AI is a broad topic that has exploded over the last few years. Today there are many types of AI, from general chat bots, that scrub the internet for information and return the best guess solution to the user, to domain or product specific chat or voice products that only respond with answers specific to that product.

The difference in AI technologies, can be demonstrated in the different results a user would receive when asking ChatGPT about the best health Insurance for a young family with 1 child, and asking the same question to a chat bot from Anthem Blue Cross. The ChatGTP result will include all of the info that it could find from any provider across the internet, while the Anthem Blue Cross bot will just provide answers from the Anthem menu.

I used the word “menu” because the database that are searched by these chat bots can be compared to a menu. Product or service specific AI usually draws from a smaller database of questions and answers, therefore having a smaller menu, than a more ubiquitous AI like ChatGPT.

Think about a hamburger. If you tell someone at McDonald’s or another specific restaurant that you want a hamburger, you can expect a specific result. A hamburger that looks like the burger from that restaurants menu, but if you google hamburger, you will see thousands or Pattie based sandwiches with a plethora of different toppings. Why? Because google is not constrained to the same menu, as 1 restaurant, so when googling if you expect someone more similar to a McDonald’s hamburger, you have to be more specific in your search. To get what you expect you have to specify your prompt.

For the sake of this article we will be using the OpenAI ChatGPT

Image of Google results of “McDonald’s Hamburgers”

Prompts

Continuing with the menu and restaurant analogy, prompts, are like your order. Remember, you are wording food from the endless possibilities at the Restaurant at the End of the Universe, so in order to get exactly what you want you have to make sure to order exactly what you want.

If you don’t care what you get, then give the AI a vague prompt. Let’s look at some examples of Prompts, AI results (based on OpenAI ChatGTP), and how these results can be used to help productivity.

Example 1 : Gathering Data and Kickstarting a project.

Sometimes the hardest part of any project is just starting it. For example, as a UX Researcher, there have been multiple times when a client, or company, has asked me to create a survey to help with qualitative research. Usually most of these surveys are the same, 3 demographic questions, and up to 10 product or service specific questions, including things like sign up, ui intuitiveness, and the always present NPS question, however even though most of the surveys are the same sometimes coming up with the questions can be tricky. This is where generative AI is helpful. As a professional, I know that we need ~ 10 questions, reported on a 7 point Likert Scale, and sometimes I even have specific questions that must be asked, about a specific website or feature.

Why not let AI do a draft for you and then utilize your time to edit and perfect the survey?

Prompt: Create a 10 question quantitative survey around the sign up and login process of ESPN + , part of the espn.com universe. Using a 7 point Likert Scale, generate 10 questions about the user experience. Have 3 questions ask about the sign-up process, 2 about the value of the product , 4 about the login and return visits, and 1 NPS question.

Number the questions and put them each on their own line, make the output something that I can copy and paste into a word document.

Result: Sure, here's your survey:

On a scale of 1 to 7, how easy was it to find the sign-up option on ESPN+?
Please rate the clarity of the instructions provided during the ESPN+ sign-up process.
How satisfied are you with the overall sign-up experience for ESPN+?
How likely are you to recommend ESPN+ to a friend or colleague? (NPS question)
Rate your perception of the value you receive from your ESPN+ subscription.
How often do you encounter issues when logging in to ESPN+?
Please rate the convenience of the login process for ESPN+.
On a scale of 1 to 7, how likely are you to return to ESPN+ after visiting the platform once?
How satisfied are you with the options provided for recovering forgotten login credentials on ESPN+?
Rate your overall experience with ESPN+ in terms of user-friendliness and functionality.

How I use them: Immediately you notice that the order does not follow the order of my prompt. You also notice that the 7 point Likert Scale is not written out, but with my domain knowledge it is easy to add. The questions also need to be converted to statements for use on the Likert Scale. The results are not perfect, but they do give you 10 questions and a great starting point to edit, reword, and reorder the survey questions.

Follow-up Prompt : Sometimes some of the prompts are neglected by the GPT or need a follow-up for clarification.

What are the main demographic questions asked when conducting a quantitative survey?

Result: When conducting a quantitative survey, researchers often include demographic questions to understand the characteristics of the respondents. Some common demographic questions include:

Age: "What is your age?"
Gender: "What is your gender?"
Education: "What is the highest level of education you have completed?"
Income: "What is your approximate annual household income?"
Ethnicity/Race: "Which racial or ethnic group do you identify with?"
Employment status: "What is your current employment status?"
Geographic location: "In which country/state/city do you reside?"
Marital status: "What is your marital status?"
Household composition: "How many people live in your household?"
Occupation: "What is your current occupation or job title?"

These demographic questions help researchers analyze survey data across different groups and understand how certain factors may influence responses.

How I use them: Now I am able to add the demographic questions that my team deems important and finalize the survey. Due to the Sample nature of this survey, I only used the 3 most important questions as chosen by my team, Age, Gender, and Education

The Final Survey can be found here.

This is a great way to kickstart a project. It also saves time when creating files that will can be quickly edited, modified, and reused.

Example 2 : Checking your work and standards

I wish I had this tool at my disposal when I was moving Pointchain from waterfall to agile. A transformation like that, aligning and rescheduling 2 development teams, is a rather large undertaking, combined that with the resource constraints of a start-up, it is inevitable that something will fall through the cracks. At the time it would have been a game changer if I was able to quickly check my work and schedule against the industry standards. Thanks to the current iteration of AI, I can do just that.

First, I create the transitions process and timeline. Now looking at the timeline and work that our teams have in their queue, I start to formalize our transition process and build out the initial sprint, making sure that all the agile ceremonies are accounted for. I am also in charge of breaking larger epics down into stories and provide guidance for how to properly write a user story and an agile ticket.

So how can I use AI to help? In this instance I will use AI to check my work. I will begin by making sure that I have accounted for all of the agile ceremonies in my initial sprint.

Prompt: I am setting up an agile development team based off the standard agile framework, what ceremonies are required in every development sprint? Provide the answers in table format with columns listing the specific agile ceremonies, their goal, the deliverable of the ceremony, what employees and roles are in the ceremony, the meeting length, and meeting frequency. Use the Rows to put the ceremonies or meetings in the correct order with the first ceremony at the top and the last ceremony in the sprint last . Allow me to copy and paste the table .

Results:

A raw version of the result table, listing Agile Ceremonies, Participants, Lengths, and Frequency

How I use them: Once again the data must be cleaned, but once it is cleaned it is ready to use. After comparing the output from the AI with the ceremonies that I have already accounted for and scheduled, I can fill in the blanks and make sure we are implementing best practices

See the usable format below

A cleaned version of the previous table, listing Agile Ceremonies, Participants, Lengths, and Frequency

Now that I have all of the required ceremonies accounted for, the correct participants invited, and the meetings scheduled, I can use AI to make sure that our first tickets have all of the required information the is standard according to Scrum Alliance.

Prompt: According to Scrum Alliance and Jira best practices, what are the requirements of a well defined agile development story ticket? List these requirements in a bullet format with the title of the section or requirement followed by a “:” and the definition of the requirement and an example of a well written requirement. Next please define What is a Spike Ticket? What is an Epic Ticket? Also please present the standard pointing system used by most agile teams, how is a 1 point, 3 point, 5 point, and 8 point ticket different?

Results: Requirements of a Well-Defined Agile Development Story Ticket:

Title: Clearly states what the task or user story is about.

Example: "Implement User Authentication Feature."

Description: Provides detailed information about the task or user story.

Example: "Develop a login page with email and password fields. Validate user credentials against the database. Implement error handling for invalid login attempts."

Acceptance Criteria: Defines the conditions that must be met for the task to be considered complete.

Example:
User should be able to successfully log in with valid credentials.
User should see an error message when providing invalid credentials.

Priority: Indicates the relative importance or urgency of the task.

Example: High, Medium, Low.

Estimate/Story Points: Assigns a point value to represent the effort or complexity of the task.

Example: 3 points.

Spike Ticket:

A Spike Ticket is used to conduct research, exploration, or prototyping to gather information or reduce the uncertainty of a task.
Example: "Investigate the feasibility of integrating a third-party API for weather data."

Epic Ticket:

An Epic Ticket represents a large body of work that can be broken down into smaller, more manageable tasks.
Example: "User Management Module."

Standard Pointing System:

Most agile teams use the Fibonacci sequence (1, 2, 3, 5, 8, 13, etc.) for pointing system.
The point values represent the relative effort or complexity of the task.
1 Point: Very small, straightforward task.
3 Points: Small task with some complexity but manageable within a sprint.
5 Points: Medium-sized task with moderate complexity.
8 Points: Larger task with significant complexity, may require multiple sprints or team members.

How I use them: Since this prompt was specific and these definitions are standard across the Scrum Alliance and Agile community I can copy and paste this info, after cleaning it, directly into our agile process documentation.

I first create a sample ticket utilizing all of the Ticket Sections listed above, with the examples for new team members, to standardize our ticket requirements. Personally, I also like to add a well written user story to the ticket, though the AI results mention a user story they do not give an example, so I can either write a standard example or ask the AI for a sample, this was the sample provided by AI:

Prompt: Please give me a standard user story that could be written and given to an agile developer

Result: [ This is more of a finished ticket but it does include a user story

example, highlighted in the beginning]

Title: User Registration Feature

Description: [**** User Story***] As a new user, I want to be able to register for

an account on the platform so that I can access the full features of the

application.

Acceptance Criteria:

When I visit the registration page, I should see fields for entering my full name, email address, and password.
I should receive an email verification link after submitting the registration form.
Clicking on the verification link should confirm my email address and redirect me to the login page.
If I try to register with an email address that is already in use, I should see an error message prompting me to use a different email address.
Upon successful registration, I should receive a welcome email with instructions on getting started.

Priority: High

Estimate/Story Points: 5 Points

Now I have all of the required meetings on the schedule, and information on tickets, ready for our development teams transition to agile. We are also able to use the standards and examples given by the AI to provide team members with templates and examples in case more questions should arise in the future.

So as you can see, generative AI can be leveraged by product managers and others to kickoff projects and gather up front data, to help check assumption and work, and to set down standards and provide process examples.

How might you use AI within your company or role moving forward?

Other ways to add AI to help productivity:

Using AI to transcribe meeting and other voice recordings. (The first result of AI is a Draft transcript that can easily and quickly be audited and edited by a live employee)
Using Generative AI, to draft designs and layouts. (Again the results of visual AI generators should not be used as a final product, but instead used to generate ideas, and drafts that can be edited and modified.
Use a professional account on ChatGPT to create guardrails and your own GPT service with user specified rules and processes.

Bonus:

Add some humor to a meeting with an AI generated Joke.

Vague Prompt: Tell me a dad joke based in the Star Wars universe.

Result: Why did the Jedi bring a ladder to the bar?

Because they heard the drinks were on the house!

Slightly more Specific Prompt: Tell me a dad joke about Darth Vader and the Star Wars universe.

Result: Why did Darth Vader go to school?

Because he wanted to learn how to use the "force" of gravity!

Find some Trivia to engage your team

Prompt: Give me a music trivia fact about today’s date, Feb. 6th

Result: On February 6th, 1958, rock and roll legend Buddy Holly, along with Ritchie Valens and J.P. "The Big Bopper" Richardson, tragically died in a plane crash. This event became known as "The Day the Music Died," immortalized in Don McLean's song "American Pie."

AI - Prompts, Results, and Getting what you ordered