Artificial Intelligence, chatbot

Rule-based versus AI-based chatbots: Which is best for your business?

17 Oct, 2017
6 Min Read Dan Ward

Rule-based versus AI-based chatbots: Which is best for your business?


Recent Posts

Subscribe to our blog

Chatbots are fast becoming the must-have marketing tool of the decade, with companies including Starbucks, Unilever, Sephora, Uber and even the Singapore Government hurrying to deploy them.

Capable of simulating a human conversation via text or voice, chatbots are making it easy for organisations to quickly engage with customers, effortlessly answer customer queries, and cost-effectively nurture customer relationships.

As chatbots grow in popularity, competition between chatbot development companies has increased. This has significantly reduced the cost of building and implementing a chatbot. And with bots capable of much more than they were just five years ago, the benefits are rapidly outweighing the costs.

More than a marketing tactic or gimmick, consumers view chatbots like personal assistants – able to help them wherever and whenever they need it. ‘Immediate availability’ has become a requirement, and brands that respond are quickly setting themselves apart from the crowd.

If your organisation is contemplating a chatbot, here are a few factors to consider…

How will you use it?

It’s vital for a bot to have a specific purpose and fulfil a specific function.

For example,takeaway stores such as Dominos have chatbots that answer customer queries, take orders and facilitate payments through Facebook Messenger.

Digital media companies use bots to encourage notification sign ups, and then allow the user to ask for the news they want to see, rather than having to search for it.  

Travel and online shopping companies use bots to help customers find the best online deals, fast access to flight or accommodation information, or product suggestions based on requests or previous behaviour.

Tech and internet companies are creating bots capable of troubleshooting problems with a device or network.

Other companies are simply deploying chatbots as brand characters to make people laugh and ensure a memorable customer experience - such asStarbucks’ divisive pumpkin spice latte bot.

By pinpointing how you intend to use a chatbot, you can make an informed decision about the type of chatbot to use.

Which chatbot suits your needs?

The most popular way of deploying a chatbot is by integrating with a mainstream messenger app, such as Facebook Messenger, WhatsApp or Slack. Most chat apps have made it easy for developers to code in a bot which, once installed, can communicate with anyone who uses the platform.

The real question then becomes: Rule-based or AI-based?

Rule-based chatbots

Rule-based chatbots are question-answer bots. They follow specific rules (i.e. “If the customer says this, I say this”). They are the most popular type of chatbot on the market today.

Rule-based chatbots are cheaper and faster to implement than AI-based chatbots, and work well in situations where requirements are relatively basic.

Unfortunately, they cannot understand intent or context, so they must be programmed very specifically to account for the varied ways humans communicate. For example, how many ways can you ask for a pizza? “Order a pizza”, “Buy a pizza”, “I need pizza” – plus consider other languages, and different cultural nuances.

Rule-based chatbots are best used for simple tasks in which there is no need to simulate human intelligence.

AI-based chatbots

AI-based chatbots learn as they go rather than following a script.

Developers are now able to create algorithms using advancements in machine learning and natural language processing, which prompt chatbots to draw valid assumptions from masses of data rather than following a set of rules.

So when an AI-based chatbot asks a question, it uses the answer to learn more about the customer in order to provide future personalised responses. In essence, it becomes aware of the user’s needs.

AI-based chatbots can also learn about a wide variety of topics with minimal input from the developer. This makes them a better long term solution for fulfilling multiple functions.

However, artificial intelligence, machine learning and natural language processing are still far from perfect. AI-based chatbots have a long way to go before they can simulate true human intelligence. But given the simplistic requirements for most chatbot uses, AI capabilities – even as they are now – are usually unnecessary.

The best way to determine which type of chatbot suits your needs is to work with an experienced implementation partner. They can assist on everything from quality and pricing to comparing rule-based and AI-based capabilities, and integrating with a messenger platform.

As we head into this exciting new era of customer engagement and conversational commerce, both types of chatbots will form an important part of many organisational marketing and customer support strategies.  


For more information on how to choose and deploy a chatbot, download our white paper: