Ask HN: Is anyone using Watson in their business?

124 points by WorksOfBarry ↗ HN
Analytics, or language, or anything Watson

50 comments

[ 2.0 ms ] story [ 92.9 ms ] thread
Thank you for asking this question.

I work at a large bank and we've seen the Watson sales pitch every year for four years now. Every time they visit we ask them this question.

It would be great to hear from folks using it.

What do they say when you ask them that question?
Since it's IBM, they probably have duped a few logos into buying a trial subscription as part of another sale of a more traditional product.

Voila, now everybody is "using" it

where "using" = sitting dormant.

Watson is probably not for companies whose core competency is software. There's plenty of less glitzy and more functional machine learning tools out there. Have you tried scikit-learn and nltk?

Watson seems to me like it's mostly leadgen to sell IBM consulting services to companies that don't know how to build software in-house.

>Watson seems to me like it's mostly leadgen to sell IBM consulting services to companies that don't know how to build software in-house.

This is exactly what it is. It's a marketing product designed to help IBM sell the company overall as a big data company. The only people the commercials are targeted at are C-level execs who don't know the first thing about data science but know they have to invest in it to stay competitive. I don't work for IBM, but I'd wager a lot that the idea for it was created by a sales team and not an engineer.

That said, IBM is certainly not the only business to do this and it's not even clear that it's not a smart business strategy. I'd wager that most enterprise software sales use deception as a major part of sales strategy.

How would a solo dev use it if they were interested?
get a free 30-day trial of Bluemix. The Watson tools are accessible through Bluemix and there's a free tier that allows you to try out things for yourself.
I worked with a large company that was one of IBM's 'partners' for watson. They mostly used the partnership for the free publicity, but complained about how terrible all of the disparate watson services were. In the end they ended up replacing most of the services with google ones, but maintained the public partnership.
I saw an advert this morning on Twitter for Watson Content Hub. I checked it out[1]. Honestly it looked underwhelming as the only part that Watson came into play was auto-tagging the contents of an image when uploading. Using the Watson branding for such a narrow aspect of the product seemed more like marketing wishes than actual innovation.

[1] https://www.ibm.com/products/watson-content-hub

I believe Fark uses it to auto-generate story tags when they go green.

I'm not actually sure if this is a compliment, now that I think about it.

Yes. We use the Natural Language Understanding service at www.bustle.com for doing categorization of content. It maps decently to IAB categories: https://www.iab.com/guidelines/iab-quality-assurance-guideli...
Did you guys build a custom model or is the IAB mapping something Watson offers (I can't seem to find any details)?
The mapping is custom. Not something they have out of the box.
Thanks for giving an actual example. Looks like it's a tiered model, first classing into broad categories and then the specific labels? What was integration like: fairly straightforward or was there a long setup/tuning process?
Very straightforward. We send text, they send back categories. We then map those to IAB categories in our own code
I'm familiar with 2 large public companies that got the sales pitch, bought in and results were underwhelming or failed. From an inside sales person "we are great at selling vision and lackluster at delivering results".
Not by any chance in the pharmaceutical industry? Because that story seems eerily familiar.
Has to be the financial industry... because it is also eerily familiar.
Watson was much hyped for Cancer research and it has remarkably failed, there was a news article a long time back. I remember that everyone in my company was like "Watson Watson" but then the AI craze stopped. Good day are back!
What exactly is Watson except a marketing term? I mean, as far as I understood it, it's Apache UIMA with some bare bones NLP bolted on, using RDF ontologies to power some inference. But I've never seen it being used in the wild except as a tool to win over people to the IBM-side, raking in large contracts ultimately doomed to fail due to lack of domain knowledge. It all seems so ... familiar. Now everyone is rushing into Deep Learning for NLP, but apart from things like sentiment analysis on synthetic data sets ... it's still /hard/ and needs constant attention. I really don't see how you can "IBM Watson" stuff unless you're already heavily invested on integrating NLP/ML into your workflow ... and that's tricker than it looks. The stuff on BlueMix seems nice and all, but far from a silver bullet or golden hammer when it gets down to the nitty gritty of real world data.
"IBM Watson" is a group of consultants and a brand. It's not an algorithm or a service or an "artificial intelligence". Don't fall for the marketing.

https://news.ycombinator.com/item?id=14766793

https://news.ycombinator.com/item?id=14682816

https://news.ycombinator.com/item?id=14767407

I'd give some insights, but I don't want the stock price to go down before I liquidate. (Or get sued) Hoping Q3 earnings are good b/c of mainframe sales, fingers crossed!

I was hired to work on a 8+ digit contract to apply Watson, office mate and I (only engineers on the team) were laid off 2 months into our new jobs. Do what that company did, put cancellation clauses in any contract should IBM not deliver.

With Watson, IBM appear to be trying to stop "nobody ever got fired for buying IBM" from being true any more.
I’ve been using Watson’s speech-to-text libraries to automate the transcription and tagging of videos of our usability tests.

Mixed results, so far on how accurate it is, but I also recognize that there are likely a number of improvements that can be made to the parameters that I’m setting for each processing run. I also have only started to play with the custom modeling.

Right now it’s a toy, but promising for our small automation needs.

If you are processing big volumes of speech data its much better to setup Kaldi ASR on your own infrastructure. Same or better accuracy, ability to train and many other nice things.
I use it to transcribe voice messages that come in through our PBX. The PBX sends the voicemail mp3 to Watson, it sends back the transcription along with a guess (%) on how accurate it is (then the PBX emails the transcript attempt and voicemail attachment to whoever). It's almost always good enough to know what the message is about, sometimes perfect, often hilarious.
Does it return a confidence % for the whole transcription or does it rate each word/phrase?
The PBX usually chunks the send, Watson returns a % for each chunk it processes and then the overall average upon completion.
We brought the Watson consultants on for the pitch, total cost for our org was going to literally be in the millions and it was just going to be simple NLP and integration with a couple systems. The audio NLP seemed cool, the text NLP seemed easily replicated with modern libraries.
Why does IBM keep separate branding for Watson and Bluemix AI? It seems like most "Watson-in-the-field" deployments are leveraging the Bluemix pre-baked models for specific workflows like document classification, transcription, etc.

Seems like they'd better align the marketing and the offering if Bluemix was the "Watson gateway" so to speak. They could still sell and market Watson as the custom consulting they do, with Bluemix branded as an entry point to basic Watson functions. There's some speak of this in their product copy, but it could probably be more obvious.

I have a friend that works at a famous credit checking company in their data research team, here in Latin America.

He told me that IBM's data service (ie Watson) was not at all impressive. At the time (about a year ago) their service was not worth the pay. He told me all the solutions they provided, and the techniques they used (in terms of algorithms and infrastructure) were very easy to implement themselves (I mean, it was an internal data research team).

But, in our conversation, he told me that there were reasons to go for some AWS dervices (such as integration and provided apis) or Google (the amount of data their stuff in trained on, the infrastructure they have) could be pointed out as reasons to hire AWS or Google. He couldn't find any that suggested hiring Watson.

I thought IBM was a blockchain on mainframe company?

Really though, I don't think bluemix/watson has anything GCP/AWS/Azure don't have at this point. Plus it looks like a bunch of their machine learning research got the axe.

I was loosely involved with a project at a very large company where they were using Watson (all on-premise, no cloud) to classify customer feedback (verbatims) into an existing taxonomy. This was actual customer feedback that came through service channels, not social media chatter.

The taxonomy basically determined the escalation process within the company. As an example, think about issues with a cell phone. "Sometimes my phone volume goes down automatically" isn't as critical as "Sometimes my phone gets really really hot in my pocket."

The Good: They felt Watson was doing a good job classifying the customer feedback. Overall, they felt they could rely on Watson and could defend using Watson instead of human raters in a legal situation.

The Bad: The amount of time/resource investment required to get it up and running had been way beyond what they initially expected. They also felt various parts of the system could have been more open (e.g. Watson's back end database) to inter-operate with downstream systems. They also were exploring if other solutions (e.g. Data Scientists + Python) could do this and had started to see positive results.

Sounds like the perfect fit for machine learning. Let people continue classifying the input text, store the input text and classification, keep doing this for a while and you get a large training set. Regularly run ML on it and compare results. Switch to ML when its results are good enough.
As a contract developer on a social media project we started to use Watson for analysis of posts about companies (that were clients of our client) and we had to go another route due to the time constraints of using Watson as well as the fact there were easier ways to use existing libraries that did "one" ML / algorithmic task to accomplish what we were looking to do (sentiment mostly).
Isn't that particular offering basically just the Alchemy API they acquired and rolled into the platform?
I believe Watson, the AI engine, is still an early concept for healthcare and the reason I say this is that my friend is working on a healthcare startup that collects a vast amount of information. Their partnership with IBM means they are required to share the information that they're collecting, possibly to provide Watson with training datasets in the healthcare space.
It is glorified robbery. My experience is that they are smooth talkers, roll deep, excellent sales people, and the execution and product are lackluster. Try to roll your own where possible or find another vendor.
Watson has a classifier, NLC. Anyone that knows what type of classifier they are using in NLC? Take long time to train, but the respons are quite fast.