Projects
Introducing CaptionAI, an AI-driven caption generator, as the latest feature in Cloud Campaign – a versatile SaaS social media management platform catering to both B2B and B2C sectors.
With key features including free white labeling, simplified social media scheduling, and an integrated client approval process, Cloud Campaign lives to help marketing agencies scale. This project added an AI powered caption generator, CaptionAI, to that list of key features.
Challenge, Goal, Opportunity: Utilize AI to further Cloud Campaign on the mission to help marketing agencies scale.
For this project, the challenge was also the goal and as well as an opportunity.
With that, the solution was a new feature that would strengthen the company’s existing mission.
Give time back to our users to focus on social strategy and growing their agency.
Make social media caption generation effortless and fast.
Streamline content creation and strengthen creative momentum.
Increase product-market fit as the platform for all things social media.
Bettering the experience for our users creates an opportunity for business growth through the attraction of new customers as well as creating a stronger relationship with our current customers.
Constraints: For this project, as with many within a start up, there was a constraint with regards to time, resource, and access.
In an effort to reduce the gap caused by the above mentioned constraints, I carried out a design sprint. The goal for the sprint was to gain access to a greater thought pool in a shorter time frame.
Role: UI/UX Designer
I led the design process for this new feature as a design team of one.
For this project, I worked closely, but remotely, with the product manager, leadership, and members from all teams across the organization.
CaptionAI was released on April 12th, 2023 post a beta release on March 29th, 2023.
Market research of competitors and user interviews supported and validated business decisions to prioritize the new AI feature.
Movement of competitors aligned with the responses received from our users. Caption generation was the first step of AI incorporation within the industry. Through interviews, we were able to pinpoint how our users create copy content which served as foundational research for feature prioritization.
How users currently create copy content:
Manually outside the platform then uploaded with the need of an outside tool.
Manually within the platform, one by one.
Through outside copywriting tools.
User interviews reveal key customer insights that guided the structure and prioritization of this feature introduction.
Functionality most often asked for:
Ability to edit captions (most value)
Bulk operations ability (future phase)
Ability to generate platform specific captions (nice to have & feasible)
Ability to train the AI generator to suit their industry (nice to have & not yet feasible)
First company design sprint clarifies the 3 main locations for the caption generation feature as shown through the user flows below.
The user flows were used during conversations with the engineering team to highlight the differences in the experience, but also to validate the design differences as well.
Sprint consolidated the experience and visual direction by gathering crucial insights previously dispersed throughout the organization.
With our limited resources, I orchestrated a design sprint that included representatives from all departments, such as marketing, sales, customer success, engineering, and product.
The sprint started with a product requirements document, and the cross-functional team took part in the brainstorming and ideation stages of the process.
The following details the sprint process executed by the sprint team.
Below reflects the ideal process. I would have liked to go through this process with the sprint team. However given limited resources, the parts reflected in blue were carried out with the product manager and leadership.
Sprint Results:
Defined the 3 main entry points.
Clarified key signifiers to improve current user experience and limit confusion for customers, especially those who are self serve.
Determined key design decisions.
Additional gains of the sprint relevant to the Cloud Campaign organization:
Encouraged collaboration amongst the organization.
Opened up passageways for future communication and sharing of information.
Increased transparency and understanding of the product design process.
Could the sprint have been better? Most definitely.
Ways to improve the sprint:
Advocate and allocate for more time.
Structure the meetings based off lessons learned from this round.
The more information is shared and defined, the better the jumping off point.
Facilitate more sprints. Learn more each time.
Design Decision 1: Internal and external A/B test in pairing with design and backend feasibility discussions determined the location of the new feature.
One of the key considerations was dealing with the limited space where the integration was to take place. It came down to the following two options.
Discarded: Option 1
The first option was to add the caption generator form to the bottom of the existing captions tab.
Pros:
User don’t have to navigate to another screen to access the caption generator.
Navigation would be easier on the frontend as well as the backend.
Cons:
The space can become overcrowded with future updates to the generator.
It may already be too crowded.
Selected: Option 2
The second option was to introduce an additional screen upon clicking the action button, Generate Captions.
Pros:
Overcrowding of space issue would be resolved.
Enough space for future feature updates as AI continues to evolve.
Cons:
There is an additional click to get to the new feature.
Cautious of the location of the entry point, make it easy to locate and understand.
Greater lift and increased scope.
The value of a better user experience was high enough to push forward with the increased scope.
Design Decision 2: To accelerate conversation and understanding for customers and internal stakeholders during user interviews and usability testing, I opted for high level screens.
In order to do so, I moved quickly to replicate relevant interfaces as previous design files were unavailable.
Design Decision 3: With the need to ship fast, I used existing patterns to lower the engineering lift.
Existing patterns helped with learnability and familiarity of experience while lowering lift and decreasing scope.
I initiated design reviews with leadership and the internal team to address and confirm experience and interface changes.
We also held a final design review before the release to work out the last bit of design discrepancies that may have been overlooked.
Through design reviews and usability tests, the following changes were made to the main screen of the caption generator.
Cloud Campaign provides a feature in which users are able to create captions platform specifically. Given two options, A/B preference tests guided the decision for option number two.
The first option shows all platforms in caption edit mode even if the platform was not selected.
The second option shows only the selected platforms. This version caused less confusion and higher learnability. Often times, less is more.
CaptionAI was released successfully and is being utilized by a majority of Cloud Campaign customers.
With that, there has been an inflow of positive reviews with regards to the design and the functionality of the feature.
Testimonials
I would recommend CaptionAI to anyone looking to create short captions for social media. It’s innovative, effective, and just plain fun to use.
It speeds everything up for us not having to switch back and forth out of Cloud Campaign to a copywriting tool!
Not only does it save me time by automating the process of coming up with captions, but it also help generate ideas when I’m feeling stuck or uninspired.
Next Steps
Gather user feedback and pinpoint key pain points.
Review updates and additions.
Organize Phase 2 requirements and KPIs.
Reflection
For this project, I stepped out of my comfort zone and reached out to internal team members and engaged more proactively with customers. This in turn led to better defined design requirements, stronger understanding of the technical feasibilities, and clearer insights into user needs. This made it easier to address scope creep and new information to quickly make necessary adjustments.
Things I would do differently:
I would ask more questions about the business decisions to clearly understand the path of the project.
I would advocate for more time invested on the sprint and the research stage.
I would actively seek out feedback from both the internal and external sources to better structure future usability tests and design reviews.
What I learned:
More minds at the beginning proved beneficial in several ways given the startup environment and one person design team
Keep documentation up to date and accessible to all necessary parties
Check in early and often with your engineers to discuss topics from feasibility to final visual design reviews
