Item: IBM Watson Discovery
Rating: 8
Author: Verified User

Overall Satisfaction with IBM Watson Discovery

Use Cases and Deployment Scope

We have a huge amount of complex events to analyze and find value on their patterns. In this way, our test environments with IBM Watson Discovery demonstrated to be really useful. We also tried to do the same with ChatGPT and other similar AI tools.

Pros and Cons

Pros

Saves a lot of time analyzing documents
Generates interesting insights from them (non evident)
Performs with an acceptable performance over complex information

Cons

Language feature (currently in Beta stage) that's really necessary for our field. We analyze lots of documents related to items sold. This data contains lots of language specific information.
Better insights on the ingestion process, in order to detect errors or failures.
Better documentation on every step of its data processing.

Return on Investment

We find its Enterprise plan expensive for a country of LATAM. For US or Europe based businesses, looks great.
A Big Data and massive queries based company would find the service expensive. Maybe a flat price plan would be helpful.
Have you thought in making a cheaper plan where you take the learning from your customer's data to enrich your AI tool?

Alternatives Considered

Azure AI Search, Azure AI Document Intelligence, ChatGPT, Google Gemini, Amazon Elasticsearch Service and Elasticsearch

IBM Watson Discovery resulted more robust and performant, also the insights were much more interesting than just an AI search from Microsoft or a prompt for ChatGPT.

Key Insights

Do you think IBM Watson Discovery delivers good value for the price?

Not sure

Are you happy with IBM Watson Discovery's feature set?

Yes

Did IBM Watson Discovery live up to sales and marketing promises?

Yes

Did implementation of IBM Watson Discovery go as expected?

Yes

Would you buy IBM Watson Discovery again?

Yes

Other Software Used

Azure App Service, Azure Blob Storage, Azure API Management, Azure Cosmos DB, Microsoft IIS, Apache Pinot, Azure SQL Database

Likelihood to Recommend

You have lots of sold items data and you look for patterns that would been useful for your customers (usually e-shops). In this case, that's great to process with stability massive amounts of information and finding answers to questions still not asked.

Also business as usual insights are provided in a reasonable time. What still should be clearer, imho, that's how to the process performs internally and if we are doing good or poor use of the tool (also an AI based advisor would be great).

Comments

Please log in to join the conversation

A good finding for a big data start up