IBM Watson IntegrationEmil Rauhe
IBM Watson Integration makes it easy to maintain and expand your chatbot’s knowledge.
Once you have set up your IBM Watson chatbot, it usually doesn’t take long before there is a need to extend it with more answers to your existing questions. As well as expanding the possibilities for answers with new questions in the database. It can be done in different ways. As you gain more and more experience with your IBM Watson chatbot, you’ll quickly discover new areas that this intelligent bot can cover – such as sales and support! The database needs to be expanded with new topics very time you start trading new products or increase the product categories in the business. Read on to learn how you easily can maintain and expand your IBM Watson chatbot.
Expansion IBM Watson manually
It can quickly become a burdensome process to expand IBM Watson’s knowledge, as every time you want to add new possible answers, you must first define new intents, entities and dialogs.
To remedy this you can choose to do this add-on manually in the built-in user interface (UI). Here you can define new intents, entities and dialogs manually one by one. It is time-consuming and you can easily loose track, ending up correcting your work again and again.
Another option is to code all of this into one or more JSON-files which are then uploaded to the IBM Cloud. This way you can create new Watson Skills! On how to do this read more about it in my guide til JSON-filer. This procedure of course demands that you can code JSON-files or have a developer who can.
Expand IBM Watson with your own databases
Luckily there are easier ways to expand IBM Watson chatbot’s repertoire of possible answers, as many of the answers the customers need you will already have these in different databases within your company. Or these could be found amongst, for example, your suppliers support- and product databases.
Imagine; that all your internal knowledge bank from the support department, plus all your product catalogue/customer database can be filled into the IBM Watson! Or that you can make Watson pull on your supplier’s support databases! This can create endless possibilities for new detailed and more advanced answers.
This can be done in two different ways, either by using IBM Watson Discovery Integration or the IBM Watson Integration that we have developed here at LHCom.
Both methods make it possible to do a dynamic integration. This means that as the data bases you integrate with Watson receives more data, the more IBM Watson Chatbot continually gets expanded with new options for helping customers.
IBM Watson Discovery integration
IBM Discovery is IBM’s intelligent AI-data base solution. It comes with the great advantage of being able to analyse your data as it continually finds new and sensible connections in even the biggest databases.
In this way it can continuously find new search-words (Keyword Extraction), new semantic connections (Semantic Role Extraction) and new emotional indicators (Emotion Analysis). This makes it possible to build new intents, entities and dialogs based on the search-words, semantic connections and so on, as IBM Discovery identifies them.
Unfortunately IBM Discovery is not free, as soon as you need to search in more than 1.000 records.
The biggest challenge is however, that if you want to use the system’s Enrichment Model to draw new possible answers from the data base, you must convert and upload it to Discovery’s own database format for it to work.
IBM have developed a few integrations e.g. SalesForce, Microsoft SharePoint or Box, if you apply other database formats than these you’ll have the tedious job of programing the integration; or will have to convert and upload data manually to Discover’s database format. It is here, LHComs IBM Watson Direct Integration becomes a helping hand.
LHCom IBM Watson Direct Integration
With LHCom IBM Watson Direct Integration, you can integrate any database with the IBM Watson using a special webhook which pushes the data in a dynamic integration up to IBM Watson, by converting the data to JSON-format.
The JSON-format is the modern standard for data exchanges and has largely taken over this status which the XML and CSV-files had earlier. Therefore today most databases and database systems support the export of JSON.
This gives unimaginable possibilities for feeding your IBM Watson chatbot with new possible answers.
Simultaneously you get the option for adding logic to the answers and making them more varied too, as you can ‘sense’ and manipulate the JSON-files which feeds the way it affects the answer, ensuring that you do not get stuck in a certain database system, but can continuously add new databases with new relevant knowledge. No matter which format they are in.
Nevertheless, you must still define new intents for unknown questions, but the integration solution opens up for infinite options as you can quick and easy pull in for example new product groups or support answers directly from your suppliers. The IBM Watson chatbot can serve the customer as soon as you’re open for the sale of new products from the webshop.
Let us help with your Chat bot!
A chat bot can answer your customers questions, and increase your sales.
We are experts in IBM Watson Chat bots and can help your to develop your own chat bot.Contact us
LHCom IBM Watson Direct Integration makes it also possible to present HTML-formatted answers to your own customers. It makes it easier to present a step-by-step guide or an index for customers in a more organised and user-friendly way.
Here we have developed a solution for the organisation Hallberg Rassy Club. In this example, all the members own an exclusive boat from the Swedish boat manufacturer Hallberg Rassy. They have an online community where data is located in a HCL Domino database. From here is it especially easy to send data to IBM Watson due to HCL Dominos database that is already using JSON.
In the database, the IBM Watson chatbot can both find answers to a question about the different types of boats, and it can also collect answers from the member-database.
So if you as an example ask “Boats in England?” IBM Watson will answer with a HTML-formatted list of all the approximately 30 Hallberg Rassy boats that lie in England. As well as which harbor it belongs to, and which member is the owner also.
The underlying live search uses the LHCom IBM Watson Direct Integrations webhook. It pulls the data from HCL Domino database and collects this data in a JSON-object. Which then gets HTML-formatted and placed in a link that IBM Watson chatbot then gives as the answer to the user.
IBM Discovery integration with LHCom IBM Watson Direct Integration
Hallberg Rassy Club also had a wish to give its members access to the clubs internal Q&A database, which is also in the HCL Domino.
However, the answers in this database are not as plain as for example membership lists and boat data. Therefore the club wanted to store this data upon an IBM Discovery-database so they could exploit its enrichment-possibilities.
This issue can also be managed with our LHCom IBM Watson Direct Integration by using the webhook to automatic push data from the clubs HCL Domino database onto IBM Discovery. After this, it’s processed as before.
In this way members could get answer to more avanced questions like:
What treatment is best for my teak deck?
This question starts the bot searching for the answer in the databases of members and boat data. But because there is no useful answer to be found there, it forwards the Watson question to IBM Discovery.
It then returns a list of possible answers, where the words teak, treatment and deck are included. Those answers are rated after its relevance and presented as a HTML-formatted list for the users. Based on the rating value in Confidence-column the users can easily see which answer seems most relevant to them.
Do you want to make your IBM Watson chatbot smarter?
If so, I can help you integrate all the knowledge that you already have in your company’s and suppliers databases using the LHCom IBM Watson Direct Integration.
This integration also allows for logging all the answers that IBM Watson gives as well as providing insight into which questions the users asks where the chatbot does not give an appropriate answer. So by implementing LHCom IBM Watson Direct Integration you can also enrich and expand Watson so the users get a better experience with the chatbot.
Furthermore, in the future you can expand and maintain the IBM Watson chatbot using internal resources, instead of hiring expensive consultants.
Therefore, you will over all get easier, cheaper and a much smarter chatbot with LHCom IBM Watson Direct Integration. It makes it easier to maintain the IBM Watson chatbot, and it will benefit your customers experience with the company’s website and webshop.
And read my PDF presentation here:
LHCom are experts in chat bots and IBM Watson. We can help you in building your next chat bot for your website.
Contact us and lets talk about your project.