2Know Magazine: Sharing KM Knowledge
2Know: Sharing KM Knowledge
March 2020 - Magazine No. 246
March 2020 - Magazine No. 246

Nowadays, many businesses tend to incorporate Chatbot components into their websites to improve customer service and leverage sales. You've obviously encountered the  icon inviting you to ask any question you might have regarding the website, an invitation actually referring the business's Chatbot service.

What is a Chatbot? A Chatbot is a computer application which simulates a human conversation via audio or textual commands.

How does the Chatbot operate? When audio commands are received, the bot converts the audio data into text using automatic speech recognition technology. The software then analyzes the textual input by retrieving central words and classifying and inspecting the phrases. Based on modules it is familiar with, it then considers the best response to this query. Finally, the Chatbot responds to the user via text, audio output or by performing the desired task.

There are three main categories of such tools:

  1. Business principles-based tools: this type of Chatbot's operate according to a predefined group of principles for each specific type of situation/dilemma. These tools try to foresee what the user might and accordingly define the tool's answers relying on these scripts. They make use of if-then logic and review the words that appear in the question, their order, the use of synonyms or mapping different ways to ask questions that can be answered by a similar response.

This main disadvantage of this type is that it is confined to a set of predefined rules and scenarios and cannot answer any question asked beyond this narrow scope. Furthermore, it cannot learn and draw conclusions based on previous conversations. This tool is practically a more interactive display of the familiar FAQs.

  1. AI-based tools: AI-based chatbots rely on two central elements: Machine Learning and Nature Language Processing (NLP). Machine Learning is comprised of different layers of data learning and analysis. Inspired by the human brain, each layer is comprised of a network artificial neurons of its own which interact with one another. Each connection is calculated according to previous learning patterns or previous events, and each input of data initiates another "learning" experience. The tool discovers new patterns in the data without using any previous knowledge, then retrieves the information and stores it in a template.

NLP is a branch of AI that assists computers in thinking, interpreting and analyzing human language. This branch allows computers to communicate with people in their language and evaluate misunderstood data related to language. For example, NLP enables computers to read a text, hear speech, interpret it, assess emotion and determine which parts of said text are most important. For more on the matter, click here.

AI-based Chatbots are usually more sophisticated, interactive and personally adapted than principle-based Chatbots. As their databases grow over time, they become more 'aware' of the content's context, leverage their understanding of natural language and implement their response patterns better to provide a personally adapted User Experience. Furthermore, they are able to answer questions beyond their area through context. These bots are an innovative solution that can handle multiple, simultaneous calls and thus improve the UX and in turn increase sales. For examples of AI-based chatbots, click here .

However, in order to attain a AI-based chatbot of even the most basic level requires large amounts of basic data and designated professionals in the field to perform the required adjustments. These all require many resources not every organization can afford.

  1. Hybrid tools: the hybrid tools attempts to bring out the best of both worlds and merge AI technology's abilities with the intellect and linguistics of a human representative. The representative can interfere occasionally, especially when a more human response is needed. Thus, by collecting questions that couldn't have been previously been answered by bots and exposing the bots to the appropriate human answers, can the organization expand the bots' base of knowledge.

This type has other advantages: the bot can channel the messages to a human representative only when the questions are of high importance or when they seem too complex for the bot to handle. They can also turn the call to the most relevant human representative according to the nature of their call. The bot can also notify the representative when the caller is about to give up on the call and allow the representative to get involved.

Chatbots' other advantages include their 24/7 availability, consistency of response, and their ability to withstand any overload. That said, one must remember that handling complex responses is still a task performed better by human representatives. Businesses should consider how their representatives can complete their customer service array.

According to a study by IBM, 80% of customers' regular questions and tasks can be handled by bots. The key to success is comprehending the customers' request and responding in a relevant and personally adapted manner. Therefore, when incorporating this tool into a business, it is vital to define the needs and objectives the tool is meant to support and adapt the tool to these purposes. once a tool is chosen, the most critical stage is the handling of the "knowledge" and information which are central components of the tool. Without a reliable base of knowledge and mapping the required templates, the machines won't be able to optimally expand, create or use the knowledge.

We are far from solving all of our problems, and you might still be wondering which type of bot is most preferable in your organization's current state. However, any organization managing calls, even if not yet ready to purchase any bot, should start learning, following and becoming familiar with this technology. It is a matter of time till it becomes a widespread phenomenon.

Related Articles











Written By Sarah Karsenti

Cognitive accessibility is a term referring to a simplification of an environment so to make it easier to understand or navigate.

Cognitive accessibility is meant to assist the cognitively impaired population, such as those suffering from neurological developmental disabilities, autism, brain injuries, dementia, learning disabilities, or people dealing with psychological disabilities. They all involve some damage to their abilities to perform processes that involve thinking: difficulties in understanding instructions, remembering, symbolizing, abstracting, quantifying, reading and writing. These difficulties affect their ability to conduct themselves independently throughout their everyday life.

Cognitive accessibility can be implemented on several levels, from the most basic level to the highest level. Hereby are some examples:

  1. Content pages: phrasing content simply along with graphic accessories.
  2. Pictures/signs: simplification of the environment so to make it easier to navigate through or comprehend. For example, one can place easily understood signs instead of written instructions or color each flight of a large building's parking lot a different color so to make it easier to locate one's parking spot.

Signs with simply understood pictures instead of written instructions







  1. Tool Tips: subtitles which pop every time one points at an icon or a word displayed, containing information on this word.



Examples of tool tips:






  1. Manuals: a manual phrased simply along with graphic accessories.

I wish to share an example from my personal experience of a project in which I am taking part this year.

In short, I've learned that cognitive accessibility can be implemented not only simply or for cognitively impaired audiences (as seen in the aforementioned examples), but also for wider audiences. This project is a complex one with all parties involves located in a different country. The processes itself is a repetitive cycle, though each cycle involves different participants. Due to the multitude of participants and the location complication, the discussions between the participants and myself (as project manager) are conducted online. We also keep our shared documents on a cloud on a SharePoint page, so that everyone can edit, and update said documents simultaneously.


Here's one issue that requires some improving. I believe that saving the file on the cloud so that all parties can work on it simultaneously is critical for an efficient management of such a project. However, over time I've discovered that some teams prefer other work methods. Some have never worked with shared files saved on clouds and can work better with features on a document and save precious time. Some were just intimidated by the size of the document and did not know how to start working on it.

What did I do? I created a work manual on working on documents saved on clouds. The manual includes an explanation how to edit and save the documents on the cloud, a link to the SharePoint page on which the documents are saved, graphic accessories, screenshots, tips on how to work with navigation and comment windows, emphasis and directions regarding efficient and correct work on these documents.


an example of said manual



In conclusion, I've explained what cognitive accessibility is and how it is implemented in familiar environments. Furthermore, and this is this article's main value added, while its main contribution was to the cognitively impaired (and is rightfully defined as such)- but we all use it on a daily basis. We all use a manual, and we all use manuals- fine examples of cognitive accessibility.








Written by Rom Knowledgeware
Fax 077-5020772 * Tel 077-5020771/3 * Bar Kochva 23 st., Bnei Brak Postal: 67135