AI generated job posts for mobile
Overview
| Company | |
| Product | LinkedIn Job Posting (Mobile) |
| My Role | Sole UX Researcher (Contract) |
| Team | 1 Product Manager, 1 Designer, 5 Engineers, 2 Content Designers (content testing support) |
| Timeline | 5 weeks |
| Methods | Moderated usability study, think-aloud protocol, content comprehension testing |
Context
LinkedIn’s job posting tool was one of the most widely used features on the platform, yet it lacked AI-powered capabilities for generating job descriptions. For small-to-medium business (SMB) hiring managers—many of whom don’t have dedicated recruiting teams—writing compelling job descriptions from scratch is a significant challenge. These hirers often struggle with knowing what to include, how to make their posting stand out, and how to attract qualified candidates.
With Generative AI emerging as a powerful tool, there was a clear opportunity to explore how it could simplify the job description creation process for SMBs. This initiative set out to determine whether AI could help hirers generate quality job descriptions with minimal effort—while ensuring the experience felt transparent, trustworthy, and gave users confidence in the output. Critically, this product was designed for mobile and was planned to ship globally across the US, Canada, UK, and India markets.
Research Statement
“How might we leverage AI to simplify the job description process for SMB hirers while ensuring transparency, accuracy, and user trust?”
This research statement gave the team a North Star that went beyond just building a feature—it centered the experience around trust and transparency, which were critical given the novelty of AI-generated content in this context. Unlike the recruiter messaging study, this project also required evaluating a mobile design that would ship globally.
Research Goals
Working with the product team, I defined goals that balanced shipping confidence with genuine user understanding:
Business Goals
- Provide a tool that helps SMB hirers create job descriptions quickly and easily
- Ship an AI-powered feature with confidence, validated by real user feedback
Research Goals
- Evaluate the usability and intuitiveness of the AI-assisted job description mobile design
- Assess users’ comprehension of the AI-generated job description features and available actions
- Assess users’ understanding of copy related to their responsibility for the final output—a critical trust and transparency consideration
Research Questions
Our research questions spanned two dimensions:
Content: “How can LinkedIn increase users’ understanding and trust in the AI-generated job description process and what actions are available?”
Usability: “How can LinkedIn improve the usability of the AI-generated job description tool so users can confidently navigate the entire experience?”
Sub-questions explored discoverability of AI features, comprehension of AI processing states, and clarity of copy around user responsibility for AI-generated content.
Research Methodology
I selected a multi-method approach combining a moderated usability study with think-aloud protocol and a comprehension-based content evaluation. The dual approach was intentional: the usability test captured how users interacted with the designs in real-time, while the content test assessed whether users actually understood the copy, CTAs, and what the AI was doing at each step.
For the content evaluation, I used a structured approach: showing participants static screens at each step of the flow and asking comprehension questions before they interacted. This allowed us to isolate whether confusion came from the copy itself or from the interaction design. Sessions were 45 minutes conducted via Zoom.
Recruitment
I recruited 6 participants with the following criteria:
- LinkedIn members who had previously posted a job on LinkedIn
- Small-to-medium business hiring managers (companies with ≤200 employees)
- Had personally created the job description themselves
- Mixed age and gender representation
- US-based participants
Recruiting SMB hiring managers presented unique challenges—these are busy professionals who don’t always identify as “recruiters,” making them harder to reach through standard recruitment channels.
Sample Tasks & Questions
Content Testing
Participants were shown static screens and asked comprehension questions at each step:
“Based on what you see, please tell me what you think will happen next.”
Usability Task
“You are a hiring manager at your company looking to hire a new Graphic Designer. Using this tool, begin the process of creating a job post for this position.”
Collaboration & Stakeholder Partnership
Building strong relationships with product partners proved essential on this project. When stakeholders trust you and your research process, collaboration becomes seamless. I worked closely with two Content Designers who supported the content testing portion of the study, bringing their expertise in evaluating copy clarity and comprehension. The partnership ensured we could assess both the interaction design and the language simultaneously.
Timeline
This study followed a 5-week rapid research timeline—the standard cadence for LinkedIn’s Rapid Labs:
- Week 1: Kickoff and research plan
- Week 2: Recruitment, prototype preparation, discussion guide development
- Week 3: Conduct research (6 sessions) and data analysis
- Week 4: Craft and deliver presentation
Key Insights & Impact
The research surfaced critical findings that directly influenced the product before launch:
- Users found the feature intuitive — The overall design validated well, giving the team confidence to proceed with launch
- Copy adjustments were needed — Specific wording changes were identified and implemented to improve clarity and build trust in the AI-generated content
- Users wanted more control — Participants expressed a desire for additional customization options, which were added before launch
- Mobile experience validated — The tool worked well on mobile devices, confirming the mobile-first approach was sound
The research directly contributed to a successful feature launch, with findings informing both immediate design changes and the broader product roadmap.
Reflections
- Building strong relationships is essential — When product partners trust you and your process, everything becomes easier. The trust built during this project made rapid decision-making possible and ensured research insights were acted upon quickly.
- International research matters — I performed additional research with international sample sizes to optimize for the global audience. Since the product was designed to reach users in Canada, the UK, and India on top of the US market, understanding cross-cultural differences in how people create and evaluate job descriptions was crucial.
- Usability research isn’t just tactical—it’s also strategic — This project shifted my perspective. While usability studies are often seen as purely tactical, the insights we gathered influenced not just the immediate design but the broader product strategy and roadmap for AI-powered features.