The process of obtaining cost estimates for home improvement projects can be an arduous and time consuming process. This is in part due to the lack of efficiency in the procedure for establishing a scope of work (SOW). In a typical case, following the consultation with a client, the contractor must manually create an SOW and generate the associated cost estimates. This could include going over photos, videos and notes to parse out relevant details. Current processes do not have an option to automate cost estimating.
Angela Yung, Tiffanie Horne
Lo-hi fidelity wireframe, secondary research, user flows, journey mapping
Sketch, Adobe Photoshop
How might we help make the cost estimate process more efficient for home improvement contractors?
ESTi is a smartphone-based voice assistant that converts speech into a ready-to-view, concise scope of work.
Secondary Research Round 1
With an interest in Voice User Interfaces (VUIs) we dove into secondary research centered around the instances where VUIs can improve workflow challenges. We also sought to assess the level of maturity of existing technologies that support voice user interfaces. Our research resulted in the following insights:
Significant factors such as communication repair and conversational collaboration between people and digital assistants impact expectations and perceived value.
Value depends on integration
While an increasing amount of people are adopting the use of voice-driven interfaces, the application of these interfaces are wide-ranging in only a mere handful of industries.
Scattered industry adoption
Users interactions with their phones will depend on the social situation and use of voice interface is avoided in public or around strangers. VUIs are best used when voice interactions are expected.
We then began ideating, ultimately creating 20 concepts to work from. To narrow down to our concept that we would move forward with, we considered the following criteria:
We wanted to chose the concept that aligned the best with the current capabilities of VUI technology.
We wanted to make sure there was a need for this type of product that was currently not met.
Secondary Research Round 2
Based on our criteria, we ultimately decided to narrow down helping independent contractors that are working on home renovation projects. We conducted further secondary research on this space and came to the following insights:
Prolonged Turnover Rate
Contractors often have back to back appointments, which makes it difficult to provide the scope of work and cost estimates immediately after a job, prolonging the process.
Current practices do not incorporate automaticity, therefore, requiring active documentation from the contractor during appointments reducing client engagement.
Reduced Client Engagement
Time Intensive Process
Contractors carry out administrative intensive, time-consuming tasks, which reduces the efficiency of their workflow.
Before beginning the design process, we used a journey map to walk through our user’s needs and the interactions that they would have with the product.
Micah is an independent contractor who specializes in home improvement. Business is booming but Micah is now having trouble keeping up with providing quotes to her clients.
Micah is able to focus on her client during the session rather than focus on the documentation process.
With ESTi. Micah is able to input all her appointments onto the app ahead of time to stay organized. When the appointment starts, she goes to her app and taps on the client's appointment.
ESTi listens in the background to the appointment and generates ESTimates based on the information that Micah and the client talked about.
Micah will have a chance to go through the summary to confirm or change any of the items.
She can either save this review for later or complete it on site and send the ESTimate to the client right away.
ESTi’s system implementation is comprised of speech recognition and natural language processing (NLP). ESTi uses an API to complete the task of translating speech to text. NLP is employed to determine and analyze the intent and entities embedded in the dialogue exchange.
The first component of ESTi’s system is speech recognition. The user’s mobile phone microphone is used to convert the physical sound into an electrical signal and then into digital data with an analog-to-digital converter.
Gathering Representative Audio
A collection of samples consisting of 450 conversations will make up the origination library and will support language models with the task of reducing word error rate
ESTi requires a speech processing application programming interface (API) that can address transforming human speech to text and accommodate intermittent internet connectivity.
Natural Language Processing
ESTi utilizes an NLP service to find insights and relationships between the real-time stream of audio received and the residential construction domain vocabulary and terminology.
ESTi sources data from construction price indexes and commercial costbooks, such as those published by Craftsman Book Company, a publisher of over hundreds of technical and professional references for the construction industry.
We developed a paper prototype and went on to test with potential users. Through testing the paper prototype, we reached three insights to drive our final design direction.
While transcription can create ease, any potential for error causes users to worry about accuracy.
As a requirement for forming trust, contractors want to give their final approval.
User approval is paramount
Users feel that the product takes some of the pressure off because the documentation process becomes less involved allowing them to focus on their client’s needs
The home screen provides a list of the user's appointments, with the upcoming appointments appear at the top in chronological order and the past appointments appearing below it in reverse chronological order.
Real time scope of work
Keywords are said to generate a list of items that were discussed during the session. This list will be updated in real time and can be viewed on the screen during the session.
Easy access to start recording
A circle button at the bottom can be pressed to start a session or the user can simply say "start session" to start the recording.
Edit and customize reports
The app then generates a summary of scope of work, giving quantity and estimates. After the session is over, the user is able to send it to the client or save the report and come back to it later.