chatbot_framework_comparison

There are three overarching chatbot frameworks available.

  • The 'DIY' chatbot, built using a platform like Chatfuel or Manychat.
  • The proof of concept chatbot, a hybrid between a speedy building tool and a bit of customisation. 
  • The tailored chatbot, a full-on custom solution built by a professional developer (or, more likely, a bunch of people including devs, linguists, designers, etc.).

In this article, you will learn what each framework's advantages and disadvantages are. My hope is you will use the content below to inform your decision and start your business chatbot on the right framework.

 

The chatbot framework comparison table


  DIY

Proof of concept

Tailored
Core features      
NLP Limited & Pre-set Limited & Custom Unlimited & Custom
Channels Limited & Pre-set Unlimited & Custom Unlimited & Custom
Human fallback No Yes Yes
Small talk Limited & Pre-set Limited & Custom Unlimited & Custom
Exit feature No Yes Yes
Onboarding sequence Yes Yes Yes
Multi-media Yes Yes Yes
FAQs Limited & Custom Unlimited & Custom Unlimited & Custom
Custom dashboard No Yes Yes
Advanced features      
Machine learning No No Yes
Sentiment analysis No No Yes
Authentication No Yes Yes
Integrations No Yes Yes
Webhooks Yes Yes Yes
APIs No Yes Yes
Gated flows No Yes Yes
Voice No Yes Yes
Mobile marketing Limited Unlimited Unlimited
General      
Cost None / Low Medium High
Support None / Low High High
Drag & drop Yes No No

 

Editor's note: most of the line items we use in the comparison table are chatbot features. To learn more, head over to our chatbot features page.

At ubisend, we build both proof of concepts and tailored chatbots (along with all the features you can see up there). When building a tailored chatbot, we always go down the proof of concept route first. More on this below.

I know, this is a lot of information to digest. You may also not have full knowledge of all the terms on the comparison table.

So, let's break each framework down in a few sentences. I'll also share some pros and cons for each.

 

The DIY chatbot framework

When chatbots became more popular about 3-4 years ago, chatbot platforms started popping up.

Those platforms allow users to quickly and cheaply build 'chatbots' themselves (hence the 'do it yourself' connotation in my title).

Most people who have heard of chatbots could name a few platforms, such as Chatfuel or ManyChat.

chatbot_framework_diy_chatbot 

Pros

  • Cheap
  • Easy to start
  • Easy to use
  • Great for a quick mockup
  • Great for wider technology adoption

 

Cons

  • Limited NLP
  • Limited functionalities
  • Limited (or inexistent) support
  • No data ownership
  • At the mercy of the platform

 

As we can see, going down the DIY chatbot framework route has its pros and cons.

Though we build custom chatbots and some may believe that we compete against these platforms, we have always felt they serve an important market. Sure, they are very limited in functionalities and NLP. Sure, platform users don't retain data and might even sign T&Cs that sign data over to the platform.

But, a DIY chatbot is a great framework to follow to test the waters. These platforms also do a great job at spreading the word about chatbots, which is ultimately what we believe in.

 


The proof of concept chatbot framework

A proof of concept (PoC) chatbot is a lightweight version of the fully tailored solutions we build.

It's sort of the middle ground between the DIY framework (free/cheap but rigid and limited) and the tailored framework (built to your exact requirement, unlimited features & advanced AI, but requires budget).

I can't remember the last time we did not encourage a potential client to take advantage of our PoC framework.

chatbot_framework_proof_of_concept_chatbot 

Pros

  • Cheaper than a tailored solution
  • Up and running in a matter of weeks
  • Focused on learning
  • Metrics-oriented
  • Informs future roadmap

 

Cons

  • Not fully tailored
  • Limited APIs/webhooks

 

We believe the proof of concept approach offers the best of both worlds.

It's cheaper than a fully tailored solution yet includes enough customisation to be a real learning experience. Building a proof of concept solution tends to take us four to eight weeks, which also speeds the learning process up.

Finally, as opposed to a DIY chatbot, following the proof of concept framework allows you to measure the specific metrics that matter to your business. Again, this is perfect for your learning experience.

Is a sales chatbot the right tool for your business? You could get an almost definite answer within two to three months, with the data to back it up.

Even if you do not come to us for your chatbot solution (but, I mean, why not), please consider asking your agency for a lightweight version before investing in the full version.

 


The tailored chatbot framework

The cream of the crop.

Following the tailored chatbot framework is exactly what it sounds like: getting precisely what you need.

Tailored chatbots are perfect for businesses that know what they want, have specific needs in terms of language, APIs, data structure, reporting, and more.

 

chatbot_framework_tailored_chatbot 

Pros

  • Suits your precise requirements
  • Performance-driven
  • Unique to you
  • Own everything
  • Complete control
  • Scalable

 

Cons

  • Investment (time and money)
  • Planning
  • Long term project

 

The tailored chatbot approach has clear advantages. It's all about you, your needs, your company needs, your business goals. You, you, you.

This is the core of what ubisend does, and we found that the biggest appeal (beyond the uniqueness of having your own solution) is making the solution completely performance-driven.

A proof of concept chatbot will take you half the way there. You could implement a PoC chatbot that looks at one specific user journey and build a dashboard around that. But, a fully tailored solution allows you full control over all the aspects of your chatbot, which means you can be truly performance-driven.

ubisend_chatbot_implementation_path_short

ubisend's methodical chatbot implementation strategy (learn more).

 

Of course, the other big appeal (especially for mid to large businesses) is owning all the data. By following the tailored chatbot framework, a business can own everything that goes in and out of the chatbot.

The cons of a tailored chatbot are, arguably, not really cons. A larger investment is to be expected when getting a fully custom solution. It also takes time to build and implement, which makes sense. Finally, it takes more planning and involving more bodies in the project.

 

Choosing the right chatbot framework

If you want to be granular, use the table above to pick your framework. Tick whichever line items you need and see what you come up with.

Like I said earlier, even if you do require a tailored solution, any serious chatbot development company would advise you to go down a proof of concept path first. It makes the most business sense (on both sides).

If you believe a DIY chatbot would do the trick, by all means, focus on this.

At a very basic level, the core questions you need to ask yourself when choosing a chatbot framework are:

  • What is our business goal?
  • How custom do we need our chatbot to be?
  • How data-secure does our chatbot need to be?
  • How much budget do we have?
  • How much time do we have?

Asking yourself (and your team) these five questions and working through the table above will give you the right framework. And, of course, come to us for all your proof of concept and tailored chatbot needs.