Smart Inputs Beta

COMPANY

Procore Technologies / Smart Inputs Tech + Smart Punch List Beta

MY ROLE

Research, user interviews, identify key themes; design Punch List beta for iOS; set path for future areas of focus; Opportunity Experience for visioning

TEAMMATES

Design: Me, Wojciech Peliks / Product: Luke Prescott / Engineering + AI / Machine Learning Team

SUMMARY

As a construction industry leader, Procore was always looking to improve experiences for customers by trying to speed time on task for phases of construction that have high-volume item creates, in this case, a way to use “Smart Inputs” with AI / ML. This exploration started with a proof of concept, then on the Punch List tool, shipping a beta to continue to see if this was a good long-term investment.

PROBLEM / CHALLENGE
  • Punch walks are time consuming with the hundreds of captures that occur in a short time
  • Existing experience was a Proof of Concept (POC) & there were large feature gaps that prohibited adoption
  • We were designing a new paradigm for capturing hundreds of Punch items in the field so educating users was critical
HOW WE’D MEASURE SUCCESS
  • Improve usability & learnability
  • Improve ease of use
  • Improve AI / ML auto-fill fields accuracy

OUTCOMES

HOW IT WORKED

The “old” way of doing punch walks was to create punch items one by one and complete a related form, which was time consuming.

The new solution leveraged audio, video, and talk-to-text-type functionality to automatically fill key fields/forms. We leveraged our Artificial Intelligence and Machine Learning teams to predict trade, assignees, and titles based on analyzing images or related phrases from issue description.

RESEARCH & ALIGNMENT

I spent extensive time consolidating POC feedback and new feedback for user interviews. This culminated in some spreadsheets and a Miro board to consolidate feedback, which I used to get alignment across cross-functional teams.

UPDATING ON-BOARDING

A new paradigm for doing a Punch walk was going to require more on-boarding to get widespread adoption. This required updating the POC screen to a more formal, accessible on-boarding experience.

ENHANCED LOCATIONS, DRAWINGS, CREATE

I went through a few iterations of sketches and lo-fi prototypes to iron out any kinks with new features like improved locations, adding drawings, and smoothing out the review screens.

SMART INPUTS OPPORTUNITY ASSESSMENT

As the project progressed, it became obvious that there were platform implications that could revolutionize the way the industry captures and creates items throughout the construction process. We explored this with a Opportunity Assessment and Experience with a brief visual story of how the technology could be applied in the future.

THAT'S A WRAP

FINAL OUTCOMES

  • Improved usability & learnability, measured via SUS Test
  • Improved ease of use, measured by user interviews
  • Improved AI / ML auto-fill fields accuracy but limiting low confidence fields, & improved accuracy to 80% on key fields
  • Painted a vision of the future of Smart Inputs in the Procore platform
REFLECTIONS

While working at this thought leader company in the construction industry, I enjoyed exploring new technologies in the lens of improving the speed and accuracy of construction workers in the field with AI and ML. I have noticed that too many companies are shoe-horning technology in to their stack instead of first starting with a problem to solve.