Sharing an anonymized follow-up email I sent to attendees of one of my Enterprise AI Cooking Shows (5/5 satisfaction scores from all ~30 attendees!) - where we demonstrated in just two hours that processes (in this case, new product development) that used to take days, weeks, or months can-with prudent delegation to general and specialized AI tooling-be compressed into iterative cycles that take hours or even minutes.

Team,

Thank you all for your energy, feedback, and active participation in our workshop yesterday. It was a pleasure running an "AI cooking show" with you and exploring the frontier of what's possible when you combine your deep domain expertise with powerful AI tooling.

In just two hours, we simulated a massively compressed product research and development cycle. We went from a business objective to analyzing synthetic [redacted] data, deriving data-driven personas like "[redacted]," generating marketing collateral (landing pages, product images, video ads), and most importantly, iterating on those assets in real-time based on your expert feedback.

Key Takeaways

Beyond the specific tools, I hope these core concepts resonated:

  1. AI as a Force Multiplier: We demonstrated how this integrated process can compress a research phase that typically takes months into a matter of hours, allowing for unprecedented speed and agility.

  2. The Human-in-the-Loop is Critical: AI is not an "easy button." Your expert feedback was the most crucial ingredient. It's what guided the AI to refine the landing page, improve the [product] design, and ultimately boost the simulated conversion likelihood from 70% to 88%. This partnership is where the magic happens.

  3. A New Way of Working: The most significant productivity gains often come from simple habit changes. Adopting dictation ("talk, don't type") and learning to provide rich, contextual instructions to AI systems are the foundational skills for this new era of work.

  4. Experimentation is Everything: The most valuable applications of AI for [redacted] won't be found in a generic tutorial. They will be discovered by you, applying these tools to your unique challenges and expertise. My biggest encouragement is to go forth and experiment.

Tools & Resources

As requested, here is a comprehensive list of the tools we used in the session, along with others you might find valuable.

Tools We Used in the Session:

Foundational LLMs (Your Core Engines)

  • ChatGPT Enterprise: The secure version you have access to. We used it for general reasoning, brainstorming, and image generation with 4o.

  • Claude Code: Anthropic's model in a coding interface (CLI). We used it specifically for its powerful data analysis and coding capabilities, where it analyzed our synthetic data and helped build the final website.

  • Gemini 2.5 Pro (Google AI Studio): Google's flagship model. We used it for its large context window to analyze all our data and generate the initial set of detailed personas.

Specialized Applications (For Specific Jobs)

  • Deep Research (ChatGPT): A specialized "agent" within ChatGPT that performs in-depth, PhD-level research on any topic, complete with citations. This is how I generated the initial market reports on [redacted].

  • VAPI: The platform we used to create the AI voice agent for conducting scalable, multilingual qualitative interviews. Despite the live demo hiccups, this is an incredibly powerful tool for gathering customer insights.

  • Lovable: The AI-powered website builder we used to instantly generate and iterate on the landing page for the "[new product].

  • Operator: The AI agent we used to simulate our "[name]" persona browsing the website, providing feedback, and giving us a conversion likelihood score; this is being folded into OpenAI's new "Agent" feature via Chat, so keep an eye out for that.

  • Flow: The text-to-video (and audio!) platform (powered by Google's VEO3) we used to generate the short, three-scene advertisement based on the prompts we created.

  • Gamma: The AI presentation tool we used at the end to quickly generate a branded slide deck for an executive update.

  • WisprFlow: My recommended standalone dictation tool for Windows. This is key to unlocking the "talk, don't type" workflow and providing richer context to AI models.

Additional Tools to Explore:

  • Fireflies.ai: An AI meeting assistant that records, transcribes, and summarizes your calls. Invaluable for capturing the "raw material" from stakeholder interviews and internal meetings.

  • Listen Labs: A more off-the-shelf solution specifically for voice-of-customer surveys, as an alternative to the more developer-focused VAPI.

  • ChatGPT with Connectors: This is a feature within ChatGPT Enterprise that allows it to securely connect to your internal data sources (like SharePoint, Outlook, etc.). This is arguably the single most powerful capability for an enterprise, as it grounds the AI's knowledge in your proprietary data.

Next Steps

  1. Session Recording: [redacted]

  2. Market Research Reports: As mentioned, if you'd like the links to the full Deep Research reports on [redacted], please connect with me on LinkedIn and send me a message saying "send me the reports.

Thank you again for the opportunity to work with you. I hope this session serves as a catalyst for you to discover new ways to create value for [redacted] and your clients.

All the best,

Christian Ulstrup

For context, this 2-hour session yielded ubiquitous 5/5 satisfaction scores :)

Want to get your colleagues energized about AI adoption?

Host an GSD at Work LLC AI Cooking Show:

  • Private and personalized to your business,

  • Enterprises only,

  • Up to 80 attendees per session, and

  • Flat fee of $4,999 (satisfaction guaranteed)

Send me an email with some information about your organization to get started:

christian@gsdat.work