Partners Now Can Embed IBM Watson Software Libraries Into Their Solutions

For the first time, systems integrators, ISVs and MSPs can embed select IBM Watson software libraries directly into their solutions. IBM on Tuesday released three libraries consisting of code that provides specific Watson AI and machine learning functions.

Ultimately, IBM intends to offer many of the Watson libraries for developers to integrate into their solutions or service offerings. The first available library includes IBM Watson Natural Language Processing (NLP), which lets developers add context using sentiment analysis.

The other two libraries include speech-to-text, allowing programmers to provide fast and accurate transcriptions, and text-to-speech for converting text into natural sounding audio.

IBM’s decision to offer these libraries to ecosystem partners and customers is a departure for the company. Until now, IBM only offered the Watson portfolio as products, which include Watson Assistant,  Watson Discovery, IBM Instana Observability, IBM Maximo Visual Inspection and IBM Watson APIs.

Licensing the libraries promises to offer partners a more affordable and potentially faster way to add AI to their offerings. Access to the libraries also lets developers without deep AI or machine learning skills integrate Watson code into their solutions.

Monica Aggarwal, VP for IBM’s build ecosystem and technology partnerships, said the libraries will appeal to the broad partner ecosystem. For instance, a communications provider can use it to enhance the automation of its unified communications service offering, Aggarwal explained.

IBM’s Monica Aggarwal

“The model that we serve works for an ISV to embed into their solutions and for a managed service provider to put in as part of their application and then offer it as a service,” Aggarwal told Channel Futures. “The technology that we offer works across the scale of partners that we have.”

A More Granular Approach

The libraries also give partners a piecemeal alternative to build AI into solutions, noted IDC analyst Ritu Jyoti. Rather than require partners to provide a complete Watson product and use its tooling, they can embed specific capabilities.

IDC’s Ritu Jyoti

“IBM needed to make it more granular,” Jyoti said. “This provides faster time to market. I’ve been telling them for a long time that they need to do this, and I’m glad they finally are.”

Aggarwal emphasized that IBM had never previously offered the Watson AI libraries.

“In fact, not a lot of our competition provides them,” Aggarwal said.

Some, such as Microsoft, do, she acknowledged. For example, Microsoft provides its Intelligent Document Processing APIs to leading RPA provider UiPath. The two companies last month announced a broad partnership designating Microsoft Azure as UiPath’s preferred cloud.

But many partners don’t require a holistic offering like that, Jyoti noted.

“It’s the same thing with IBM. They have all these tools, but not everybody is going to buy the entire suite of tools,” she said.

A case in point is SingleStore, an ISV that offers a multi-model database used for real-time decision-making. SingleStore wanted to add natural language processing to its database, which supports various relational, document and text data models. Embedding it into the database versus using a cloud-based AI service was critical because it’s used for real-time decision-making.

SingleStore’s Yatharth Gupta

“Our customers will now be able to take advantage of the NLP algorithms and run analytics such as sentiment analysis,” said SingleStore senior VP of products Yatharth Gupta. “Customers across industries running marketing analytics greatly benefit from this. They no longer need to move data outside the database to run these functions. Moving data, as we all know, is prone to reliability and latency issues and is a very expensive proposition.”

Other partners that IBM revealed using its embeddable Watson libraries include EquBot, CrushBank, and Sherloq.

Besides offering more options for partners, providing the software gives IBM a better way to monetize Watson in ways it could not do in its own, IDC’s Jyoti noted.

“They can’t do it all,” she said. There are too many industries and business processes.”

Aggarwal said IBM is offering various licensing options for the software libraries, such as calls per month or minutes used.

“Our licensing model is going to be extremely competitive,” she said. “It depends on what they sign up for and what they buy or what’s in the contractual terms.”