Scibler Smart Scheduler
An automatic scheduling app that provides AI powered one-tap scheduling for people on the go.
The Smart Scheduler is intended to automatically intercept meeting requests and meeting related negotiations in your email, know your preferences and schedule, check your availability, prepare responses to meeting requests and then allow you to quickly review automatic responses and reply with just one tap.
About the project
In our current pervasive digital environment, people encounter a deluge of digital communication and personal data. To get things done, one must frequently wade through large amounts of data in order to retrieve the right information at the right time. This is often an ongoing challenge. In 2013, Scibler, a venture funded, cloud-hosted startup service was looking to develop new state-of-the-art technologies for analysis, retrieval and navigation of personal data and communication.
The potential exists for technology to expose exciting possibilities and revolutionize the way in which people might interact with their personal data in the future. As a designer this offered an opportunity to partner with a team of AI, Machine Learning and NLP engineers to invent the next-generation of personal-data user experiences.
My role
As the design director, I was responsible for the end to end design evolution of an AI based app from ground-up. Which included initiating and driving a human-centric ideation and design process with the team, iterating on concepts, designing the UX for a prototype web plugin and helping define a minimal viable product launched on the iPhone app store. I helped drive design strategy, collaborated on product plan and strategy, led the creative direction, as well as provided hands-on creative guidance for all aspects of the product including social media and marketing.
Process
Conceptualizing and building a product that improves how digital information is consumed by people requires an understanding of the challenges they face in dealing with the constant stream of digital data in their everyday lives.
Research
There is no doubt that the amount of information we encounter daily is overwhelming and often hinders our ability to get things done efficiently. As technology workers ourselves we all had first hand experience in dealing with our own data, but it was important for us to go beyond ourselves and learn more. I initiated a short user research study with the team to better understand the people we would be building solutions for and their current environment. Information was gathered via surveys, in-person and phone interviews. The goal was to gather information on real situational challenges faced by people in managing their data with respect to work as well as their personal lives. The team began to engage with and interview a wide spectrum of people where we observed and documented their concerns with managing their personal data and communications. These interviews also helped sensitize all of us to the lives of the people for whom we would be envisioning solutions.
A fair amount of time was also spent analyzing and understanding the competitive market space. As a team we looked at products currently used by the people we had interviewed, as well as new products in the market that might address their concerns and needs. Competitive products were evaluated and opportunities identified. Analysis revealed that there were several products in the market aimed at making people more efficient and productive with their digital data. These fell into the following categories - integrated search experiences across data, reminders and personal assistants and email organizers. User feedback suggested that these product had not yet met people's needs completely and that there was still opportunity in simplifying and automating some of these experiences.
Analyzing data, defining personas and user journeys
I led the team through an exercise to sort the data gathered and categorize user characteristics, The information gathered from the research exercise was analyzed and distilled to define three typical types of personas.
The outcome was a set of three broad persona categories that addressed generic profiles of people who wear multiple hats throughout the day. Their daily activities may include tasks, interaction and communication related to family + work, business clients + social contacts or purely business.
We felt that these user types would be suitably benefited by any AI and machine learning based predictive solutions as it would help to streamline their incoming data and potentially provide them with efficient tools to manage and complete tasks and activities. Efficiency in managing tasks would in essence allow them more time to - spend with family, do the things they like to do, or focus on their business.
User journeys were developed for each persona and typical user scenarios identified.
Identifying pain points and ideation
Information gleaned from the research, persona and scenario definition process allowed us to collaborate on, and define a set of pain points relevant to our target audience. Using the pain points and scenarios as launch points, I facilitated a deep dive design charrette with the team, where we collaboratively brainstormed a broad set of ideas.
The many ideas that were generated were then mapped to user scenarios and categorized into potential focus areas.
Prioritizing scenarios and concepts
A couple of recurring threads in the user journeys were identified. We found that people have a difficult time finding and keeping track of the things they need to do in their email. They also find it difficult to complete tasks and actions that arise in email conversations. Data gathered by the team suggested that traffic in a typical business account exceeds 100 emails a day. Fifty percent of which are opened first on their mobile device. This results in people often needing to revisit email messages on their desktops/laptops in order to complete related actions and they often tend to overlook deadlines.
Current solutions in the market were all manual. Email apps such as Mailbox, Triage, Handle and task apps such as Any.do, Wunderlist, Teux Deux etc. We also found that these were solutions adopted by only the most organized of people who are willing to spend time using apps to manage their data.
Opportunity existed for automation to unlock a much bigger market by making the entire experience far more streamlined.
Developing user models for automatic task discovery and completion with AI and machine learning
I began to iterate on multiple user models building on the concept of an integrated end to end experience that was intended to-
a. discover email messages that have associated tasks,
b. find related information for task completion,
c. provide the shortest path to completion.
Concepts were explored and evaluated heuristically for technical feasibility and ease of use.
Testing a prototype web plug-in for gmail
While exploring the product manifestation, it became apparent that it was necessary to build a working prototype using real user data to validate and test various aspects of the product concept. We had identified a set of users (many of whom we had interviewed in the research phase), to whom we would be deploying the prototype product. We would be using their gmail data to provide the experiences that would address a couple of pain points identified by them. 1. Automatic alerts for actions and tasks in email, 2. Search and discovery of related information.The logical manifestation for this prototype would be a gmail plug-in.
This lean web based product was also intended as a proof of concept and needed to be built quickly and efficiently. For this purpose, I designed a light weight, bare bones user experience that captured the fundamental user scenarios that we had prioritized.
Building and launching a minimal viable product
What we learned and the data we gathered from testing the web plug-in helped evolve the product and refine a minimal viable product. Alerts that allow the user to discover potential information that might otherwise slip through the cracks was of definite value and identified as giving the product unique selling power. Whereas search and related information required better context to capture user's attention. Federation of personal and work email accounts was found to be useful. Anecdotal evidence validated types of alerts such as meetings, follow-ups, events and bills that people often miss in their email. It was important for users to identify different types of alerts and be able to act upon them on the go.
The general characteristics of our target users ("Life Jugglers") were people with multiple roles and responsibilities who frequently juggle home, work and social life. For whom email is a hub of everyday workflow. They struggle to keep track of things needing attention in their email and utilize calendars to schedule important events and activities and don’t really have much time to manage and track all activities. They frequently need the right information handy to get things done while at home, work or elsewhere and often share and delegate tasks to others (both home and work).
The product plan focused on individual productivity and some smart collaboration capabilities, with an aim to
target individuals to adopt Scibler. These individuals were identified as small business owners (who exemplify all of the characteristics of our target audience).
The product MVP
a mobile app built for iPhone would focus on alerts for automatic discovery of three primary types of information-
1. meetings and meeting requests,
2.tasks (including follow ups to meeting requests or scheduling), 3. feeds (updates from people important to me).
Designing the user experience
The app which was essentially a dynamic calendar, would automatically discover information relevant to your day. The UI was designed to make it easy for the user to identify types of alerts and take action appropriately.
Information was prioritized in the following order - meetings already scheduled for the day, proposed meetings or actions requiring your response or someone else's response and people important to your day. The default app view (a list view) was designed to accommodate the above order of priority and the interaction model designed so that actions could be completed in a minimal number of steps.
Although the app was a souped up AI powered rendition, it provided fundamental features of a regular calendar. Along with the ability to quickly discover relevant information through the coming days/weeks/months, the user experience also made it easy to schedule and add new items to the calendar.
I ensured that the design followed iOS guidelines.
Defining branding and aesthetic direction
The visual design goal was fairly straightfoward. There was a need to keep the design simple and clean with the intent to enhance performance and reduce effort on design and development as well as to allow for quick turnaround on refinements based on feedback.
Given our understanding of the users we hoped to target, and the kinds of apps and tools in the market, I led the visual design effort by creating a "mood board" to help direct the product's aesthetic style. Fonts, colors and other details were tested at various points using A/B testing. A direction was chosen that was reminiscent of old classic calendars in basic red, white, black and gray. I provided guidance, visual critique and sometimes hands-on design input to help create a simple yet sophisticated visual direction that defined clear information hierarchy using color, typography and iconography. The design was contemporary and clean. Feedback on the aesthetic was very positive and the impact it had on performance was noticeable.
Visual design evolution
Launching "Scibler Smart Scheduler" on the iOS app store
The beta version of the product, offered us the opportunity to test and gather feedback on the overall user experience, including visual direction and interaction. The UX was tested multiple times at various stages using split testing, analyzing user behavior and gathering feedback in the form of follow up questionnaires.
A refined version of the app was launched on the app store that encapsulated feedback from the beta version. I worked with the team to streamline and re-envision the experience as "Scibler Smart Scheduler". Scenarios and actions related primarily to appointments and scheduling related tasks were prioritized.
Detailed walk-through of the product
A series of screens highlighting key aspects of the experience were provided to assist the users while getting started.
Challenges and outcome
A key challenge we faced as a lean start-up was resources. A small team with a limited number of people working on a very tight schedule, required that each member of the team wear multiple hats and perform multiple roles. As the design director and primary owner of the user experience, I not only had the opportunity to define, drive and direct all aspects of the product's design, I also had the opportunity engage in product planning, testing, marketing and social media.
In some instances I was called upon to tread in areas that I am not usually required to, and in others I had to refresh skills that I hadn't tested in a while.
The drawback to working in a lean start-up team also means that all that you envision for a product, doesn't necessarily make the cut. Often only because there aren't enough resources to build it. The necessity to adjust and shift to changing directions requires a degree of agility and resilience. As a designer, this can be disappointing at times. But the dynamic nature of the experience can also make it exciting and satisfying.
My contribution to the product ended sometime after I had delivered a complete user experience for the final product launch. There were a few rounds of continued refinements and improvements to the product. But by the time we had launched the product, there were many significant and large players in the AI space grappling with similar product concepts. It was no longer viable for a small start up to hope to compete with the likes of a Google or a Microsoft and it became imperative for the company to change course.