Off-The-Shelf vs Custom Chatbot Development: What's Best?

31 May · 8 min read

- Hi, how can I help you?

- I want to build a chatbot

- Do you want to build an off-the-shelf chatbot or a custom chatbot?

- I’m not sure. I need some advice.

- Ok, this article will help you choose.

Today you can meet online chatbots almost everywhere. No wonder you want to learn how to create a chatbot. This technology can dramatically improve the customer experience.

Chatbots are growing in popularity, and the market offers more ready-to-use software all the time. But some businesses still prefer to create a custom chatbot. With all the massive variety out there, it is hard to tell what is the best option. We will help you understand the different types of chatbots and approaches to their development. Let’s begin by discussing what exactly a chatbot is, and why you might want one.

Reasons for creating a chatbot

A chatbot is the software programmed to converse in text and/or audio formats. Some chatbots use scripts to answer queries by searching for keywords or keyphrases. Others use machine learning (e.g. Natural Language Processing) to analyze previously received data, provide unscripted answers, and improve their own service. In general, chatbots are powerful tools that are able to significantly boost the quality of customer experience.

Here are some advantages of chatbots:

Customers get instant replies to their questions when a human manager is not available.
Customer support is available 24/7 to process requests anytime.
Bots reduce employee stress by handling repetitive tasks.
Interaction with a chatbot simplifies the data entering process for visitors, which improves their experience and conversion rates.
Chatbots enhance the visitor's experience by giving an introduction, typical information, and fluid navigation.

There are several types of chatbots:

Script-based - static in their abilities. The bot moves through a list of predefined questions and answers during the conversation. Due to this rigidity, script-based chatbots can deal only with simple tasks. It can lead a customer to a specific, concrete goal, e.g. to complete a purchase or to leave contact information.
AI-based - able to converse almost like a human thanks to natural language recognition. The bot analyzes the context of the message and then comes up with the appropriate response on its own. To achieve this, such systems process and learn from previous conversations.
Hybrid - a combination of scripted and AI-powered chatbots. Simple questions receive predefined responses, while complex requests are processed by AI. If a misunderstanding arises, customers are offered to connect with a human agent or to email the company.

How and why to develop off-the-shelf chatbots

Off-the-shelf chatbots are solutions that you can develop without any programming skills. As a rule, this software has a drag-and-drop interface so anyone without tech experience can install and set it up. This option offers basic functionality and is able to answer common questions.

What is particular about off-the-shelf chatbots:

They are less expensive. Most canned solutions have affordable prices or even free plans. Additionally, there are trial periods to test if the software suits your business.
You don’t need to be a programmer to create them. You can customize ready templates or effortlessly build a chatbot with a drag-and-drop interface.
They have pre-built design templates. There are a lot of ready designs that need minimal alterations.

They possess only basic functionality. Most solutions offer elementary features which can be a problem if your business needs something more unique.
Hard to integrate. It can be difficult to integrate an off-the-shelf chatbot with your CRM or ERP if you don’t have technical knowledge.
There is low data protection. Canned solutions can be unreliable in terms of data storing and access to these data. Moreover, it can be complicated to transfer data to a new platform.
They have limited customization. There are templates that you can use, but they allow limited modification.

Here are examples of the most popular no-code platforms:

Chatfuel chatbot is perfect for Facebook Messenger. It increases sales, personalizes marketing, and automates customer support.
Tars helps to optimize the conversion funnel and can be integrated with WhatsApp.
Botsify is easy to set up and integrates with Messenger, Slack, or SMS as well as with your own website.
MobileMonkey allows the creation of chatbots for Messenger, Instagram, and SMS. It provides efficient interaction with the audience and increases the conversion rate.
HubSpot’s chatbot has a powerful tool for turning visitors into leads.

These chatbots can help users collect information, navigate through the website, answer FAQs, or connect with a human operator. But don’t expect them to provide anything more complicated than that. The functionality of canned solutions is limited and will not suit every business.

If you are looking for something more complex and flexible then you need a custom solution.

How and why to develop custom chatbots

Custom chatbots are built from the ground up using programming languages. This approach allows you to create software with unique features and designs to meet all your needs. You have more freedom in terms of functionality and integrations.

What is specific for custom chatbots:

They are flexible and scalable. You are in total control from the beginning so you are the one who defines what it will look like and how it will function. You also can change and improve them anytime.
They are tailored to your needs. Your team will develop features that are necessary for your business and leave the rest behind.
You can have the ultimate level of security and total control of the data and code. Chatbot developers will provide you with data protection by using the best practices.

Custom chatbots can take a lot of time and money to develop. It requires more resources to create everything from scratch. The more time your team spends, the more it costs. But you don’t need to worry about the monthly fee.
It requires a team of professional developers to build software. You need to know how to code to build a chatbot from zero or hire a team to do it for you.

Of course, custom development is time-consuming and it takes months to build a whole new software, not to mention the expense. But some solutions can speed up the process. For example, you can use chatbot building frameworks, databases, or NLP libraries.

Many of these solutions are open-source and help to reduce development time and cost. They can provide system architecture as well as speech recognition technologies (ASR and NLU) and AI capabilities. Such software easily integrates with services like telephony, data storage, or user verification. It also has scalable hosting with instant data recovery features.

But the flexibility of these solutions is limited because you can’t change the pre-built architecture or fix bugs there. Therefore, sometimes it is easier to build a chatbot from scratch than to modify a pre-made platform’s capabilities to meet customer demands.

Here are the best practices:

Open-source platforms - offer databases and libraries with a large pool of collected data. AI-based chatbots will use this knowledge for training to provide relevant answers later.
Professional chatbot platforms - speed up development time, but require programming skills. Among these platforms are Microsoft Azure, AWS Lex, IBM Watson, and Google Dialogflow.

Our experience: one of our clients at Greenice asked for help developing a medical chatbot for a hospital. We built a bot that allowed patients to provide their symptoms and to schedule doctor appointments if necessary. This software allows users to choose answers from available options, enter text, or contact a human agent. The bot is AI-powered, so it analyzes previously given information to learn, make predictions, and ask relevant questions. Such a solution became possible only by starting from scratch.

Ready-made and custom chatbots comparison

Both canned and custom options have their pros and cons. But what is most important, not all of them are equally good for you. To help you make an informed decision we will compare off-the-shelf and custom chatbots. Let’s go over the features, development steps, price, and other important aspects.

Functionality

As you know, canned solutions offer predetermined features and limited customization. If you need a bot with standard functionality for simple assistance, off-the-shelf software is what you are looking for. There are a lot of options to choose from.

But premade bots suit only limited types of businesses, or even a small niche. Such bots can’t adapt to your business - you are obliged to adapt your company to the bot. Modern businesses that offer a wide variety of services need a complicated bot solution. If you haven’t found a bot that meets all your needs you should consider custom development. It’s too expensive, too complicated, and possibly impossible to upgrade an existing bot. What’s more, the maintenance of a bot that has undergone major revision is a nightmare.

Development from scratch leaves chatbot abilities up to your imagination. You start with a blank sheet and tailor the software to your vision. It might be the only right solution for some businesses with unique needs.

Development path

Considering that off-the-shelf and custom chatbots are different in their functionality, the steps for their creation also differ. In both cases, development starts with research and planning. During this so-called Discovery Phase, a company discusses chatbot goals and features and pictures the customer’s journey. However, further steps are not the same. Here is a chart featuring Off-the-shelf and Custom bots telling you how to make a chatbot:

Price

Off-the-shelf solutions charge a platform fee which depends on the plan you choose. Some platforms are free so you will definitely find a perfect match in matters of price.

Chatfuel: starts from $0 and goes to $199 per month for Pro subscription
Tars: starts from $99 per month to an individually estimated price for a custom plan
Botsify: starts from $49 per month and offers custom pricing plans
MobileMonkey: start from $0 and goes to $299 per month for teams
HubSpot’s chatbot: starts from $0 and goes to $3200 per month for the Enterprise version.

The price for custom chatbot development depends on the following:

how complex are the features
costs of paid technologies, hosting, and software
developers’ rates and scope of work

From our experience, chatbot development takes 8-10 months to complete and costs about $75,000-$90,000 if you build it with a professional platform like Amazon Lex. This price includes the work of developers, project managers, QA testing, as well as used technologies and resources.

Wrapping up

Chatbots are getting more and more popular and surely might add dynamism, excitement, and more profit to your business. Their capabilities start from the simplest tasks and go to complex decision-making using AI algorithms.

Before building a chatbot you have to decide if you will use canned or custom solutions. Both development paths have their pros and cons, and what suits one business may be inappropriate for another. Most importantly, you should consider your needs in order to determine the perfect sort of bot for you. There is a vast multitude of options out there. You’ll surely find exactly what you need in terms of functionality, design, and price.

ChatbotProgramming Language

written by