brainstorming some marketing ideas, and even designing a fully functional loyalty app for Boden. His point? AI is already powerful. But you only see the value if your team knows how to use it well.
Here are the key tips on helping your whole business use AI:
1. Don’t wait for the perfect tool. Start now.
You don’t need to spend months researching and testing the ultimate AI tool for a specific challenge. Just get your team access to a secure AI workspace. Paid options like ChatGPT Enterprise, Copilot or Gemini are affordable and protect your company data. They give everyone a consistent starting point. You can trial one and then swap to another. The tech is changing daily, so you need to be flexible. At Piper, we started on ChatGPT. We’ll probably move to Copilot, and we’re experimenting with bespoke GPTs and developing prompt libraries. The key is to start building what you need, not wait for someone else to.
2. The opportunity is people, not tech.
You’re not competing with AI; you’re competing with teams who know how to use it. That’s why some companies are offering prizes, even extra salary, for the best internal AI use cases. We’re already seeing brilliant use cases across the portfolio, including sentiment analysis, creation of product display landing pages, and personalised CRM communications.
It takes time to build great prompts and use AI in a meaningful way (more on that below). So, encourage people and create space for them to experiment. Give everyone in your business an “AI Day” or appoint an “AI lead” to help others build and test tools.
3. Teach people to use it properly.
Bad outputs stop people from trying again. (Like when ChatGPT misspelt “generative” in the image for this email, even though the prompt was correct.) We’ve hosted multiple AI sessions with our portfolio to help teams learn how to use it well.
Always remember that AI isn’t a robot. It’s a human mimic and not a mind reader. AI only works if you guide it well. Prompts are often long and detailed. They need to be crafted to get the best out of the AI.
Here are some of the best practices we follow:
- Pick the right model. Different AIs do different things well. ChatGPT has 6 models, including ones that are great for writing or reasoning. Claude reads PDFs brilliantly. Perplexity is designed for research.
- Give it a role. “You are a commercial analyst for a wellbeing brand…” helps it deliver more relevant answers. Give it a context on the business it’s working for, and how you would think about the task.
- Tell it how to reply. Be clear on output format, otherwise it will take its own approach: “Give me three ideas in a table” or “Respond using these headings.”
- Set rules. Create the guidelines you want it to follow and parameters to help it answer well. If you don’t want guesses, tell it: “Do not infer or hallucinate.”
- Reuse prompts. Good prompts are gold. Make them long and detailed. Save them, refine them, and share across your team. We use Magical to do this quickly, but a simple copy and paste works well.
- Be curious. Test and learn. If you’re not sure if AI can do a job, ask it. If you see a new button or development, test it (try Sora and Deep Research on ChatGPT). Be curious and flexible.
You don’t have to be an AI expert today. But teams that experiment now will move faster, be more efficient, and find better solutions sooner. After our first AI Geekmeet last year, we shared a list of tools to get started – read them here. But what is becoming clear is that while the tools change and evolve. If you follow the principles above, any tool can make your day-to-day easier.
We’re certainly no experts, but we’re testing, learning, and using it as much as possible at Piper. Plus, we’re fortunate to know some real specialists. If you’ve got questions or want help to start or improve how you’re using AI, let us know. We’ll support if we can.