A retail kiosk bot 🛒

A telecommunications client wanted to develop an internal innovation team specifically focused on enabling artificial intelligence within the organization. I was tasked with creating this team and working with the client to pilot this program.

problem

How might we…create an innovation (research + prototyping) culture at this company?

method (5 months)

• Phase 1: Innovation Team Foundation

• Phase 2: Artificial Intelligence Pilot in Retail Stores

my role:

I was the Innovation and Research Lead on this project, leading 3 other team members (designers, developers, research) to work with the client stakeholders to design an innovation squad and pipeline of potential projects.

phase 1 solution: Methodology Design

The first thing we needed to do is understand the innovation capabilities, existing projects, goals, and landscape at the client. After several meetings and a workshop session, a common goal was identified:

The Innovation Team will identify use cases and go through the human-centered design to perform research, design prototypes, and test in the marketplace to identify the value for future artificial intelligence work.

With this, we prioritized the client’s internal goals, their capabilities, organization structure, and customized an innovation squad and project pipeline based on our experience of successful research + testing teams.

phase 2 solution: Artificial Intelligence Pilot in Retail Stores

One of our client’s priorities was to introduce artificial intelligence in retail stores to reduce wait time for customers. My team was asked to take over for another experimentation team’s work, so we had to work with their existing prototype and constraints.

The prototype concept was a “wizard-of-oz” kiosk where someone would play the role of the artificial intelligent kiosk that the customer speaks to. The customer would get help through a digital kiosk (powered by a human) and then get transferred to a live store representative once a sale/promotion was identified. This was a high-fidelity prototype that could be tested in the market to collect data and took around ~2 weeks to develop.

We quickly realized there were a few issues with the existing prototype:

• The original developers created the concept without talking to real customers or on-the-ground stakeholders (retail representatives)

• KPI’s for the prototype were not based of store data/analytics due to lack of data access

• Key Cross-Functional Stakeholders were not aware of the prototype and how the experiment would actually run
  ex) Legal/Security was not included

Upon our recommendation, we had to backtrack this prototype testing process because the client didn’t want this prototype to run based on all the roadblocks. To do this, we decided to:

• Develop a partnership with a smaller + specific geographic area’s team with stores that are used for experimentation in order to get this prototype into the hands of real users

• Isolate certain assumptions in the experiment (customer interest, rep interest, opt in rate) and run mini-tests to validate

• Design a Business Case on how much added revenue/savings this prototype would bring based on actual data

We went out on the field in Georgia and acted as Sales Representatives to understand customer behavior, representative’s priorities, and validate this concept. We asked customer’s to interact with a low-fidelity prototype of the concept (one Wireframe screen) and iterated our prototypes as a result.

outcome

• What was meant to be a quick prototype experiment with artificial intelligence in retail stores took longer due to lack of proper governance, oversight, and processes. Our new innovation squad model addressed those previous issues

• Our research showed that a retail AI experience was something customers/representatives and Business Stakeholders were actually not interested in. This idea was quickly scrapped before it went to market

• Based on the ethnographic research, a new “north star” concept for the retail experience was designed that the team is now investing in

• The innovation team pivoted their focus to an artificial intelligent promotions experience that went live in early 2020

Additional Samples of Work Available Upon Request