CRE(AI)TOR:
A Content Creation App for
the Individual

Role: Product Designer
In the age of technology, content creation has become more than just a “good” skill–it’s both a necessary and a lucrative one. Whether you are an individual or a multibillion dollar company, the content you put out into the world is a reflection of your identity (or your “brand”), and a crucial component in the maintenance and growth of your product.

For this reason, it is more important than ever for content creators to find a voice that:
showcases who they are
follows trends
stands out
appeals to their audience
With AI in the mix, any creator can go from a low-impact, small-reaching “business” to a heavy hitter. With this in mind, I set out to determine what content creators enjoy, struggle with, and would like to see change in their relationship with content creation.

The final product? An AI-based progressive web app that helps content creators specify, maintain, and grow their audience without giving up the unique content that makes them who they are.

The Who, What, Where, When, and Why

Date: February 2022
Team: Individual Class Project
Tools: Figma, Figjam, Google Suite, Otter.ai, Voice Memos, Pencil & Paper
Sprint Length: 3 Weeks
Final Deliverable: 4 High Fidelity Screens

Skip the details and see the final screens here

Project Scope

Business Requirements: Build a user-friendly progressive web app (PWA) that leverages AI to generate better quality content for creators across mediums. Users must be able to define their audience scope as well as key words that best represent their brand.

User Challenge: Users wanted to be able to visualize and customize the size and specificity of their target audience, as well as post to various sites and mediums from one place.

Emotional Goal: Maintain the excitement about the original idea throughout the process of creating and releasing the content into the world

Research Questions

  • What issues or frustrations do content creators face throughout the content creation process?
  • How do users feel about AI as it pertains to their creative process?
  • What gaps remain unfilled in the existing market for AI content creation?
  • How can the current content creation process be improved for users?
  • How much control are users willing and comfortable giving up to an AI feature?

1. Understanding the Current Problem Space

Methods: Domain Research, Competitive Analysis

Background: The Vast World of Content Creation

What is "content creation?" According to a PEW study, content creation is "the material people contribute to the online world." But what if it's more nuanced than that? This definition lacks the sparkle and glitz of true creativity and personalization–of collaboration and empathy. Perhaps the better question is, what is content creation for?
What does content do? What is it for? A visual representation of a Weber Design Study (2017)

Content creation, at its most basic level, is "for" people

Every time a person posts on the internet, they are contributing to the great mountain of online content to be consumed by–you guessed it–other people. Youtube Shorts alone get over 15 billion views per day–nearly twice as many views as people on the planet. That's a lot of views, but quite a lot of content to compete with.

What if, instead of seeing a singular mountain of content, we could shift our mindset to see a massive mountain range? A peak for animal videos, a peak for comedy sketches, a peak for cooking blogs, and on and on until all the genres of human existence encircle us, and all creators can find their niche?
This was the mentality I set out with to help creators determine their creation space, and find and target their specific audience.

A Look at the Competition

With its seemingly endless possibilities, there are tons of companies using AI technology. Every day we move one step closer to the chair-laden, smoothie-drinking existence of a Wall-E inspired existence. In the meantime, why not learn more?

Six companies consistently popped up in my search for "AI Content Creation" that sit at the crosshairs of creation and marketing (because what are influencers, really, if not the marketing elite for products, or–more often–themselves?).
Hubspot
Indirect Competitor
‍
Offerings:
site optimization for search, content topic suggestions, data analytics, schedule social media posts
‍
Vibe:
clean, informative, accessible
Concured
‍Direct Competitor
‍
Offerings:
customized suggestions for optimal UX, personalization based on learned interests
‍
Vibe:
technical, less creative, easy to use
BrightEdge
‍Direct Competitor
‍
Offerings:
AI-based search discovery, content creation alignment, strategy to increase traffic & performance
‍
Vibe:
data-driven, clear, engaging, visually busy
Acrolinx
‍Direct Competitor
‍
Offerings:
flexible, customizable, clear + compliant branding, data analytics, advanced security, capabilities for large scale
‍
Vibe:
bright, engaging, colorful, accessible, whimsical
Automated Insights: Wordsmith
‍Indirect Competitor
‍
Offerings:
language generation tool, real-time analysis, customizable, publishing capabilities on all sites
‍
Vibe:
technical, informative, cold
Cobomba
‍Direct Competitor
‍
Offerings:
Market trend research, audience analytics, content brief suggestions, discoverability + performance improvement
‍
Vibe:
whimsical, informative, engaging, fun

Color and Font Choices Among the Competition

Key Takeaways:
  • Acrolinx & Cobomba both make a point of creating whimsy and ease for their users through color, typography, illustration style
  • All AI companies in this sphere use a variation of the same color palette
  • None of the competitors (even direct ones) are directly marketed at individuals
Opportunity: there is little to no competition in AI technology that is engaging, user-friendly, and whimsical, that focuses on content creation assistance for individuals

2. Identifying the Problem

Methods: Interviews, Transcript Coding, Empathy Mapping, How Might We,
Problem Statement Breakdown

Creators in the Flesh (through Zoom, but close enough)

I interviewed six content creators, with content varying from queer portraiture to comedic skits to thought-provoking podcasts. Each creator adamantly stood by their love for coming up with ideas and actualizing them–i.e. the creative process.

On the other hand, they all mentioned editing, research, and search engine optimization (SEO) as being frustrating and time-consuming. Thoughts on AI varied, each interviewee noting the "obvious" pros, but explaining a skittishness around giving up control.

“I enjoy the creative artistic process... coming up with an idea and executing it. And I enjoy if I get to do it with a friend. I love collaborating and that makes it a fun activity.
“I think people would like my page, but unfortunately I don’t think it really reaches a lot of people. I don’t think I can do much about that except for streamline more, which is what I don’t really want to do.”
“People fall in love with your content and with your brand. The content I use is a way to connect with my target audience. I try to embody my brand by promoting confidence with the way I present myself, living fearlessly without regrets, representing myself boldly”

“[Community management] reminds people who engage with you that you care, and that a real person cares about their engagement.”
“I don’t trust algorithm things. Like... It’s a robot.
They’re really useful but they’re not creative.
I don’t think any AI can really replace human creativity”

Transcript Coding: From Keywords to Classification

In order to organize and analyze the detailed accounts I received from creators, I coded the transcripts into manageable bites, seeking out common themes and talking points. Then I took the themes and synthesized them into takeaways, which gave me direction towards next steps.

Transcript Coding Method

From Code to Theme

Synthesis of User Interviews

Empathy Mapping: Getting Into the Mind of the User

After synthesizing quotes from my interviewees, I wanted to be sure I understood the way creators feel relative to the work they do. I created an empathy map, with careful consideration of moments in which my interviewees did not directly name feelings or refer to an act as a "pain point," but spoke around it. Conversation tone played a big role here–a great reminder as to why interviews are an important research method.

Empathy Map

How Might We: The Problem is Coming into Focus

I mulled over pain points from the interviews–most of which directly echoed those found during domain research. Could slogging through research be made easier? Maintaining an audience fun? Branding without streamlining possible? The examples below proved most useful in determining the problem statement.
How Might We
make all aspects of content creation enjoyable?
How Might We
make the research + technical side of content creation like character development? Like the sims? Like a choose-your-own-adventure novel?
How Might We
make content creation feel like a full journey with a beginning, middle, and end?
How Might We
use alternative research methods to figure out what creators + audiences want?
How Might We
lessen the load of the content creation process on the non-creative end?

Problem Statement Creation

Draft 1: Content creators, project managers, and marketing teams need assistance on the overall process of researching, creating content, and maintaining a following in order to spend more time on the enjoyable aspects of creating without losing credibility or quality in the final products, and continue to grow their audience base.
+ too large in scope: too many users & problems to solve
+ lacking specificity: difficulty in determining a solution direction
Draft 2: Content Creators need assistance in the process of researching, editing, and maintaining a following in order to spend more time creating their content and growing their audience base.
+ "need" still too large
+ insight is vague: how does this help users?
Final Problem Statement:
Content Creators need to understand the mismatches and overlaps between their ideal audience and their current audience in order to better solidify their brand and create content to match.

3. Gearing Towards a Solution

Methods: Persona Creation, Sketching, Low-Fi Wireframes, Competitive UI Analysis,
High Fidelity Wireframes, Usability Testing

A Pause to Re-Group

At this point in the process, what do we know?
  • Gaps in the market (a.k.a. path of least resistance into a competitive sphere):
    AI assistance with research, discoverability, and content tuning for individuals
  • Creator tasks (making, releasing, and marketing content) are time-consuming
  • Creators would struggle less and enjoy the process more if they had a better understanding of their audience, and how best to reach them.
  • Good, empathetic design will assist creators in their process without taking over or squashing the unique, personalized quality of their content

Persona: Meet Sophia

Sophia is a social media manager who wants to grow her audience, but has many interests and doesn’t want to put herself in a “brand box.” Understandable, right? I'd love a few more followers without having to exclusively post singing videos or hiking photos.
‍Sophia represents the collective identity of my interviewees and research.

Concept Brainstorming

I began by asking myself the "simple" questions; looking at the problem through Sophia's eyes to better understand how my theoretical problem statement could become a concrete solution. This was also useful because it gave me certain design constraints:
Q: How you determine your current audience?
‍
A: By connecting your social media accounts: your current audience(s) can be analyzed and understood through data analysis software, or–in the case of programs like instagram–it has already been done by THEIR data analysis software.
Q: How do you teach an AI program who your target audience is?
‍
A: By plugging in specific details, such as age, interest, location, etc. etc.
Q: How do you keep AI from rebranding and streamlining your content
(a concern raised by many interviewees)
‍
A: By giving users the information and the option to use AI suggestions, but not requiring it: transparency + user agency.
Q: As a content creator, how do you strengthen your brand, gain more followers, and reach more people?
‍
A: By optimizing your hashtags, keywords, and phrases; researching your competition; releasing more contet; and actively engaging with your current audience
Though I felt like I was writing an "Ask Alice" column for the back of a tween magazine, I also managed to go from, “'The world' is my oyster” (which is great and all, but incredibly overwhelming), to something more along the lines of, “'Canada' is my oyster,” which–though still grand, exciting, and full of possibility–felt much more manageable.

Concept Sketching

I sketched out my Onboarding Flow Concept:
On top of basic information, the user chooses keywords to describe their content, what they want assistance with, and who their target audience is. Finally, they can connect other social accounts and get an in-depth analysis about how their current audience and keywords stack up against their target audience.

Another way to think of it:

User Input:
  • Basic info (name, DoB, email, etc.)
  • Self-specified keyword descriptors
  • Target audience parameters
  • Access to social media profile info
  • Ideal form of AI assistance
AI Output:
  • Current audience insights
  • Keywords actually associated with user content
  • Overlap between current and target audience
  • Assistance in chosen form of content optimization

Concept Research: Keywords, Topics, Audience

I felt good about the concept, but in order to validate it, I needed to talk to users. After pulling areas of creator frustration from my initial research, I surveyed friends, family, and random followers for keywords that best describe the content they post.

Keyword Bubble Chart

Bubble Chart: 127 responses to the Instagram poll question "what keywords would you use to describe the content you post?"

Top Keyword Categories Among My Following:

20%
of participants post artistic media (visual art, fashion, film, books, etc.)
24%
of participants post people + relationships (family, friends, joy)
17%
of participants post adventure + nature (travel, landscapes, etc.)
These numbers validated my online domain research, with "love," "art," and "happy" consistently among the top 10 IG hashtags, and "nature," "travel," and "style" in the top 20.

Low-Fidelity Mock-Ups to High Fidelity Knock-Outs

With a concept model to work from and research to validate it, I began designing screens.
Img: Profile Screen over three stages: sketch, low-fidelity, hi-fidelity

Round 1: Low-Fidelity Screens

Based on the initial concept sketch, I designed low-fidelity wireframes for the sign-up flow as well as the dashboard.
Let's make Sophia our hypothetical user. As she sets up her account, (post-initial email/password/etc.) she is asked to define her content as she best understands it.
‍
After this, she defines her target audience in detail, chooses what kind of assistance she'd like from the AI, and then receives a detailed explanation of the mismatches between her current and target audience. The final screen is a low-fi dashboard with suggestions to optimize her content..

Round 2: High-Fidelity Screens

Competitive analysis showed that the common AI color palette includes white, blue, navy, grey, and orange variants, so I worked within the same palette to build initial trust with users (humans are, after all, creatures of habit).
Interviews showed that many young, individual content creators don’t trust the idea of AI because it feels like the taking over of their creative integrity and resulting excitement. Therefore, my assumption is they are less likely to use a product that feels heavily technical or “robotic,” and are more likely to trust a softer, more engaging and accessible site. This translates to fonts with more curves (such as Nunito or Avenir Next), more focus on color (less grey & white), and illustrations.

Cobomba and Acrolinx both leverage bright colors to create whimsy and engaging screens, so I saturated the colors and tested screens in navy, off-white, and orange.

Dark Vs. Light

I surveyed a small cohort of people about the light vs. dark color theme.
64%
preferred the dark theme, calling the added contrast and pop of orange "fun" and "engaging."

The "Target Audience" Screen

This screen proved to be a complicated one:
  • Users felt the navy background was jarring, but the white was boring
  • Users preferred more screens with less text to one text-heavy screen. Accessibility-wise, it's also more difficult for users to scroll through and click smaller buttons.
  • The rectangular "next" button was less visually pleasing and harder to see than the round arrow.
  • An instructional screen at the start of the "target audience" section was preferable to tiny text on the screen where the user makes his or her choices.

The Dashboard

Dashboards are difficult because users know what to expect from them. We interact with dashboards every day, across several apps. They must be intuitive and easy-to-use while simultaneously showcasing a large amount of information.

First vs. Second Iteration

  • First Iteration objectively uglier
  • Users found stripe pattern visually overwhelming and hard to read
  • I used a familiar pattern for iPhone users in Iteration Two
  • Information hierarchy is established better in second iteration

Round 3: High-Fidelity Screens, But Better This Time

Bearing all edits in mind, I created four high-fidelity screens for CRE(AI)TOR:
the sign-up, profile info, target audience, and dashboard (+ bonus: intro page!)

Final Thoughts, Next Steps, Areas to Improve:

This project was an excellent opportunity for me to try my hand at all aspects of the product design cycle. I created an app that identifies mismatches between content creators' target and current audiences, and then suggests language to strengthen and grow their audience in either direction.

Improvements:
  • UI Usability Testing: I was able to test among classmates and instructors, but would need to test among content creators out in the world to get a real sense of how users interact with it
  • Diversity of Interview Pool: all interviewees, though lovely, were women aged 18-45. To get a broader, more comprehensive scope of the content creator space, I would need to interview a wider range of participants.
  • Communication with Developers: in my ideal world, I would work on this product using an agile methodology, informing and learning along the way which aspects of the product are viable, affordable, possible, etc.
Thank you for reading. For any queries or interest in working with me, shoot me a message!

Click here to return to the top of the page