AI Training Resources

Further Resources 

In June 2026, Orchestra's June AI Innovation Challenge: Client GPT Edition,  featured an hour journey of learning, innovation, and inspiration. Our 7 finalists shared practical use cases of their GPTs to compete for the top spot. More information and resoruces can be found at the Orchestra AI Innovation Challenge page. 

AI Fundamentals 201 

Pete Pachal and Melissa Flynn, The Media Copilot

September 25, 2025

In part two of the Orchestra AI Fundamentals training series, Pete and Melissa cover the following: 

1. Assistant recap

2. Fast breakdowns of topics with NotebookLM

3. Media list building with Deep Research

4. DIY media monitoring with ChatGPT Tasks

Key resources:

-> Download the deck: covers NotebookLM, deep research tools, media monitoring automations, and agents
-> Watch the recording (Passcode: ?+9=UY+1): full session with demos and Q&A

We recommend logging into your ChatGPT Enterprise account to follow along. Pete and Melissa use some other tools as well to highlight what's possible. You can follow along live with these, too. Just be sure you don't share proprietary data with unauthorized services.

• Google NotebookLM

• GPT for Work

For NotebookLM, it's helpful to have ready a particularly dense report or study (public-facing documents only—nothing confidential). If you’re unsure what to pick, here’s an example report you can use.

Takeaways

• NotebookLM is your fast primer. Upload files and let AI create summaries or even podcasts to help you get up to speed quickly.

• Deep research is about accuracy, not speed. Tools like Gemini, Perplexity, and ChatGPT Deep Research produce highly specific outputs but take time. They work best when you refine the prompt together with the model.

• Media monitoring can run itself. ChatGPT Tasks let you set up daily, automated reports on your chosen topics — a big step toward consistent, proactive comms.

• Agents are the future. Browser-based assistants like ChatGPT’s Agent and Perplexity’s Comet can surf the web and take actions for you, confirming crucial actions (like buying stuff).

Tips

• Start small: choose one of these tools and test it with a real work problem.

• Treat AI like a collaborator: steer, refine, and double-check the outputs.

• Think in workflows: AI can support research, monitoring, and execution — not just text drafting.

AI Fundamentals 101 

Pete Pachal and Melissa Flynn, The Media Copilot

September 18, 2025

In part one of the Orchestra AI Fundamentals training series, Pete and Melissa set the stage for how AI can strengthen PR and comms work:

• AI mindset & risks: What AI is, where it excels, and where human judgment must stay primary

• Prompting essentials: Output vs. collaboration, plus a prompting framework for better results

• Assistants 101: Building a “5-minute assistant” workflow

• Applied demos: Executive Voice Simulator and Client Feedback Simulator in action 

-> Download the deck: Includes the POP prompting framework, assistant-building steps, and use case demos

-> Watch the recording (Passcode: @=&exUh0): Full 90 minutes, with Q&A

Key takeaways

• AI is software that makes choices. It’s not magic — it’s pattern-matching. Knowing that helps us decide what tasks to give it (and which to keep human).

• Weaknesses don’t remove your agency. Hallucinations, bias, and sycophancy are real, but you can actively counter them through how you prompt, check, and ground the model.

• Prompting is strategy. Output prompts (press release, tweet, blog) get you drafts; collaborative prompts help you think through strategy, crises, or creative ideas in real time.

• Assistants = context + repeatability. Custom GPTs are great for consistent voice and functionality; Projects are better when you need to ground outputs in shared files like client dossiers.

• PR use cases are here, today. From simulating an executive’s voice to stress-testing a pitch against client feedback, AI can directly support high-stakes comms work if built thoughtfully.

What we built together

• Executive Voice Simulator (Custom GPT): Drafts or refines copy in a leader’s authentic style, using speeches and past posts as training material.

• Client Feedback Simulator (Project): Evaluates campaign pitches against brand guidelines, compliance docs, and past client feedback — anticipating objections before they happen.