Is the Cognitive Space a game changer?

Sebastian Gurt

Hi all,

 

what do you think about the cognitivespace? What use cases do you see in existing solutions and in the future especially for the nlq (native language querying)?

 

How do you explain and demo this to new / existing customers?

 

Do you see the need to change development guidelines like using prefixes to categorize cubes (because "Sales Budget" is searchable, but "REP Sales Budget" not so much)?

 

Looking forward to your thoughts and input.

 

BR

 

Sebastian

Gabriele Gallo

Ciao Sebastian,

i think cognitivespace and BOARD functionalities like nlg might be a useful instrument for approaching data querying and business intelligence from a different perspective.

It sets the basis for a new approach to data navigation which is less "structured" (i.e. start from a predefined or existing report) and more "tailored" or "specific", therefore based on a specific question (i.e. "show me the budget and actual for year 2017" typed or dictated in our cognitive space area).

 

The way i would envisage its use today, would be more like an advanced search or a sophisticated "query" feature, that would allow users (who don't navigate and open the BOARD solution on a daily basis), to find the information they are looking for, in few seconds.

Think about a CEO or a C level executive, who might have access to the application once a month or once per quarter, and needs something specific, today: say that he doesn't remember where this piece of information might be placed or if he needs to create an ad hoc analysis.

 

With the cognitive space, he doesn't need to browse our capsules or think about where the cashflow analysis might reside: all he needs to do is "ask" for the "cashflow report" or the "cash balances for november 2017" and potentially get a chart with the cash balances per month or day, and, a list of screens and capsules where this analysis resides. So he might get to the information faster, without anyone's help, and potentially find additional useful information which is related to the question.

 

As of today, i would say it is important to "tag" or "name" metadata and data in a consistent and "easy to browse" way, which you would need anyway, if you would like end users to "play" with reports or to fully understand every bit of information that resides into the BOARD database.

 

What do you guys think?

Helmut Heimann

Hi Sebastian Gurt and Gabriele Gallo

 

plus, we'll have to bear in mind that language is crucial in the querying process - everything you want to search for (and find) has to be defined in the language you're using in the query...

My idea to cope with that would be some sort of "tagging" the underlying database and making those tags searchable by the cognitive space as well.

I'm not quite sure if that's what you meant, Gabriele.

 

Kind Regards

Helmut

Unknown

I agree with the comments above. nlg and cognitivespace are promising features. I think there needs to be some time for users to feel more comfortable with the experience before it will catch on. Making information more accessible is certainly a benefit, but there needs to be a way of capturing the context elegantly. When someone asks a question about cashflow, the right answer is always, "it depends". Trying to answer a question without sufficient context may confuse folks, or not fully address their concern.

Daniel Zillmann

I absolutely agree with all the comments above. But I must say that the current state of the cognitive space had some negative impact on demonstrations for us in the past. Siri, Google Assistant and Alexa pushed intelligence personal assistance into the focus for everyone and people are expecting more when it comes to asking questions to a computer. I think Helmut Idea about using "tags" for cubes and entities is great, I assume that this would make it a lot easier for users to type in questions and receive the right answers.