From AI Curious to AI Fluent: Building Your Own AI Partnership

Getting started with AI assistance can feel overwhelming. As I speak on AI, often the biggest takeaways are the concrete steps to begin an AI collaboration journey. Let's begin with actionable steps, starting with choosing your AI partner.

Finding Your AI Partner

The best way to understand AI assistants is to experience them directly. I now recommend going to Poe.com because it offers an excellent starting point - you can think of it as a "try before you buy" platform for various AI models. By bringing together Claude, GPT-4, Llama, DALL-E and other LLMs under one roof, Poe lets you:

• Seamlessly compare responses across different models - with the click of a button

• Better understand each assistant's unique strengths

• Experiment on multiple LLMs without multiple subscriptions

What I've found particularly valuable is how different models handle the same prompt. While Claude excels at nuanced writing and analysis, GPT-4 might offer more technical and coding precision, and Llama or Mistral could provide a fresh perspective. This comparison shopping helped me find the right fit for my workflow. As I mentioned in my last post, I use Claude most of all, but still find incredible value in testing and using other assistants.

Getting Started: Safety First

While AI assistants are incredibly helpful, they're also prone to certain, shall we say, quirks. One crucial habit to develop early: verify everything, especially links to sources. Keep the Human In the Loop!

AI models, regardless of their sophistication, often generate plausible-looking but non-existent links and citations. (They’ve been trained to suggest probable tokens / parts of words, and that training unfortunately carries over to link suggestions.) At most, think of their suggestions as research starting points rather than direct references. I’ve found too often that even the most believable links can be hallucinations. <— Talk about a quirk!

Speaking Their Language: Effective Prompting

Think of prompting as having a conversation with a very capable but very literal colleague (we’ve all had them, now’s the time to appreciate what you’ve learned from them, or to finally learn something new about communication styles and preferences here...)

Here are key strategies that consistently work for me:

1. Set the Context (You Are Statements)
   Instead of just asking a question, frame the context:
   "You are an experienced grant writer reviewing my proposal. Please analyze this draft focusing on clarity and compelling narrative."

2. Provide Examples (Few-Shot Learning)
   Before: "Write me a professional email."
   Better: "Write a professional email following this style:
   Example 1: [Short example]
   Example 2: [Short example]"

3. Define the Tone (Support Your Personal or Professional Brand)
   "Please respond in a casual, conversational tone similar to a trusted mentor explaining concepts over coffee."

4. Specify Output Format (Get What You Want Sooner vs. Later)
   "Structure your response as bullet points, with each main idea having 2-3 supporting details."

5. Request Reasoning (Help You Both Learn Along the Way)
   "Walk me through your thinking process" or "Explain your rationale for these suggestions."

6. Shake Things Up (Surprise Me Most Often Works Delightfully)

   “This feels a bit pedantic. Surprise me with a different perspective.”

Building Consistent Workflows

The real power comes from integrating AI assistance into your daily routine. Here are three workflow examples I've found particularly effective:

The Writing Partnership

• First Draft: Brain dump your ideas

• AI Review: "Review this draft for clarity and structure, suggesting specific improvements"

• Revision: Incorporate useful suggestions while maintaining your voice

• Final Check: "Compare this revised version to the original, highlighting major improvements"

  The Research Assistant

• Topic Exploration: "What are the key debates around [topic]?"

• Independent Verification: Always fact-check AI-suggested sources and find primary references

• Counter-Arguments: "What are the main criticisms of this perspective?"

• Synthesis: "Help me organize these verified findings into a coherent outline"

The Problem-Solving Partner

• Problem Statement: Clearly define the challenge

• Framework Check: "What frameworks or approaches would be relevant here?"

• Solution Brainstorming: Generate multiple approaches

• Critical Analysis: "What are potential weaknesses in each approach?"

Pro Tips:

• Save your most effective prompts for reuse

• Build context over multiple exchanges rather than trying to get everything in one shot

• Develop a verification habit: treat AI responses as helpful suggestions requiring validation

• Regular reflection: What worked? What didn't? Adjust accordingly

• Mix AI assistance with human collaboration for optimal results

• Maintain a collection of verified sources separate from AI suggestions

Common Pitfalls to Avoid:

• Assuming the AI knows your full context without providing it

• Accepting responses without critical evaluation

• Using overly complex prompts when simple ones would work

• Forgetting to verify factual claims

• Trusting any links or citations without verification - AI models often generate plausible-looking but non-existent URLs and citations

• Taking technical advice, especially code, without testing

• Following legal, medical, or financial advice without professional consultation

The Learning Curve

Remember, building an effective AI partnership takes time and experimentation. Start small, perhaps with one workflow, and gradually expand as you become more comfortable. Pay attention to what works for your specific needs and adjust accordingly.

Next Steps:

• Choose your platform (I recommend starting with Poe)

• Test different models with the same prompt

• Save prompts that work well for you

• Start with one simple workflow

• Develop your verification process

• Bookmark reliable primary sources for your field, and provide the assistant your source preferences

• Create a system for tracking verified versus AI-suggested information - check every link

• Gradually expand your AI collaboration across workflows… PLAY!

In our next post, we'll look at how AI agents might further transform our work processes. But for now, focus on building that foundation with an AI assistant - it's the best preparation for the exciting developments ahead.

Governance, Ethics & Bias in AI

Resources from AIGG on your AI Journey

Is your organization ready to navigate the complexities of AI with confidence?

At AIGG, we understand that adopting AI isn’t just about the technology—it’s about doing so responsibly, ethically, and with a focus on protecting privacy. We’ve been through business transformations before, and we’re here to guide you every step of the way.

Whether you’re a government agency, school district, or business, our team of experts—including attorneys, anthropologists, data scientists, and business leaders—can help you craft Strategic AI Use Statements that align with your goals and values. We’ll also equip you with the knowledge and tools to build your playbooks, guidelines, and guardrails as you embrace AI.

Don’t leave your AI journey to chance.

Connect with us today for your free AI Tools Adoption Checklist, Legal and Operational Issues List, and HR Handbook policy. Or, schedule a bespoke workshop to ensure your organization makes AI work safely and advantageously for you.

Your next step is simple—reach out and start your journey towards safe, strategic AI adoption with AIGG.

Let’s invite AI in on our own terms.