Skip to content
popart__comic_style__propel_spaceman_cracking_an_egg_and_making_a_cake_-ugly__deformed__noisy__blurry_3676993815
Ben Ross | Propel Ventures Co-FounderMay 30, 2024 3:34:11 PM2 min read

When it comes to AI Product Design, Make Sure We Allow People to Add the Egg

Incorporating AI into software products presents a unique challenge for product designers: finding the balance between automation and user agency. This balance is crucial for creating products that are not only efficient but also engaging and satisfying for users. To illustrate this, let's look at a story from the 1950s involving General Mills's Betty Crocker instant cake mix.

Back in the 50s, General Mills introduced a ready-to-use cake mix called Betty Crocker. All you had to do was pour in water, mix it up, and bake it. An hour later, you'd have a delicious cake. Despite the convenience and quality of the product, it didn't sell well. The reason? It was too convenient. People felt guilty serving something that required so little effort. They wanted to feel involved in the process.

General Mills responded by removing the powdered egg from the mix and instructing users to add a fresh egg themselves. This simple change made users feel like they were truly baking, not just following a simple step. Sales skyrocketed as a result.

This story highlights an essential lesson for AI product design: while automation can greatly enhance efficiency, users still need to feel like they are contributing to the process. Here are some key considerations for product designers when incorporating AI into their software products:

1. Identify Critical Decision Points

Determine which parts of the workflow benefit most from human input. For example, in a content generation tool, AI can suggest options, but the user should make the final selection. This keeps the user engaged and ensures the output aligns with their preferences.

2. Maintain Transparency

Users should understand how AI makes decisions and have the ability to override or adjust those decisions. Transparency builds trust and allows users to feel more in control.

3. Provide Customisation Options

Allow users to customise AI behaviour to suit their needs. This could mean adjusting the level of automation or personalising the AI's responses based on their style and preferences.

4. Encourage User Interaction

Design the product to encourage regular user interaction. This could involve periodic prompts for user input or opportunities for users to refine AI-generated suggestions.

5. Balance Efficiency and Engagement

Ensure that while AI handles repetitive or complex tasks, users still engage with meaningful parts of the process. This maintains a sense of ownership and satisfaction in the final outcome.

In conclusion, the story of Betty Crocker's cake mix teaches us that even in highly automated systems, there's value in allowing users to "add the egg." By thoughtfully integrating user input into AI workflows, product designers can create more engaging and satisfying products that users feel connected to.

Ben Ross | Propel Ventures Co-FounderMay 30, 2024, 8:18:48 AM2 min read

Balancing AI Investment: How to Stay Ahead Without Overcommitting

In the fast-paced world of software product development, the decision on how much to invest in AI can be challenging. For ...
Start Reading
Amy Johnson | Chief Product Officer PropelOct 26, 2023, 12:37:10 AM2 min read

Why is no-one doing Product Discovery?

Product discovery is an invaluable phase in the product development process. It's all about uncovering what users truly want ...
Start Reading
Ben Ross | Propel Ventures Co-FounderFeb 27, 2025, 8:31:41 AM3 min read

Why AI is Exposing Weaknesses in Agile – and How a Strong Product Operating Model Can Fix It

For over eight years, Propel Ventures has been at the forefront of product thinking in Australia—helping businesses not only ...
Start Reading