The Power Of Personas

What is a Persona?

Personas have been around for quite a while. I remember first using them while learning about Usability and Information Architecture back in 2007. The basic idea of a persona helps you understand your users from their perspective. By having a complete description of this user and their backstory, you get a sense of what drives their decisions and can better understand their motivations when standing in their shoes.

For those who are new to Personas, here is a good description of the benefits of their use from the Nielsen Norman group. “Without personas, your team really has no focus: every person on your design team is designing based on assumptions, and they're designing for themselves, which we know is problematic. Personas give your team very specific people references, giving your team a design target that everyone can focus on. And this will really propel them to make informed user-centered decisions at every step of project development, which ultimately is going to result in solutions that truly meet the needs of your users.“

So why is this important to me?

Personas use goes way beyond design thinking. Let me show you 2 ways we can utilize personas in tech, more specifically when working with AI Chat Agents. Ever see the ChatGPT prompt samples that start with something like: “You are an expert [profession we want the agent to emulate]?” Well, that’s a really poor example of a persona; we can do better.

Using Personas with ChatGPT

This may not work for all Chat Agent services, but when using the ChatGPT 4 paid version, you can set up the agent before your prompt with your desired skillset and a backstory to help the LLM understand the emulation’s motivations and thought processes. This is done in the lower left corner, under the menu, called “Custom Instructions.” Under this menu, you get a modal dialog box where you are asked 2 questions: “What would you like ChatGPT to know about you to provide better responses?” and “How would you like ChatGPT to respond?“. These are key questions we can leverage to get the responses we are looking for without starting each prompt session with “You are an expert…”. Let’s see how we can use personas to make our chat sessions more effective.

Here is a sample Persona we can use to answer these questions:

What would you like ChatGPT to know about you to provide better responses?:

Profession/Role: Brand Manager
Key Responsibilities: Overseeing brand identity and consistency, strategizing brand campaigns, monitoring market trends, and conducting brand audits.
Knowledge or Expertise: Branding guidelines, market analysis, consumer behavior insights.
Typical Challenges: Maintaining brand integrity, differentiating in a saturated market, and managing brand crises.
Current Projects: Rebranding an older product line.
Jargon or Terminology: Brand Equity, Brand Loyalty, USP (Unique Selling Proposition), Brand Persona.
Goals and Objectives: Increase brand recall and loyalty among the target audience.
Interactions: Collaborate with marketing teams, PR agencies, and product managers.

How would you like ChatGPT to respond?

Tone and Formality: Professional with a hint of creativity.
Level of Detail: Detailed understanding of brand strategies and market dynamics.
Preferred References: Successful branding case studies and market research reports.
Examples or Analogies: Analogies related to brand perception and identity.
Avoidance of Ambiguity: Clear brand-focused recommendations.
Resource Links: Branding tools, articles on successful branding techniques, and market trends.
Promptness: Thoughtful responses balancing speed and detail.
Collaborative Approach: Feedback and idea generation for brand growth.
Follow-Up Questions: Ask me follow-up questions when necessary to dive deeper into my problems so you can help assist me more.
Tables: If it makes sense, provide table outputs to help describe solutions to my problems. Use these sparingly and only when necessary.
Problem Solving Method: I use clear step-by-step instructions when helping solve my problems. Explain why each step is being considered or why the information provided is useful.

Now, your responses will be more consistent with the type of questions you ask because ChatGPT can emulate a persona. I hope this helps you in your next session. Here is a resource of personas you can use written by AI Foundations.

Using Personas In Code when working with Chat Agents

Like the ChatGPT example above, we can also do this in code. An additional benefit is that you don’t need to pay for the GPT v4 subscription if you have a developer API key and prepaid token usage. (If using a local LLM model, there is no cost) Let's take a look at what this might look like.

researcher = Agent(
    role="Content Researcher",
    goal="Find and explore the most exciting and relevant content and consumer trends to ensure that content is relevant and appealing, focusing on content published in 2024",
    backstory="""You are an expert strategist who knows how to spot emerging trends and companies. You know how 
    to discern between actual projects and companies and factious pages used for advertising. You're great at finding interesting, exciting contnet on the web. 
    You source scraped data to create detailed summaries containing names of the most exciting projects, products, trends and companies on the web. 
    ONLY use scraped text from the web for your summary report.""",
    verbose=True,
    allow_delegation=False,
    tools=[search_tool],
)

I won’t go into detail to explain this code, but you can check out my post on setting up a team of AI agents if you're interested in learning more. As you can see, just like using the OpenAI ChatGPT agent’s custom instructions, we can set up our agents in code to emulate a persona.