Behance App

A new feature : AI talents recommendation to empower recruiters to scout creators easily !
Year
2024
Category
UIUX / Product
Client
Parsons, The new school
Director
Chuck Allison
Role
Product design, user research, UI design, UX design

Behance, a premier portfolio website, hosts a diverse range of creators and their artwork, making it a vital platform for recruiters seeking collaborative design projects. However, its extensive content and manual search process can overwhelm recruiters. To alleviate this, a new AI-driven feature is in development to provide tailored talent recommendations based on project needs. This will simplify the process of finding suitable collaborators for various project owners, catering to recruiters with or without design backgrounds.

Discovery

Context

 My Apple work experience

During my tenure at Apple, sourcing suitable creative talents for art direction projects was consistently my most challenging task. This experience led me to select "portfolio platform" as the focal theme for this project. Drawing from my past needs and experiences, my goal is to enhance the matchmaking process between creators and recruiters on the chosen product.

Industry audit

The uniquness of Behance

Through my experience, I discovered that Behance stands out as one of the few project-based third-party portfolio platforms in the market. It provides deeper project context, enabling recruiters to better understand the design thinking behind creators' work. Due to its unique positioning in the market, I chose Behance as the primary product for this project.

Focus

Recruiters' JTBD

As a portfolio platform, Behance's users includes not only creators but also recruiters. Each user group has different objectives they aim to achieve on the platform. Therefore, I applied the JTBD (Jobs to Be Done) framework to help me focus on the needs and challenges recruiters face when hiring creators.

Product audit

Two current hiring features

To effectively identify the shortcomings of the current recruitment features, I conducted desk research on the current product. I gained a deep understanding of two existing features related to hiring creators: the "Hire Creatives" feature, available on both the app and web versions, and the "Jobs" feature, which is only available on the web version

The flow of " Hire creatives "

In the "Hire Creatives" feature, recruiters can search for project keywords on Behance to find suitable work. They can view project details or visit the creator's profile to see other works and their resume. If the creator fits their needs, recruiters can inquire about collaboration through the chatroom

The flow of " Jobs "

Recruiters can access the "Jobs" feature on the web version, create a job post with relevant details, and publish it after Behance's approval. Once creators apply, recruiters can use Behance's project management backedntool to contact them and discuss project details directly.

Define Problems

Meet users

5 Recruiters' hiring experience

After fully understanding the industry and the current product, I invited five frequent users of Behance who are recruiters for the first round of user interviews. These participants included marketing personnel, entrepreneurs, and design managers, etc. The goal of these interviews was to generally understand the pain points and needs of recruiters when hunting for creators, as well as the challenges they face using the current Behance.

Meta insight 1

It is challenging to find mutiple creators with a similar style to address availability problem

Many interviewees mentioned that in the recruitment process, they typically need to find 2-3 creators with similar styles to prevent availability issues. However, Behance's current search functionality doesn't recommend more creators based on the same style." It is unfriendly for the stage of exploring potential collaborators list. As a result, they tend to use Pinterest and Dribble in the initial stages of a project to find multiple creators because of smarter recommendation.

Meta insight 2

It is challenge to identify creators' abilities to match the project needs

I discovered that non-design recruiters, such as small business founders and marketers, face greater challenges in recruiting creators. They struggle to assess whether a creator can complete their commissioned projects by reviewing portfolios alone. For instance, two non-design background interviewees mentioned that they often rely on their designer friends for professional advice when hiring creators for their projects.

Opportunity

Bridge the gap between the process of "planning" and "recruitment"

Based on recruiters' experiences, I've learned that creative projects typically involve planning, recruitment, and execution phases.  The "Hire Creatives" feature aids recruiters in exploring creators and inquiring about collaborative willingness during recruitment. Additionally, the web version's "Jobs" feature bridges the gap between recruitment and project execution, enabling recruiters to manage the project easily after it has kicked off.

However, from the planning to the recruitment stage, recruiters still heavily rely on their expertise and judgment to select suitable collaborators, which can be demanding and time-consuming. This finding highlights an opportunity for product improvement.

HMW statement

To ensure I effectively focus on the product's opportunities in the subsequent design process, I transformed the two key insights from the user interviews into a How Might We statement. At this stage, the design direction is eatablished, which is to create a feature that streamlines the process from project planning to creator recruitment on Behance.

How might we help recruiters find multiple suitable creators based on the specific needs of the project to ensure its success?

Solution

Low-fi prototype

Idea : Project-based AI creators recommendation

Based on the previous HMW statement, I brainstormed various ideas and visualized them with low-fi prototypes. The idea below allows recruiters to create upcoming creative plans. By detailing the project, AI analyzes requirements and matches them with creators' portfolios, recommending the most suitable talents.

Idea testing

Insight : lower the barrier to describe the project details

To assess the feasibility and verify the value of my idea, I conducted a second round of user interviews. In the interview, participants were divided into two groups based on their backgrounds: Group A included 2 recruiters with design expertise, evaluating the feature's suitability for advanced users, while Group B consisted of 2 recruiters without design backgrounds, assessing the feature's ease of understanding.

Overall, all participants viewed the idea as effective and practical. However, three individuals noted challenges in describing creative projects with text. They all suggested using images directly to illustrate visual styles.

Mid-fine refinement

A plan moodboard enables users to add visual references

According to feedback from users, I've incorporated a new process into the flow of creating a creative plan : the ability for users to build a "plan moodboard" using images from Behance. Recruiters can select suitable images from the Behance images pool to establish visual references specific to their creative plan. These images will directly assist AI optimization, thus refining the final list of recommendations.

Final testing

6 insights to optimize the usability

Finally, I conducted the last round of interviews, including three previous potential users and three new participants. The goal of this stage was to have users complete tasks with a mid-fi prototype—from creating project requirements to finding potential creators—and observe their behavior to identify usability improvements. Ultimately, I identified 5 insights for optimization across three stages of the feature.

Final Design

Homepage

Explore & recommend modes

In the "Hire Creatives" tab, recruiters can switch between two modes at the top of the interface. The first is the "Explore" mode, which is the existing functionality that allows users to broadly explore all creators on Behance. The second is the "Recommend" mode, where the new feature is located.

Phase 1

Create the plan

Entering the "Recommend" mode, users will not see any creator recommendations until they create a new plan. When users click the plus button, they will be prompted to fill in the project title, creative field and project description. Additionally, they can tag the project with categories for easier management of multiple plans in the future.

Phase 2

Add plan moodboards

Next, users can create a visual moodboard for their plan by searching for suitable visual references using keywords or filters. During usability testing, some users expressed concerns about their ability to use precise keywords when searching for images. To address this, the system will automatically generate recommended keywords based on the text descriptions provided.

Phase 3

Receive talents recommendation from AI

When a plan is created, users can access the plan page to see the project description. Under the "Candidates" tab, they can view a list of AI-recommended creators. On each creator's card shows four preview images and a match percentage to help users quickly decide if further exploration was warranted. They can also add interesting creators to their favorites list directly.

Review

Edit plans

Users can switch to the "Moodboards" tab to review previously added visual references for the plan. Each plan page includes a fixed-position floating button, allowing users to edit the plan details and the images in the moodboard. Every adjustment made will continuously refine the system's list of recommended creators.