“Covid has shifted the world to a digital first mode. Today every consumer’s relationship with brands is built on the digital medium, through digital first products.”
Today Digital is the new human! It’s every company’s survival tool against extinction. But with so many different software tools out there, how do you even make the right calls? What is good for your business? How do you stitch together different solutions to achieve your business objectives? And how do accelerate this process so that you don’t get left behind?
Feeling stuck with these questions?
We have been there before and we completely feel you!
It was the biggest motivation behind creating this guide. To cover every aspect of a modern day digital customer engagement solution, starting with the basics, laying out the software tools of choice today, trending strategies and technology stacks we have got it all covered.
This is our complete guide to digital customer engagement using Conversational tools in 2021 and beyond. Born, proven and time tested through the pandemic induced digital era, startinging March2020.
In this comprehensive guide you’ll learn:
- About modern day Digital-First Tech stacks that companies use today to win in the digital space.
- How to get your business Live on high density, everyday use digital channels like WhatsApp, Facebook, Instagram – the Backbone of Conversational Automation.
- How to serve a million customers with zero in-person interactions, by automating customer engagement touch points, with “multi talented” Bots?
- The top software tools used for digital business.
- Technology stack templates which you can start using right away.
💡Is it just another article in the Conversational automation space? Well, no. We won’t flood you with tons of options, software suggestions and tactics. Rather we publish the EXACT actionable stuff that 200+ companies, many of whom are our customers, use to generate insane business outcomes leveraging Conversational Automation. And if you replicate the strategies and guides, you’re bound to get amazing results too.
So if you want to turn your Business into a Digital First Growth Machine with a Conversational Makeover, you’ll love this guide.
Let’s get started.
TABLE OF CONTENTS
- Anatomy Of Conversational Customer Engagement
- Bots – The Backbone of Conversational Systems
- Power Boosters – Expanding Platform Capabilities
- Security – Keeping Everything Safe
- Winning Conversational Templates
1. Anatomy of Conversational Customer Engagement
1.1 Old Way v/s New Way
One of the most significant things to have happened in recent times is the awakening to the fact that the classic way of doing business is obsolete.
The now almost defunct way of doing business involved:
- Organising capital.
- Selecting outlet locations.
- Setting up a store.
- Setting up customer care centres.
- Hiring a large number of people to manage the show.
This was a slow, time consuming, expensive process involving a lot of moving parts. And extremely difficult to scale.
Enter Direct to Consumer, Digital First Business Model.
Entrepreneurs and Business owners have now discovered a brand new way of doing business which is much more in sync with how the world operates today.
A refreshingly new way of doing business to win in the Digital First Era.
- It does not require too much capital.
- Messaging and social channels can be enabled as sales and customer care “channels”.
- It is easier to reach a large distributed buyer segment from day 1.
- Requires very less manpower to operate, with an Automation First approach.
- Helps reach customers directly, give personalised attention and get raw, unfiltered feedback and insights.
- Easy to scale rapidly.
1.2 The rise of Messaging & Social
Social and messaging platforms have become default channels where people go about their daily lives today. Right from relationships to finance, entertainment to news, travel and work it’s all happening on these high density digital channels.
And now they are looking to engage with brands and organisations on these channels for convenience. It is imperative that brands and organizations must scale their digital capabilities to connect with their customers on the social and messaging platforms their customers use every day.
Today social and messaging apps like WhatsApp, Facebook, Twitter etc have virtually replaced several of the known players of digital customer engagement from the bygone era.
❌ Web Portals
❌ Call Centres / IVRs
❌ Contact Centre Walk-Ins
The Messaging and Social revolution in the business segment is well and truly on.
1.3 Getting Started
To introduce a Conversational Customer Engagement Stack into your business operations you require 3 key things
- A clear goal
- A slick strategy
- Informed choice of tools
Modern Conversational systems can help a company automate every department within the organisation. However without clearly articulated macro and micro level goals it can become overwhelming to drive this transformation. For example in a consumer electronic company a macro level goal could be Customer Service Automation but within that domain micro level goals could be Support Automation, Extended Warranty Process Automation, Feedback Automation etc.
Here is a chart with various Conversational Use Cases 👇
Once a goal is clearly defined and use case chosen, it’s time to strategise.
Here are 3 key principles that help you formulate a robust Conversational Customer Engagement GTM strategy.
👉 Target Audience
Defining a a core group of your customers with specific desires, geography and behaviour is key to building a compelling Conversational Customer Engagement Channel. It helps you design the look and feel of the channel, decide what functionalities to include and which messaging and social channels to target as a delivery vehicle.
Here is a chart showing the most popular messaging apps across countries. A quick glance shows that if your target audience is in the USA then the primary delivery platform should be Facebook Messenger, the No.1 channel in the region.
👉 “Personality of Your Conversation Channel”
Every brand/company has a certain tone, language and conversation style in their day to day activities. It is important to reflect that in your Conversation channel. With nearly 100% automated dialogues likely to take place between your customers and your systems, getting your brand personality right in your conversation channel is critical.
Factors that you can take into consideration are language, tone (humour, witty, quirky, professional, etc), long text responses v/s short text responses, adding warmth etc.
Here are a few illustrations of Personality driven Chat Bots:
👉 Choosing the Right Tools
The choice of tools is directly proportional to the desired goals and specific use cases. Modern day Conversational Stacks are usually a combination of different business applications(like service ticketing tools, live chat applications, analytics tools, etc) ERP systems, Commerce applications, messaging and social apps, Bot Platforms, NLP platforms, Automation tools, Payment gateway etc.
We will be covering various tools in this article that you can mix and match to craft your perfect Conversational Customer Engagement Stack.
Here is a glimpse of the different type of tools in a Conversational Stack:
Chapter 2: Bots – The Backbone of Conversational Systems
2.1 Types of Bot
Bots can be of varying types and are usually constructed based on the use cases. Let us take a look at the 4 main types of Bots, with examples of leading brands and how they capture the essence of each type of Bot.
Bot — is an automated program that runs over the internet, typically behind the scenes. This places it at the bottom of the scale of intelligence and user engagement. It can do minimal functions like sending out information or capturing data.
Chat Bot — can look for patterns in text and respond with automated answers, usually best at very simple exchanges. They are usually also simple menu based Bots with multiple functionalities. In these cases they only need user inputs in the form of menu and sub-menu number selection to respond.
Conversational Bot — engages consumers with its ability to have free flowing conversations, precisely understanding the intent of the conversation, and then guiding users to complete their goal. A conversational agent is always in self-learning mode and uses basic AI and Machine learning to get better over time.
Cognitive Bot — knows the consumer deeply and uses that knowledge to anticipate needs, placing it at the top of the scale for intelligence and user engagement. These types of Bots are usually heavy on AI, machine learning and Sentiment analysis. They are typically used for high end, deep tech activities in the fields of enterprise grade customer interactions, back office automation, healthcare, finance etc.
2.2 Building A Bot
While there are various ways of building a bot we will take you through a step by step guide of the most commonly used Bot building process with a real world example.
For the purpose of this segment we will use the example of a healthcare bot for Case Hospital.
Step 1: Define the Bot Structure
Bot are typically structured in 2 ways: A. Menu Driven Structure B. Conversational Structure
Menu Driven Structure
This structure is followed by majority of the Bots with its user friendliness and intuitiveness. An example of the Menu Driven structure is the Case Hospital Bot. This show the different menus (or services) which the hospital is providing though their Conversational Channel.
Conversational structures try and replicate an actual human conversation pattern. For example imagine calling up customer care or walking into a store and being received, greeted and shown around by a representative. This is the exact process followed in a conversational flow.
Here is an example of a Bot used in the beauty segment with a conversational flow structure:
Step 2: Give your Bot a name and identity picture.
To give a human touch to Bot many brands prefer giving the Bots a Name. Case Hospital has named their Bot “Casey”. In addition to a name you can also add a picture to make your Bot recognisable. Most brands prefer to use their logo as the identity picture, however some also creatively use Avatars to give a unique feel.
Step 3: Define Services.
The advancements made in the last few years in terms of Bot technology now makes it possible to offer an incredible amount of services through this channel. Bots are now replacing entire contact centres, web portals, walk-in counters and in some cases even very sophisticated mobile apps.
Here is list of services that one can provide through Bots:
👉 Information Helpdesk
👉 Automated Query Handling
👉 Appointment Booking
👉 Service Ticket Management
👉 Live Human Chat For Support & Advisory
👉 Commerce – New Sales & Order Management
👉 Commerce – Add On Selling
👉 Bill Payments
👉 Account Management (e.g. Broadband Plan Management)
👉 Financial Transactions
👉 Digital KYC
👉 Claims Processing & Salvage Auction
👉 Agent Based Insurance Quotations
👉 Student and Course Management
👉 Notification and Alert Services
👉 Ai Based Advisory Services (Beauty, Finance)
👉 Lead Generation Activities
👉 Feedback and Customer Insights
You need to choose one which suits your business and is aligned with the goals set out earlier. Once you decide on the services to be offered it has to translated into a crisp menu. A well designed menu goes a long way in setting the stage for a captivating Bot experience.
Here are a few examples of Bots Menus, ranging from simple to visually rich.
Step 4: Assemble Content
Once you have defined the services to be given through the Bot channel and finalised a look and feel its time to gather content. The services (Menus) which collect information would not require any content. However various other services can be filled with content to give users a wholesome experience.
All Bots can accommodate the following types of content:
👉 Media Files (Video, Audio, Images)
👉 Plain Text and Rich Text Content
👉 Urls for Content Display
👉 Document Files (PDF/DOCX/XLSX)
👉 Live Streaming Widgets
👉 High Volume of Question Answers
Content can be tactically used in various ways. They can be used to provide instant information, step by step guides, detailed specifications of products and promotional/educational/entertainment material.
Here are a few examples of how pre-loaded content looks like on Bot interfaces:
Step 5: Design Conversation Flow
Once all the ingredients have been assembled, its now time to bring it all together and design a simple yet captivating Bot Flow.
This is easily the most important phase of Bot design process as it involves a lot of moving parts. It is during this phase that one may need the services of several specialists. Some of the services which we can avail and will add value to the production level Bot are:
👉 Conversational Flow Designers: They help in crafting user journeys and simplifying complex interaction flows helping to get users to intended results sooner.
👉 Conversational Script Writers: Professional writers specialising in Bot scripts. Once a Bot conversation theme is defined they get down to not just writing a script but also infusing the Brand’s tone, theme and style into the script.
With channels like Upwork, Freelancer, Fiverr, Boldly one can easily find these experts on hourly contracts.
To get started with a Bot flow it’s important to get a flow chart in place. A flow chart helps visualise the various journeys a user might undertake, identify different interaction touch points and craft intuitive user journeys. A flow chart also helps in defining the shortest path to achieve goals.
Here is an example of simple commerce bot flow chart:
Most Conversational Platforms have an in-built Bot Flow designer with multiple functions enabling 100s of design permutations and combinations. These Bot Flow designers usually have a drag and drop interface or a flat, step by step non-linear interface, with the latter becoming more popular as it allows more flexibility.
Bot Flow designers allow admins a large collection of tools to design the Bot and Flow of their choice. Some of the tools available at their disposal are:
👉 Step by step Linear and Non-Linear flow creators.
👉 Branch creators.
👉 Single line, multi text, multiple response input fields.
👉 Mathematical formulas, Multi-Choice Questions.
👉 File Inputs (Documents, Media, Pictures)
👉 Header and Footer Messages
👉 Language Options
👉 Rich Text Boxes for Adding Emojis, Bold Text, URLs
👉 Payment Gateways
👉 Location Data
👉 API Blocks – To fetch or push data in real time to/from 3rd party applications.
Here are some screenshots of Bot Conversation Flow Designer interfaces:
Step 6: Define Activation Keywords
Every Bot needs an activation call sign to “wake it up” for conversations. These activation call signs can be simple alphanumeric keywords, numerical combinations or special characters.
Some of the most commonly used activation keywords are:
👉 Talk To Me
2.3 Training Bots – With Basic AI
Now that you have designed your Bot its time to up its “Talent Level”. Modern technology allows Bots to be skilled up, just like humans. To incredibly high levels of Smartness. This injection of AI not only makes the Bots smarter but also ensures that they start behaving like humans in real life.
In this section we take a look at one key subset of AI, called NLP (Natural Language Processing) which makes it possible for any Bot to easily understand natural human language, in multiple languages.
Natural Language Processing
The easiest examples of NLP in action are virtual assistants, like Google Assist, Siri, and Alexa. These sophisticated systems with in-built NLP understand and translate the human language. Remember the times when you asked Siri to “Book an Uber, for me from home to St.Cathedral Church on Robinhood Street.” or Alexa to “Play Coldplay Station in shuffle mode“. That is NLP in action.
Now imagine having such free flowing natural conversations with your Bot. And your Bot understanding (in English and various other languages), interpreting the intent behind it and responding with an appropriate contextual answers. Some NLP engines are going beyond understanding intents to analysing “Moods” with Sentiment Analysis. Sentiment Analysis identifies emotions in text and classifies user inputs into 3 major categories: positive, negative, or neutral. Based on this assessment Bot response is further improved in terms of the quality and tone of the response.
The top 2 tools of choice as far as NLP systems go are:
Dialog Flow By Google
Offers capabilities to build speech to text, human language understanding tools. Supports majority of the platforms like Facebook Messenger, Slack, Alexa, and Google Assistant. It supports multiple input channels where Bots are usually deployed. Currently supports 20+ languages.
Lex By Amazon
is an AWS product for building conversational interfaces into applications using voice and text. Amazon Lex, uses the technology layer of its popular Amazon Alexa engine. It supports less languages than Dialog Flow, but is equally capable as far its Natural Language processing capabilities are concerned.
An Overview of Where an NLP System fits in the Conversational Platform
2.4 Deploying and Distributing Bots
Your super talented Bot is now waiting to be unleashed to serve your customers. It’s decision time for you. Choosing the channel(s) to deploy the Bot. Technically the Bot can be deployed in more than 50+ channels today, however we will look at the most high density channels which are used by over 90% of the target audience.
Here is a list of the most popular Conversational Channels for Customer Engagement:
👉 Facebook Messenger
👉 other mediums like Twitter/Google Business Messaging etc
Let’s now take a look at the unique characteristics of each channel and what is needed to activate them as customer engagement channels.
Today WhatsApp is the most popular business conversation channel and is growing rapidly. To get your Bot running on WhatsApp requires the following activities:
👉 Getting a KYC for your registered company with Facebook
👉 Registering for a WhatsApp Business Account(WABA) with an authorised WhatsApp Business Service Partner
👉 Connecting your Conversational Platform with WhatsApp Business Service Partner APIs
👉 Implementing WhatsApp’s strict Opt-In protocols
Creating WhatsApp Business Account
A WhatsApp business account setup is a two step process.
Step 1: Getting a Verified Business Manager Page with Facebook. Please note that Facebook is the parent organisation of WhatsApp.
Step 2: Using the Verified Business Manager Page details to register for the WhatsApp Business Account which gives your brand an end point access to use WhatsApp Business APIs.
Pre-Requisites and Steps for Verified Business Manager Page with Facebook
👉 Working website + Personal Facebook Account login + Facebook Page of Your Business + Work Email
👉 Use the personal Facebook Account login to create a Facebook Business Manager Page.
👉 Provide Business Details like legal name, contact number, work email, address.
👉 Provide supporting documents for Legal Business and Address Proof.
👉 For Business: Business License / Company Registration Document (any one)
👉 For Address: Tax Certificate, Water/Electricity/Utiility Bill/Bank Documents (any one)
👉 Wait for approval.
Pre-Requisites and Steps For Registering WhatsApp Business Account (WABA)
👉 Choose an authorised WhatsApp Business Partner.
👉 Submit the Verified Facebook Business Manager Page ID, along with with screenshot.
👉 Provide a new Cell Number to be tagged to the WABA.
👉 Provide a Display Name and Display Picture.
👉 Mention the purpose of use. (Bot, Live Chat, Outbound Messaging)
👉 Mention the industry segment, category where it will be used.
👉 Mention the uses cases. (Please note WhatsApp has strict restrictions on certain industries and use cases.)
Role and Need of WhatsApp Business Partner
WhatsApp does not provide access to its Business API services directly to customers. It works through a network of partners. Customers can directly approach these partners or use the services of the Conversational Customer Engagement service provider, most of whom have partnerships with official WhatsApp Partners.
WhatsApp partners provide the technology layer for receiving and sending messages and connect to the Conversational Platforms through APIs. All the other functionalities like Bots, Analytics, Messaging etc are managed over the Conversational Platform.
WhatsApp also has a structured messaging scheme where by it classifies all messages sent/received through its API as Notification Messages or Session Messages. WhatsApp Partners provide these messaging delivery services at a cost.
What is Notification Message?:
Any message sent by Bot/Conversational Platform to user without the user initiating a dialogue with Bot. These include shipping updates, OTPs, transactional confirmation messages, alerts, etc. The Cost of these Notifications Messages, per country, is defined by WhatsApp and updated daily on their website. The base cost of all Notification Messages is the same for all WhatsApp Partners. However some partner may add a small markup and offer it to customers.
What is Session Message?
Session messages are any messages which are user initiated and the Bot responds. These include all general messages which are 2-way. WhatsApp does not charge any fees for Session Messages. However all WhatsApp Partners charge a fee. This could be Per-Message Fee or Per-User Fee(with unlimited to and fro messages).
WhatsApp Business Partners also provide the messaging gateway server which handles all the high volume WhatsApp messaging traffic.
Going Live On WhatsApp
Once the WhatsApp Business Account is approved, it can be connected with any Conversational Platform to take the Bot live to the target audience.
Here is how it looks once once it is Live.
Distributing the WhatsApp Channel
Once the WhatsApp channel is live you can adopt various ways to distribute it and make it easy for your customers to access it. Some of the common ways of distribution are:
👉 QR Code
👉 WhatsApp Chat Widget URL
👉 Missed Calls
👉 Save Number and Access Service
👉 WhatsApp Buttons
👉 Share Contact
Using your WhatsApp Business Account Number and your preferred activation Keyword (Hi, Hello, Start, Menu, etc) you can create ad distribute QR codes in leaflets, e-brochures, emails, newspaper, text messages, etc.
Here is how a sample QR Code looks like for a Brand(NHANCE NOW) + Cell Number (+91 91084 99376) +”Hi” as Activation Keyword.
WhatsApp Chat Widget URL
WhatsApp provides a url service which also helps to generate links to distribute Bots. Called WhatsApp short link it captures necessary information and provides an easily shareable short url link.
Here is how a sample short link looks like for a Brand(NHANCE NOW) + Cell Number (+91 91084 99376) +”Hi” as Activation Keyword:
Clicking on this link takes the user to the Bot on WhatsApp and pre-populates the input field with the activation keyword.
Missed call service is another easy way to promote distribution of Bots. A combination of Missed call service and WhatsApp short link is used tactically to effectively distribute Bots.
A Number for Users to give a Missed Call –> Call Disconnects after 2 rings –> Within 5 seconds user gets a SMS with the WhatsApp Short Link –> User clicks on Short Link to access Bot with Activation Keyword
Save Number and Access Service
This is the easiest way to access Bots. Company promotes the Bot number and the Activation Keyword. Asks users to save the number, access the account over WhatsApp and type the Activation Keyword to get started.
A sample promotional message could be: “Please save +91 91084 99376 in your WhatsApp Contact list and Type “Hi” to access our services.”
Many 3rd party developers publish interfaces to create simple WhatsApp Buttons which use simple programming and WhatsApp short links to create WhatsApp Button widgets. These widgets can be placed on Websites, Social Media Pages, Landing Pages, Blogs, etc to help users easily connect to the Bot over WhatsApp. Clicking on these button opens the Bot on WhatsApp (Desktop, Browser, Mobile App)
Here is how it looks on an Website:
Contact sharing is a process whereby users are incentivised or “coaxed” into promoting the Bot to friends, family, colleagues. It is similar to sharing a regular contact over WhatsApp.
Here is how it looks like:
Facebook Messenger is another highly popular Bot channel. Facebook Messenger has the added advantage of not just working in its own independant channel but it can also be used, via its chat widget, as Bot delivery channel on Websites and Apps. Getting your Bot live on Facebook Messenger involves the following steps:
👉 Creating a Facebook Business Page. Fill it with company information.
👉 Create an APP in Facebook Developer Account.
👉 Connect Facebook Business Page and APP.
👉 Connect the Conversational Platform to both using Messenger API.
👉 Publish Facebook Page, which contains the Bot.
Facebook Business Page
A Facebook page is the primary interface through which your customers will access the Bot over Facebook Messenger. It contains all the details of your brand, logo, display name etc.
Facebook offer a Business Manager Page which is where you can manage a Facebook Business Page.
Here is how a Facebook page looks like on web interface:
Facebook Page offers plenty of configurations. Some of the key settings include:
👉 Country restrictions – page would be visible in selected countries.
👉 Age Restrictions – page is shown to users above a certain age.
👉 Colour theme, language, profanity filters amongst others.
Once your Facebook page is setup, its time to create the Facebook Developer APP which will take additional permissions from Facebook for enabling the Bot.
Facebook has separate developer segment where these APPs are created. Here is a sample view of one such APP:
A number of permissions are required by Facebook to activate an APP for production mode.
Information requested by Facebook range from location, gender of target audience, timezone, step by step guide to use the Bot, sample Bot flow, explainer videos amongst others.
👉 Page ID
👉 Page Access Token
👉 App Secret Key
Once connected here is how the Bot looks when the Page is accessed via Facebook Messenger.
Please note that this page has been create for the Europe and Middle East (EME). With the country filters in place you can hide it from users in other regions. This helps is delivering highly localised and personalised content and services.
Distributing Facebook Channel
Once the Facebook channel is live you can adopt various ways to distribute it and make it easy for your customers to access it. The methods are similar to WhatsApp channel with some modifications.
👉 QR Code
👉 Messenger URL
👉 Missed Calls
👉 Page LookUp
👉 Facebook Messenger Buttons
Using an unique page url you can create and distribute QR codes in leaflets, e-brochures, emails, newspaper, text messages, etc.
Here is how a sample QR Code looks like for a Brand(NHANCE NOW) + Page (NHANCE NOW)
Facebook provides an unique url for each page which also helps to generate links to distribute Bots. Called Facebook url it provides an easily shareable short link.
Here is how a sample short link looks like for a Brand(NHANCE NOW) + Page(NHANCE NOW)
Clicking on this link takes the user to the Bot on Facebook Messenger to start conversations.
Missed call service is another easy way to promote distribution of Bots. A combination of Missed call service and Facebook short link is used tactically to distribute Bots.
A Number for Users to give a Missed Call –> Call Disconnects after 2 rings –> Within 5 seconds user gets a SMS with the Facebook Short Link Url –> User clicks on Short Link to access Bot on Facebook Messenger
Page LookUp is a simple way of directly accessing the Bot. Brands promote the page name and Bot activation keyword. Users lookup the brand’s page in Facebook Messenger search bar and type in the Keyword to start conversations.
Many 3rd party developers publish interfaces to create simple Facebook Messenger Buttons which use simple programming and Facebook short links to create Facebook Messenger Button widgets. These widgets can be placed on Websites, Social Media Pages, Landing Pages, Blogs, etc to help users easily connect to the Bot. Clicking on these button opens the Bot on Facebook Messenger (Desktop, Browser, Mobile App).
Here is how it looks on an Website:
In recent times Instagram has emerged as one the most popular and fastest growing channels for commerce and simple consumer-business messaging. With a recent update to the Messenger API, you can now use the same Conversational Platform and Bots to not just connect with WhatsApp and Facebook but also add Instagram to the mix.
The now powerful Messenger API allows the existing conversational workflows to enable commerce, customer service and lead generation communication to be extended to support customers, reaching out via Instagram.
Here is how a Conversation flow works in Instagram:
Although websites are not the most used channels for customer engagement in comparison to apps, they still offer a complimentary channel which you should not ignore. The Web channel usually has less functionalities than the messaging and social channels.
We will look at two ways in which Conversational Bots can be deployed on Websites:
A. Facebook Web Chat Widget
B. Standalone Web Widget
Facebook Web Chat Widget
This is one of the easiest was to activate a Bot and Conversation Workflow on your website. It simply replicates your Bot designed for the Facebook channel and puts a copy of it on the Website. Any changes made to the primary Bot workflow also reflects on the Web Widget. It is for this simplicity that the Facebook Web Chat widget has gained immense popularity.
Deploying the Facebook Web Chat Bot takes less than 10 minutes. A configuration screen and code is provided on the primary Facebook Page Settings section. You have to simply make the customisation, like initial Welcome message, colour theme, copy the code and paste the code it on all the webpages that you want this web chat widget to appear. A developer support is only needed for copying the code to the individual HTML pages of the website.
Here is how a Facebook Web Chat Widget looks on a website:
Facebook’s Webchat Widget also allows to extend the capabilities of Bots and Conversations to any iOS and Android APP. Once again the big advantage is the reusable widget code. Just like in webpages, the code can be attached to any menu, button or tab in a mobile app. It pulls up the same Conversational interface that we get in Facebook Messenger and Facebook Web Chat Widget. Changes made to the primary workflow also shows up instantly in the App interface of the Bot.
Here is how its looks in an APP interface:
3. Power Boosters – Expand Platform Capabilities
Having got the Bot organised, its time to enrich the capabilities of the Conversational Offerings. Different Conversational Platforms provide different capabilities and you can choose one which best meets your intended goals.
Here are some of the functionalities that can add significant power to your offerings, make the Bots more powerful and increase overall customer experience.
👉 Outbound Messaging
👉 Service Ticketing
👉 Live Chat
👉 Payment Gateways
👉 3rd Party Integrations
Let us take a closer look at them.
Outbound lifecycle messaging is an important part of most Conversational Platforms. With outbound messaging one can send simple time based notifications to more complex event/action based alerts and notifications. This functionality can also be used for sending offers, feedback collection requests, commerce shipping updates, service ticket updates, calendar reminders, OTPs and various kinds of other useful information.
Outbound Messaging modules work in two ways:
A. Standalone – Typically programmed to send messages based on time or a scheduled event.
B. Via APIs – Connecting to 3rd party applications, picking up information and relaying it through your Conversational Platform to the end users.
There are 4 types of rules commonly used for outbound messaging in Conversational Platforms.
- One Time – Messages are sent immediately or at a certain time.
- Recurring – Same message is sent multiple times at pre-defined time intervals.
- Triggered – Messages are sent when a certain condition is met.
- Consumer Segmentation – Sent to all consumers or rule based segments.
Here is series of screen mockups to give you a sense of these rules of outbound messaging, with each having a section for user segmentation.
One Time Messages
You can also see the total number of messages sent, delivered, viewed and clicked (if they contain action items like urls, quick reply buttons).
Live Chat is the default fall back module of Conversational Systems. You have the power of involving humans to resolve complex queries or issues which the Bot, AI or NLP systems fail to handle. A quick handover to humans helps not just save the day but also results in higher customer satisfaction.
Live Chat modules are typically used for:
👉 Fall Back against inadequate Bot / Ai response.
👉 Human based personalised advisory services.
👉 Handle issues not covered in Conversational Systems
As Live Chat modules are immensely complex systems very few Conversational Platform have a full fledged Live Chat Module. The best case scenario is an integration with an established 3rd party Live Chat tool, like LIVECHAT or CRISP CHAT.
There are 3 ways in which a Human Agent handoff is done:
A. Through a Menu – A menu to connect directly with an agent is provided which users have to choose.
B. Through Keywords – like “Talk to Agent“. Users can type this phrase at any stage to connect with agents.
C. Negative Bot Response – Workflows can be configured to automatically re-direct users to agents if there is more than 2 consecutive negative responses.
Here are some of the key functionalities of Live Chat tool. For the purpose of reference we shall use LIVE CHAT, one of the most used solutions.
Create Team and Roles
Add agents, supervisors, admins based on the expected incoming chat volumes.
You can assign shifts to each agent for each day, have weekly offs, plan lesser scheduled hours new joins.
Incoming Chat Routing
Chats can be automatically routed to agents based on rules or selected manually from the queue.
Automatic chat routing – Each new chat goes to the first available agent. When all agents have one chat each, the next round is distributed. Manual chat routing – All agents receive a notification when a visitor wants to start a chat. Agents then can select the unassigned chat from the queue.
Reduce Repetition. Your customers can pick up a chat where they last left off or reread a conversation anytime they need to. It saves time because they don’t need to ask again about something you’ve already answered.
Measure the performance of each agents and see who are the winners and who needs improvement.
Apart from these you can also manage pre-defined responses, share files, allow transfer of chats, view centralised chat dashboard, chat in local languages, set advanced routing and chat rules for agents.
Service ticket modules are a key component of the customer journey. These platforms assist the activities of help desk teams. Tickets are managed with shared team inbox, which support team members can access through an individual portal with a unique login. Customer requests can be tagged, modified, assigned a relevant status, and closed without hoping between multiple systems. Service Ticketing systems are used when the volume of inbound queries are large. These are used along side Live Chat tools by many organisations.
Service Ticket Modules are typically used to:
👉 Collect and process customer issues, requests and suggestions.
👉 Maintain customer support history for better to develop best practices in future.
👉 A centralised repository of customer tickets so that nothing slips through any crack.
👉 Overall better management and faster response to service ticket.
If you have a high volume of service tickets implementing a robust service ticket management tool is vital. Some of the lead players in this field are Freshdesk, Kayako, Helpscout and more. All of these systems can be easily integrated into Conversational Platforms. This allow you to collect service requests directly from end users using social and messaging apps and respond back to them via the same channel.
Here are some of the key functionalities of Service Ticket tools. For the purpose of reference we shall use Kayako, one of the most used solutions.
Managing customer conversations becomes easy with the shared inbox tool. With custom views, tags, and conversation assignments, your customer service team is able to stay on top of open issues and automatically assign conversations to the best people.
Customer Interaction History
Visual timeline provide lists every interaction your customer has had with your business. No need to ask additional questions. The information your team needs to respond quickly and personally is right at their fingertips.
Great customer service happens when teams work together. Collaborators can come from any internal team, such as tech, success, sales, shipping, billing, etc to help resolve complex customer service challenges. Your agents remain the main point of contact for customers while collaborative pooling of inputs happens if resolutions are ongoing.
You can get deep insights on how the support team is performing and how it is impacting customer satisfaction rates and generate custom reports. Dashboard and charts help you keep an eye on your team, see where they need help , check on key metrics and measure efficiency of every conversation.
Beyond this you can also group team members into specific segments, build ticket assignment rules, share pre-defined responses, files and respond in multiple languages.
Payment systems help you enable Bots to collect payments without any manual intervention. Conversational platforms are widely used for commerce and the availability of payment interfaces add significant value for digital payment collection via credit card, debit card, internet banking, wallets, mobile money, UPI transfers.
Payment Gateways are typically used for:
👉 Purchases / Ecommerce
👉 Renewals and Subscription fees
👉 Bill Payments
Some of the most popular payment gateways that you can use are Razorpay, Stripe, PayU, Worldpay, Paypal, DPO, Pesapal, Payfort and more.
Here is a list of features which typically a payment gateway provides. For our reference we are using Razorpay.
👉 Create and Send Payment Links
👉 Maintain list of Invoices, payments made, failed payments.
👉 Triggered or action based payment links via integrations
👉 Integration APIs
👉 Detailed reports of collection, refunds, reconciliations.
Analytics and Insights
In Digital First environment, with little or no human involvement, getting automated statistical data and deep insights is like gold dust. By centralising data collation from various modules it is not just possible to generate analytics for individual customers but also get a holistic view of how the Automated Conversational System is performing.
Data and Analytics can be generated in the form of Excel/CSV Reports, Charts and numerical mini dashboards. It is also possible to generate lists of keywords used, most used intents and utterances from customers and trending topics.
Most conversational systems have some form of Analytics Generation. This can be further augemented by using 3rd party Analytics tools like Botanalytics and many more.
The Analytics modules typically offers these 5 types of output
👉 Conversation Metric
👉 User Analytics
👉 Bot Analytics
👉 Live Charts
Here is an explanation of the most commonly used terms in Conversational Analytics.
Total Users: Total number of users in the system. Unique user is a unique phone number.
Daily Active Users – DAU: Total number of users (phone numbers) who have pinged the system at least once a day.
Monthly Active Users – MAU: Total number of users (phone numbers) who have pinged the system at least once in a calendar month.
Engaged Active users per month – EAU: Total number of users (phone numbers) who have messaged the system more than 15 times a month. The messages could be in one session or multiple sessions over a period of a calendar month.
Daily New Users – DNU: Total number of new users joining per day. A new phone number added to the system.
Monthly New Users – MNU: Total number of new users joining per month. A new phone number added to the system.
Definition of Session in the Context of Analytics Generation:
A session is defined from the time a Bot or a User starts an interaction and ends when there is no activity between the bot and user (no exchanges of message) for a period of 30mins.There can be multiple sessions between a user and a Bot in a day.
If a session is in progress from 10.00am to 10.15am and then from 10.15am to 10.45am and there is no further message exchange then the session ends at 10.45am. Please note that there would scenarios where a question has been asked by the user or a Bot at 10.15am but the response comes at 12.00pm and subsequently the discussion carries on from 12.00pm to 12.10pm, then in this case there are TWO session: 10.00am to 10.15am and 12.00pm to 12.10pm (even though it could be the same topic)
Session Summary: Total number of sessions from the start date of the account till present.
Total Outgoing Messages: Total count of all outgoing messages (session and notification) from the start date of the account till present. If the same message has been seen on both Whatsapp and FB (or nay other channel) it is considered as one SENT MESSAGE. (For example a Notification message has been sent and it has landed in both WhatsApp and FB, then it is treated as one message).
Total Incoming Messages: Total count of all incoming messages (session and notification) from the start date of the account till present. Every message hitting the server is considered as one message.
Average Daily Session per User: The average number of sessions over the past 24 hours
24hrs: 12.00AM – 11.59PM UTC.
Formula: Total number of Sessions in the Past 24 hrs / Total number of Daily Active Users in the past 24hrs
Average Session Time Per User
The average time each daily user spent in a session.
24hrs: 12.00AM – 11.59PM UTC
Formula: Total duration of Sessions in the Past 24 hrs / Total number of Sessions in the past 24hrs
Average User Messages Per Session
The average count of incoming messages per session.
24hrs: 12.00AM – 11.59PM UTC
Formula: Total number of messages sent by users (Incoming) in the Past 24 hrs / Total number of Sessions in the past 24hrs
Monthly Session: Number of Session Per Month
Daily Average Session Time
The average time each daily user spent in a session.
Formula: Total duration of Sessions in the Past 24 hrs / Total number of Sessions in the past 24hrs
Monthly Average Session Time
The average time per month each daily user spent in a session.
Formula For each month: Total duration of Sessions in the month / Total number of Sessions in the month
Daily Average Messages Per Session
The average count of incoming messages per session.
Formula for each day: Total number of messages sent by users (Incoming) in the Past 24 hrs / Total number of Sessions in the past 24hrs
Monthly Average Messages Per Session
The average count of incoming messages per session for the month.
Formula For each month: Total number of messages sent by users (Incoming) in the Month/ Total number of Sessions in the month.
Trending Keywords / Intents: The most used words, topics or sentences discovered from customer inputs.
In addition to these metrics advanced rules can be applied to build custom analytics and reports.
Here is a glimpse of various charts and graphs which help in visualising key data and statistics.
APIs & WebHooks
Conversational Platform can be made more dynamic by using APIs and Webhooks. This allows real time exchange of data (to/fro) between various interconnected systems, enhancing the user experience by several notches.
An API can be used from a Conversational system to communicate with an external web application. Through that communication, the API can List, Create, Edit or Delete items. The API needs to be given specific instructions on what action to perform and what parameters to touch..
Webhooks, on the other hand, are automated calls from an external web application to the Conversational system. Those calls are triggered when a specific event happens on example.com. For example, if a new user signs up on example.com, the automated call may be configured to ask the Conversational system to send out a welcome message on WhatsApp.
Here is an example of how Stripe uses both API and Webhook calls for different functions:
How do APIs and Webhooks help to enrich your Conversational Offering?
Lets look at this with a real life example, using the function of Stripe as an example, beautifully explained by Vero.
To use a real world analogy, APIs would be likened to you repeatedly calling a retailer to ask if they’ve stocked up on a brand of shoes you like. Webhooks, would then be like asking the retailer to call you whenever they have the shoes in stock, which frees up time on both sides.
Webhooks are less resource-intensive because they save you time on constantly polling (checking) for new data.
Here’s a diagram that shows how APIs and webhooks handle new data:
Even though only a single event occurred at 1 pm, you’d have to constantly poll your provider for new data with an API. With webhooks, you just get notified once at 1pm. It’s vastly more productive.
APIs are still popular for a number of reasons:
- Not every application supports webhook integrations.
- Sometimes you only want to know about the end result, rather than every event (i.e. every permutation) that’s changed an object.
- Webhooks can only notify you about an event, so if you want to make a change based on new information, you’ll need an API.
- A webhook payload may not contain all the data you need about an event.
So APIs are still really useful, which is why a lot of applications support both APIs and webhooks.
Lets Look at a few more examples of Integrating 3rd Party Tools using Webhooks and API to enhance Automation over Conversational Systems
Using Shopify APIs to Automate Post Purchase Commerce Interactions.
Shopify is the most used e-commerce application with over 1.4mn online stores in 170+ countries. Along with its standalone application Shopify can be integrated with multiple business applications, including Conversational Platforms. It offers a suite of APIs that allow developers to extend the platform’s built-in features. These APIs allow partners to read and write merchant data, interoperate with other systems and platforms, and add new capabilities for their end users.
Shopify Store Front API
Storefront API gives you full creative control to add Shopify buying experiences anywhere your customers are, including websites, apps and conversation channels like WhatsApp and Facebook. Storefront API is useful when Shopify merchants have custom requirements not met by existing channels. Most conversational platforms do not have a full fledged commerce backend and this is where they can leverage the capabilities of Shopify backend and pull / push data to their Conversational End Points. This API can access product information, create customer orders, complete purchases, manage orders, cancellation and check up historical data.
- Sell products from a Shopify store through a non-Shopify channel like WhatsApp / Facebook Bot.
- Build standalone functionalities like Cart management, product catalogues, order management, FAQ lists using Shopify as the underlying commerce platform.
Here is a example of a Conversational Interface pulling / pushing data entirely from Shopify to complete e-commerce transactions using Product Catalogue, Cart Manager, Order Manager:
In many cases a combination of Shopify, In-Built features of the Conversational Platform and other 3rd party tools can be used to craft a complete end to end Commerce and Customer Experience for users.
This diagram illustrates how a compelling experience can be crafted for end users on their personal messaging or social app:
There are many other ways in which APIs and Webhooks can help create interconnected solutions to solve specific business needs. We will be covering more of it in a future article.
4. Security – Keeping Everything Safe.
Including Conversational Platforms as part of your Digital Business Strategy needs to come with adequate protection of data, privacy and overall security. With regular exchange of sensitive and personal information it is vital that strong security policies are in place. There are various forms of securing these sensitive interactions and most Conversational Platform use a mix of them to keep everything safe.
Let’s take a look at some of them and what to look out for when assessing the security features that come with a Conversational Platform Service Provider:
👉 General Security
👉 Data Privacy & Information Security
👉 Data Encryption
👉 Audit Trails
Usually all employees sign a strict Non-Disclosure and Data Access Policy that binds them to the terms of standard data confidentiality policies. Access rights vary on employment role. Office premises are typically secured via access control system which logs visitor entries and exits.
All networks and access points to all active networks are tightly monitored. Collection of logs produced by networking devices such as firewalls, DNS servers, DHCP servers and routers are maintained in a central place. The network logs are retained for the security appliance (firewall), wireless access points and switches. Intrusion detection systems are enabled to identify and alert potential security threats.
Application and Data Center Security
Most platform use the services of established cloud server providers like Amazon Web Services, Microsoft Azure or Google Cloud Platform to host their application. These companies employs a robust physical security program with multiple certifications, including an SSAE 16 certification and many other globally recognised protocols for data center security
Backup and Disaster Recovery
Platforms usually maintain geographically separate environments to ensure data availability and uptime. In the unlikely event of simultaneous failure of both environments, a third manual backup in a localised or tertiary environment is kept to minimise system recovery time.
Data Privacy and Information Security
There are many standards being followed in different geographies. While they are all robust and provide significant protection to customer data and information we will take a look at two of the widely followed ones globally, GDPR and ISO 27001. Please note that these standards will give you a general sense of global best practices but your Conversational Platform Provider may be using other protocols which may be equally robust.
GDPR can be considered as the world’s strongest set of data protection rules, which enhance how people can access information about them and places limits on what organisations can do with personal data.
If a platform is GDPR compliant they would be adhering to these key protocols:
ISO 27001 was developed to help organizations, of any size or any industry, to protect their information in a systematic and cost-effective way, through the adoption of an Information Security Management System (ISMS).
Why is ISO 27001 important?
Not only does the standard provide companies with the necessary know-how for protecting their most valuable information, but a company can also get certified against ISO 27001 and, in this way, prove to its customers and partners that it safeguards their data.
Conversational Solutions Providers who are ISO 27001 certified usually follow these protocols:
Data Encryption & Integrity
When data travels between multiple systems, it needs to be highly encrypted to ensure that there isn’t any “leakage” and it doesn’t fall in the hands of hackers. (e.g. Data flowing from a Conversational Platform to a CRM, fetching details from CRM and passing it further to a messaging platform like WhatsApp.)
Some of the protocols you need to watch out for are:
👉 HMAC – (Hash-based Message Authentication Code) is a type of a message authentication code (MAC) that is acquired by executing a cryptographic hash function on the data (that is) to be authenticated and a secret shared key
👉 AES – The Advanced Encryption Standard (AES) is a fast and secure form of encryption that keeps prying eyes away from our data. We see it in messaging apps like WhatsApp.
👉 IP Whitelisting – It is a security feature often used for limiting and controlling access only to trusted users. IP whitelisting allows lists of trusted IP addresses or IP ranges from which your users can access your domains.
The audit log is the report which records accessed information and resources like target URLs, source lists, tags, and user account details. It helps you track all admin and customer activities right from login to all kinds of navigation, offering information about who did what and when.
Here is how a typical audit trail in a Conversational Platform looks like:
5. Winning Conversational Templates
Crafting winning tech stacks, with so many options to choose from, is as much an Art, as Tech. To give you glimpse of how technology tools are used to create value based Conversational Customer Engagement Stacks, we will look at some real life cases in this section.
In the section we will look at what tools are used by some of the customers of our company, NHANCE NOW, segmented by use cases. It’s an unfiltered behind-the-scenes look which you can dip into and craft your winning Conversational Tech Stack.
Here is an example of a typical full-fledged Conversational Customer Engagement Stack. Companies mix and match tools to meet their specific business goals in their Digital First Transformation journey.
Now let’s take a look at some custom conversational stacks. This compilation is a result of data collection from over 300+ companies that we examined during our research.
💎 Commerce Stack
💎 Customer Service Stack
💎 Lead Generation Stack
💎 KYC Stack
💎 HR Services Stack
💎 Financial Services Stack
💎 Insurance Stack
Use Case: Sell Products and Services
Industries: FMCG, Food, Fitness, Personal Electronics, Clothing, Grocery, Automobiles, Office Stationaries.
Turnaround Time: 12 – 17 days
Customer Service Stack
Use Case: Post Purchase Customer Service
Industries: FMCG, eCommerce, ISPs, Financial Services, D2C, Educational Institutions, Real Estate, Electronics.
Turnaround Time: 10 – 12 days
Lead Generation Stack
Use Case: Automate Lead Generation on Websites, Social Medial, Email and Via Missed Call Services.
Industries: ISPs, Financial Services, Educational Institutions, Real Estate, Events.
Turnaround Time: 5 – 7 days
Use Case: Digital, Presence-less KYC
Industries: ISPs, Financial Services, Telecom, Loans.
Turnaround Time: 5 – 7 days
HR Services Stack
Use Case: Hiring Assistance & Employee Engagement.
Turnaround Time: 10 – 15 days
Financial Services Stack
Use Case: Financial Transaction, Support, Account Services.
Industries: Payments, Digital Banking, Loans, Neo Banking, Investments.
Turnaround Time: 15 – 30 days
Use Case: Selling Products, Claims Processing, Salvage Auction, Digital Assistant For Agents
Industries: General Insurance, Vehicle Insurance, Health Insurance
Turnaround Time: 15 – 30 days
Digital Transformation is an ongoing activity. With new technologies coming up all the time in the conversational automation space you will continue getting more power in your hands. And learn to use messaging tools in new and exciting ways to race ahead of competition with a Digital First approach.
If you even follow just 30% of the technology and strategies covered here you can develop a viable conversational customer engagement strategy, and can go a long way in staying relevant to your digital era customers.
Look you will come across other view points that talk of simply focussing on growing sales and improving product/customer service and ignore this pandemic created buzzword “Digital First” and the need to get a complete digital makeover. If it aligns with your strategy, sure, go ahead, maintain your current status. But you’ll often find that it’s better to go with the market trends. And today it’s loudly and clearly a Digital First market. And we all have to find a way to fit it. By still focussing on sales and customer service but with Digital Only approach.
I hope this guide has been helpful. I would love to hear your thoughts. What more could have i added to this article? What should i write about i future articles? And i keenly await to hear your digital first stories, points of view and challenges.
Nishith Patnaik, co-founder NHANCE NOW. You can connect / DM me on Twitter, anytime.
And do stay tuned for more helpful guides and handbooks from the Fast Forward blog by signing up for our newsletter below.👇