Why I decided to work on this project?
The idea to work on this project came of out necessity. There is a very good Salon near my office in Dublin which provides excellent service but they have minimal digital presence like a website or booking system.
At the moment you have to physically go to the Salon to gather all kind of data like getting a reservation, closing and opening times and price ranges.
“I wish there was a solution which can enhance the customer experience for such small salons by automating the booking and answering frequently asked questions.”
What is the main problem I am trying to solve?
Currently the customers have to physically visit the salon to get a booking or go get relevant information.
“I want to automate the booking system and answering frequently asked questions.”
Who are the users I am solving this problem for?
- Customers with limited technology knowledge but familiar with facebook messenger
I decided to design a facebook chatbot to help with booking and answer simple queries.
Why a Facebook Messenger chatbot?
Facebook now lets your followers have an automated conversation with you over Facebook messenger.
This means that your clients can immediately get info about directions, opening hours, services, book an appointment and more right over messenger without you needing to personally respond to each question. Messenger bots make this possible.
Plus, they’re super cool.
Points to consider —
- Can certain customer service questions be automated?
- How much time do you spend answering questions on Facebook every day?
- Do you have a staff member that’s always immediately available to answer questions?
- And is this something that your target audience will understand and be interested in?
Advantages of having a chatbot —
- It’s available 24/7, even when a receptionist can’t pick up the phone to answer questions or you don’t have a receptionist.
- Clients and lead can engage with you right from social, where they have most active presence and after seeing your photos and videos are more likely to interact..
- It shows that you use the latest technology and are catering to a certain younger target.
- You can encourage more online booking requests: when things are immediate they are more likely to happen and as it is integrated with your facebook, you can push custom offers based on their preference.
User Research — Understand the problem
Who are the users?
What are their pain and gain points?
How will the solution attempt to solve the pain points?
What is the user journey?
Understand the conversational aspects like how a booking is done and what information customers ask for.
Design — Chatbot design
The UI part of the chatbot involves the visual part.
User Testing — How effective is the solution?
Set the KPIs (Key Performance Index) to measure the success of the solution.
Conduct usability testing with some users to understand the knowledge gap which exist in the solution.
Learn from each step and iterate the product for constant improvement.
How to include A.I aspects to the solution, given most people get haircut on a periodic basis (Every 15 days or monthly), how the chatbot will gather data on the user and learn patterns.
As personally I visit the Salon, it gave me lots of insights of the customers who regularly visit the Salon. So the initial research method is Field study. The important points I noticed was -
- Most customers were not native English speakers and majority of them were East Europeans.
- The salon is always busy with customers as they allowed Walk-ins. I visited them thrice so far but always have to wait for a while to get my haircut. As they don’t have a formal booking system or technical infrastructure.
- All of the employees of the Salon are immigrants and speak Polish or Croatian.
- They have regular customers and the customer experience post haircut is good and everyone is treated really well.
- The salon is not digitalised as they don’t accept digital payment and they don’t have a computer. But all of the employees have cellphones.
Based on the Field study, following points can be inferred to be included in the solution -
- The solution should be mobile friendly as the Salon employees doesn’t own a computer in the store.
- As the Salon employees are not tech savvy, the solution should integrate with their existing technical knowledge such as use of facebook messenger.
Based on the field study, here is a persona of a user which best captures the conditions observed.
User Needs —
A Chatbot which can book reservations, give information about opening and closing times, price range.
The chatbot should integrate with Facebook messenger as most users have a Facebook account.
Constructing a user journey with a chatbot involves not only understanding the user’s state of being and possible steps, but also mapping potential answers and scenarios the chatbot must address
Common Points to consider
- If the bot isn’t done well, it may confuse potential clients and put them off.
- The bot could seen false and not friendly if you don’t inject your own sense of humour and style into it
Points based on research
One biggest challenge will be to make it available in multiple language as most users were not highly proficient in English, but the field study suggest they have basic level of English so the conversations has to be really simple and easy to understand. In future, the bot may be able to converse in different languages. At the moment, the solution is implemented only in English.
The biggest challenge is to make sure that users are opting for inputs they are presented with and keep the conversation contextual.
When someone is talking with a Salon bot, it is important to understand why they are interacting ? What are their needs? (i.e. No customer will order pizza from a Salon bot, but the bot must be well designed from conversational aspect to guide user to keep the conversation contextual.)
The solution must make sure that the customer is keeping the conversation contextual and presented with all the information he/she will need.
The user journey tries to handle such other user input and keep the conversation relevant to the use cases.
As the solution is for Facebook messenger, the UI is standard as per Messenger guidelines.
Improvements and Iterations
Through user testing, I was able to uncover other questions which the users may ask such as about hair products. The test participants were comfortable with the conversation flow and appreciated the contextual conversation as they visited the bot messenger for very specific reasons and it was able to answer their questions and meet their requirements in an efficient way.
I am continuously making improvement to the conversation flow and adding more humane touch to the content of the conversation to tackle user inputs in most delightful ways.
“Simple and efficient.” — Test participant
“It will be killer if it suggests me a hairstyle like with Snapchat filters, lol!” — Test Participant
KPI to measure once the solution is live -
A good KPI is the measure of successful customer journey completion. For e.g. , in Uber, the KPI is number of rides successfully completed.
The north star metric for this solution will be Number of customer who had haircut after booking via the chatbot.
Minimum Viable Product vs. Ultimate Goal
During the brainstorming phase, I found several features gap between the minimum viable product with specific response to specific choice and ideal AI chatbot which can anticipate user needs.
Some future directions includes-
- Based on user’s Facebook interactions, giving suggestions to the users. For example, if the user is reading a post titled “Top 10 hairstyles currently trending”, the chatbot will text the user if he/she needs a haircut and even better will show the latest hairstyles available in the Salon.
- Possible integration with AR features that will help the user to try hairstyles before getting one.
- Smart scheduling based on previous patterns
Project Probable Timeline
- Chatfuel AI Platform / Dialogflow and Node.js NLP
- Facebook Account
- Facebook Messenger