Prompt Engineering for Ideation: Customer Satisfaction

By Tojin T. Eapen

In this post, we'll examine the potential of using large language models (LLMs) like ChatGPT in the idea generation process, with the case of a coffee shop that is looking for ideas to enhance customer satisfaction. Specifically, we'll delve into the use of design heuristics in the process of crafting prompts. In this post, we explore the use of the vary principle a.k.a. the attribute dynamics heuristics. Our prompts are in blue and are italicized, and the response of ChatGPT is in green.

D-Café, located in Columbia, Missouri, is looking to improve customer service. We have identified output attributes, such as the number of options, ambiance, friendliness of staff, taste of coffee, cost of products, and selection of food items. Can you provide additional output attributes we should consider?

Additional output attributes that could be considered for D-Café could include:

• Speed of service: How quickly customers are able to receive their orders and be served.
• Cleanliness of the café: The overall cleanliness and maintenance of the café, including the bathrooms.
• Accessibility: Whether the café is easily accessible for customers with disabilities.
• Wi-Fi availability: Whether customers are able to access Wi-Fi while they are in the café.
• Sustainability: Whether the café is taking steps to reduce its environmental impact, such as using reusable cups, recycling, or sourcing ingredients from local farmers.
• Personalization: Whether the café offers the ability for customers to customize their orders and drinks.
• Special events: Whether the café holds special events, such as live music or poetry readings, that customers may be interested in attending.
• Additional services: whether the café provide additional services such as printing, scanning, or mailing services, or whether they offer a loyalty program.

Keep in mind that these attributes would vary based on the type of café, their target market and location. It's important to understand the current state of the café and target market.

Great! We have identified some input characteristics that are relevant to D-Café, such as weather conditions, time of day, loyalty of customers, and age of customers. Can you provide additional input characteristics we should consider?

Additional input characteristics that could be relevant to D-Café include:

• Location: The physical location of the café, including whether it is in a high-traffic area or a residential neighborhood.
• Competition: The number and type of other coffee shops and restaurants in the area.
• Advertising and promotions: The effectiveness of the café's advertising and promotional efforts, including social media and local media.
• Staffing levels: The number of staff members working at the café at any given time, and whether they are able to handle the volume of customers.
• Seasonal fluctuations: Whether the café experiences fluctuations in customer traffic due to changes in weather or other seasonal factors.
• Product sourcing: The quality and cost of the ingredients and products used by the café.
• Reputation: The reputation of the café among customers, including reviews on social media and other review sites.
• Employee training and culture: The level of training and support provided to staff members and the overall culture of the café.

It is important to keep in mind that these input characteristics can interact with each other and impact the café's performance and customer service. A thorough understanding of the current state of the café, as well as the target market and location, can help identify the key input characteristics that are most relevant for D-Café.

Great, now please generate five ideas where an output attribute is modified or changes based on a change in an input attribute. For example, one idea could be that the output attribute (price of coffee) changes based on the input attribute (weather).

• The output attribute, selection of food items, is modified based on the input attribute, time of the day. For example, the café could offer a different menu for breakfast, lunch, and dinner.
• The output attribute, ambience, is modified based on the input attribute, weather conditions. For example, the café could have outdoor seating during sunny weather and indoor seating during rainy weather.
• The output attribute, speed of service, is modified based on the input attribute, staffing levels. For example, during peak hours, the café could have additional staff members to ensure that customers are served quickly.
• The output attribute, personalization, is modified based on the input attribute, loyalty of the customers. For example, regular customers could have the option to save their favorite drink orders, or have a loyalty card that gives them discounts.
• The output attribute, special events, is modified based on the input attribute, age of the customers. For example, the café could host events such as trivia night, poetry reading or live music that are more likely to attract a specific age group.

It's important to note that these ideas are just examples, and it would be important to test and evaluate the feasibility and impact of these changes before implementing them.

The ideas provided by ChatGPT provide a useful starting point to further improve and refine the strategies for D-Café. They can be used as a foundation to generate more specific and actionable plans. Additionally, they can serve as a benchmark for evaluating the effectiveness and feasibility of new ideas.

To learn how leading Fortune Global 500 companies such as ABB, Bosch, Google, Samsung, and NetApp have used Innomantra's Functional Innovation Methodology to turbocharge their idea management process, schedule a meeting today at calendly.com/innomantra.