Would be interesting to know future roadmap. I reviewed IBM Bob for our company, currently we mosly use Claude opus 4.6 for almost everything, which is both expensive and overkill.

But i need to babysit the code a lot more with IBM bob. And it doesn't always make good judgements. For example we have a csv of apps from the google play store and I asked it to find the ideal candidates to make a replacement paid app. And Bob decided to look for apps with high stars and price between $1 and $9 that has been updated last 12 months. Which is ridiculous because if there is a app with terrible ratings, and super high price that has not been updated in 10 years but are still popular then it would be excluded. Claude Opus managed to at least not pre-filter out potentially great candidates.

Bob also is terrible at leaving all sorts of scripts everywhere and just keep making new scripts instead of fixing old ones. Also I could not straight away use custom skills with slash commands, or at least the UI didnt make it obvious. Although bob understood what i was trying to do when I attempted.

It feels like going back to how AI was a year ago. So at the moment there are better alternatives for our use. But I am very interested in keeping track of Bob and following it along. Is there a way to do that? What is planned and so on? Would be good if IBM developed some benchmarks to compare to alternatives so that we know when it is good enough or surpass the competition, which I assume is just a matter of time.

To me, this sounds like a use case issue. I don't see IBM BOB as a AI Chat agent, but as a coding assistant to do work around code, including modernization, security in code, quality in code, modernization, and so on.  

For items that you asked to compare applications is a ChatBot style question that would be geared more for Claude or Perplexity or ChatGPT.