I want to ask about just the enrichment of the data - not the functionality (or lack thereof) of the AI Assistant.

We support hundreds of Cognos instances on version 11.2.4 and 12.0.4. Some of our customers have large amounts of data and some customers have not much data. All customers have the same ~30 Framework Manager packages each with ~50 tables/views with complex joins.  

(1) how long does enriching a large package take? Would large be the complexity of the package and/or amount of data?

(2) how reliable is the enriching process? does it often fail without finishing?

(3) have you ever had a glitch while enriching where it has affected the cognos content store?

(4) does data enriching take a lot of system resources. will it interfer with other instances of Cognos running on the same server.

(5) each time that we have a release with a new deployment of FM packages, does the data have to be enriched again?

Any other information that you have would be appreciated.

------------------------------
brenda grossnickle
BI Programmer Analyst
FIS
------------------------------

Hi Brenda

1 - This is going to be dependent on the server (how much horsepower it has) - and how large the models/data is.  There is some documentation/best practices available that you will want to follow to really utilize the AI Assistant.

2 - Reliable?  Hmmm -- Ask google this question "Can I register for the kids Pittsburgh marathon in person the day of the event"

Gemini AI provides an answer of "no".   The reality is "yes you can" for I did this last year and "AI provided a No answer to me last year and I was able to register 30minutes before the heat took off".   The saying of "garbage in, garbage out" is still true with AI -- maybe worse.   In Cognos, if the data is garbage, you will get some bad results returned.

Does it fail?  Yea, it can fail and when it does, you will need to do the enriching manually (table by table). 

3 - I have not noticed this behavior.

4 - Enriching is time and memory intensive.   Other instances on the same server?  I guess it comes down to how many instances you are running on each server - horsepower of the server and how much memory is available -- it is memory intensive.

5 - If you make a change to the package which changes the behavior of the data/model, yes you will need to enrich it again. 

Here are some links to the documentation regarding the assistant.

https://www.ibm.com/docs/en/cognos-analytics/12.1.x?topic=packages-enriching

https://www.ibm.com/docs/en/cognos-analytics/12.1.x?topic=SSEP7J_12.1.0/com.ibm.swg.ba.cognos.ug_ca_dshb.doc/t_ca_assistant_tips_best_practices.htm

Hope this helps.

------------------------------
John Cusack
Analytica iQ
------------------------------