HP AI Scan enhancements
40% faster scans, 35% fewer errors, and 92% OCR accuracy for dependable results.
Project Type
Integrate new AI scanning interactions and functionality.
Team
Product, Research, Design, Platform Strategy, Engineering, Analytics,
Deliverables
Strategy, User Research, Wireframes, Prototypes, Dev Ready Output
Problem overview
The most frequently used scan destinations, Scan to Computer (35.36%) and Scan to Email (26.57%), require a lengthy setup. The current workflow for setting up scan destinations is cumbersome, resulting in a frustrating user experience and limiting the adoption of advanced scanning features. Users struggled with configuring scan settings, leading to inefficiencies in daily use.
Objectives
How might we design a cohesive, future-ready HP scan experience that reduces manual tasks through automation, allows user adjustments and leverages AI/ML to recognize content, automate workflows, and apply intelligent enhancements with over 90% accuracy, continuously improving over time?
User Problems (Ethnographic Qualitative Data)
Users rarely move away from the default settings and are afraid to make changes that may alter future scans.
Many of the scan options are not well understood by customers - they are not experts
Too long to scan multiple files at once, therefore low retention rate
Quantitative data research:
AI user experience trajectory
-
The initial phase of AI adoption is behind us, and most users are now familiar with AI and its general capabilities.
-
Users are less concerned with the fact that a feature is powered by AI; what matters to them are the enhanced features and outcomes AI can deliver.
-
Older users may not prioritize "AI" as a concept, but they do appreciate the ease of use and accessibility that AI enables.
-
The future of AI experiences is leaning toward seamless integration and delivering meaningful results.
2.5%
Innovators
13.5%
Early adopters
34%
Early majority
34%
Late majority
AI experience principles
Guiding principles for all HP AI Experiences
Empower and elevate
HP AI empowers users' goals and capabilities, giving them control to undo, replace, and edit AI-generated content.
Blend in seemlessly
Seamlessly integrates into the user’s workflow, boosting efficiency without disruptions. The UI emphasizes user benefits, not the AI tool itself.
Empower and elevate
HP AI empowers users' goals and capabilities, giving them control to undo, replace, and edit AI-generated content.
Market research
Current AI trends in feature awareness
For feature awareness, AI tools use naming, badging, and grouping.
Emerging AI trend in UI integration
AI should be integrated, not spotlighted, when driving outcomes. By positioning AI as just one tool among many, the focus stays on the user's success, not the technology behind it.
Sketches
Sketching solutions that would help users refine their output and seamlessly choose a destination, whether to share, email, or save.
Testing high-fidelity wireframes
After defining the initial objectives, I focused on validating concepts and testing users' ability to create an AI scan smoothly and quickly.
Testing methodology:
-
Session Type: Remote, moderated interviews conducted via UserTesting.com
-
Location: US
-
Test Stimuli: Figma Prototypes (Photo, Document) Participants were shown both versions in a randomized order
-
Participants: 12 people (6 versatile + 6 personal productivity)
-
All people use the HP Smart app to Scan
-
Objectives:
-
Understand users’ expectations/attitudes towards AI Scan
-
Understand users’ sentiments on the new AI Scan feature
-
Usability for locating the “Scan” function in the app, AI scan vs Manual/Basic scan comprehension, and AI scan flow for photo & document
Prototype wireframe A
Prototype wireframe B
User testing key takeaways
Despite the initial skepticism about integrating AI within scan, all participants were delighted to experience it.
The majority of participants (9/12) preferred Option A for its easy discoverability of features like AI Scan, non-AI Scan, and scan preferences.
The settings icon felt familiar, making it easy to discover.
All participants appreciated the edit options on the scan preview screen below the image and liked that the buttons were outlined to promote enhancements.
Participants were able to navigate without any challenges.
All participants were able to understand that AI would detect the type of file that they want to scan and then suggest relevant edit options accordingly.
Participants who saw this version first were excited to try out AI scan.
Participants felt that they were forced to use AI Scan.
‘Preferences’ with wand icon felt unfamiliar.
Many participants found “Original” misleading in the tabs above the image.
Participants wanted the option to view all three versions. Without AI enhancements, with basic enhancements, and fully enhanced for specific use cases, such as intentionally distorted artistic images.
For documents that needed to be scanned as-is, this version felt inconvenient due to the extra steps required to disable AI settings.
For files that require participants to scan the document as is, this version felt like an inconvenience due to the additional steps it would take to turn off AI settings
Design solution
Users can greatly benefit from automating essential yet intelligent tasks.
Enable users to seamlessly discover and maximize all advanced scan features without manual effort.
Easy switch between cloud & legacy scan.
User does not need to pre-select scan capture mode to get the perfect scan.
Suggests next steps based on AI/ML analysis of the document's context.
People can easily share output through a centralized bottom sheet.
Value
Users achieve the perfect scan on the first try, effortlessly benefiting from the hardware and software imaging pipeline without needing to understand every advanced setting. Suggested options enhance awareness of available features and encourage personalization.
Business impact results
Scan Time Reduction
The enhanced scanning experience cuts scan time by 40%, reducing the average from 15 seconds to just 9 seconds.
AI Scan Utilization Rate
55% of users adopted AI-powered scanning.
Error Reduction Rate
Scan errors, such as incorrect cropping and failed text recognition, decreased by 35%.
OCR and Auto-Optimization Accuracy
Boosts optical character recognition accuracy to 92%, for faster, more reliable workflows
Incredible team members: Mike (UX manager), Whitney (AI Product manager), Durba (Testing), Mark and Jose (Engineering)
