IBM Think 2023: Watsonx Promises Enterprise AI At Scale With Trust
Wade Tyler Millward
‘If you look at AI adoption, it’s doubled since 2017. And that was even before this move towards generative AI and foundation models. I think if you fast forward another four or five years, penetration is probably going to be, I don’t know, if it’s not 100 percent, it’s pushing 100 percent at that point. So 80 [percent] to 100 percent. All of that is opportunity for partners,’ says Rob Thomas, IBM senior vice president of software and chief commercial officer.
IBM unveiled the Watsonx artificial intelligence and data platform during its annual Think conference and promised the first two product sets will become generally available in July.
The Armonk, N.Y.-based vendor also said that a third Watsonx product set and a host of other AI-informed offerings will become available later this year. Those other offerings include a Cloud Carbon Calculator, a Center of Excellence for generative AI, a graphics processing unit (GPU)-as-a-service infrastructure offering and the Watson Code Assistant code-generating AI.
IBM unveiled the offerings at Think 2023, which runs through Thursday in Orlando, Fla.
IBM Introduces Watsonx AI Studio
Branden Voegerl, the IBM solutions director at Carmel, Ind.-based partner CleanSlate Technology Group, told CRN that AI has been coming up more in conversations with customers, including by those asking for AI tools with more security guardrails than that offered by ChatGPT, a text-generator program that has helped popularize generative AI.
“There’s opportunities within the IBM Cloud to introduce this IBM Watson solution along with digital workspaces that customers can take advantage of,” Voegerl said. “It allows their customers to work from any device from anywhere in the world. So I’ve seen a lot of chatter about that. A lot in the systems space in terms of the AI, ML [machine learning] kind of thing.”
During his keynote address Monday at the vendor’s Partner Plus Day, IBM CEO Arvind Krishna said that customer interest in AI comes in part from the high cost of labor and desire for tools that can make sense of data collected by divisions within an organization with their own individual data standards.
“We need to have technologies that can ingest more data … and be able to put it all in and be able to very quickly digest it for some purposes. … And then leverage AI to do a task that used to be so manual,” Krishna said during the keynote. “And that is what we are wanting to do with our Watsonx.data platform. And hopefully as many of you get your hands on it over the next few months, you will give us feedback.”
Kate Woolley, IBM’s channel chief, whose formal title is IBM ecosystem general manager, told CRN that expanding partner access to the training, badges and enablement as internal IBM sales employees is an example of how the vendor is investing in bringing partners along as it unveils new cutting-edge technology offerings such as Watsonx.
“For anything, as soon as that becomes available for IBMers, that will be available for partners,” Woolley said. “And I think it’s going to be really important as well for—it’s not just the skilling, the badging. But it’s also all of the client content. So it means that when we’ve got an IBMer and a partner going to a client to have a conversation about generative AI and what that means for them, or going to a partner client to have a conversation about Watsonx.ai or Watsonx.data, they’re using all of the same content in that conversation. They’re having the same conversation with our clients.”
When asked by CRN during a virtual press conference about the opportunity for IBM services-led partners with Watsonx, Rob Thomas, senior vice president of software and chief commercial officer, said MSPs and systems integrators will play “a big role” in bringing Watsonx to market.
“The No. 1 thing I hear from clients, once they establish we’re interested in this, they ask about use cases, then they say, ‘Can you help us?’” Thomas said. “And there’s a technical angle to the assistance, which is, ‘Help me build a model, clean my data, translate this into business outcomes.’ There’s also a big change transformation piece, which is, ultimately, if you’re using foundation models, you’re probably going to change the business process, eliminate a business process, automate something. That requires a lot of assistance, typically from an integrator or a consulting-type partner, to go do that type of work.”
Already, consulting giant EY has worked with AI to streamline human resources tasks while rival Deloitte has used AI in sustainability, he said.
“There’s probably an unlimited opportunity for service providers as part of what’s happening with foundational models,” Thomas said.
Other generative AI work that would involve services-led partners, Thomas said during the press conference, includes cleaning data, finding data, structuring data, building a foundation model, customer success and more business consulting on change management.
“If you look at AI adoption, it’s doubled since 2017,” Thomas said. “And that was even before this move towards generative AI and foundation models. I think if you fast forward another four or five years, penetration is probably going to be, I don’t know, if it’s not 100 percent, it’s pushing 100 percent at that point. So 80 [percent] to 100 percent. All of that is opportunity for partners.”
The Watsonx Platform
The Watsonx platform aims to help enterprises train, tune and deploy foundation models, ML capabilities and AI models across the organization on any cloud environment with trusted data, speed and governance in one place, according to IBM.
Early business use cases for Watsonx will include:
*Customer and employee interactions and conversations
*Automating IT processes, business workflows and internal processes
IBM Watsonx.ai and Watsonx.data are the two product sets slated to be go into general availability in the summer, according to the vendor. Watsonx.governance has not received a specific GA release date.
The Watsonx platform will feature IBM-curated and trained foundation models and open-source models, according to the vendor. Users can access a data store for gathering and cleansing data for training and tuning.
Customers can build their own models or fine-tune and adapt available ones using company data and then deploy the model at scale in trustworthy, open environments, according to IBM.
The IBM Watsonx.ai product set is meant for AI builders who train, test, tune and deploy traditional ML and new generative AI capabilities powered by foundation models.
Users can leverage Watsonx.ai from data preparation to model deployment and monitoring. The product set has a foundation model library with large, curated sets of enterprise data backed by filtering and cleansing processes and auditable data lineage, according to IBM.
The models are trained on code, time-series data, tabular data, IT events data, geospatial data and other modalities. The first foundation models set will be in beta tech preview for select clients, according to IBM. Some of those early model categories will be fm.code, fm.NLP and fm.geospatial.
The fm.code category is for developers generating code through natural language. The fm.NLP collection includes large language models (LLMs) for industry-specific domains with curated data and the ability to customize with client data. And the fm.geospatial category is built on climate and remote sensing data for natural disaster change planning, biodiversity, land use and other processes, according to IBM.
IBM has partnered with Hugging Face to provide thousands of open models and datasets through Watsonx.ai, according to the vendor.
Watsonx.data And Watsonx.governance
A second product set for Watsonx is Watsonx.data, according to the vendor.
The IBM Watsonx.data data store is built on open lakehouse architecture and optimized for governed data and AI workloads, according to the vendor. Users will have access to querying, governance and open data formats.
Watsonx.data can manage workloads on-premises and across multi-cloud environments. IBM estimates that users can halve warehouse costs through Watsonx.data workload optimization. Users can access data through a single point of entry while applying a variety of fit-for-purpose query engines.
Watsonx.data provides built-in governance, automation and integrations with users’ existing databases and tools.
The third product set for Watsonx—Watsonx.governance—promises to mitigate risk, time and cost from manual processes and provide documentation for transparent and explainable outcomes, according to IBM.
Watsonx.governance promises to protect customer privacy, proactively detect model bias and drift and help organizations meet ethics standards, according to the vendor.
More AI-Informed Offerings
The vendor teased at other AI-informed offerings coming this year, including:
*Watson Code Assistant, which allows developers to generate code with commands in English
*A preview for a geospatial foundation model-enabled IBM Environmental Intelligence Suite (EIS) Builder Edition offering used for mitigating environmental risks
*A full-stack, high-performance, flexible, AI-optimized infrastructure delivered as a service on IBM Cloud for foundation model training and serving
*The Cloud Carbon Calculator dashboard, which uses IBM Research technology for measuring, tracking, managing and reporting carbon associated with hybrid multi-cloud
All major software products going forward will receive an infusion of Watsonx.ai foundation models, according to IBM. The vendor plans to combine a natural language processing (NLP) foundation model with Watson Assistant and Watson Orchestrate for improved employee productivity and customer service.
The vendor will also launch a Center of Excellence for generative AI with more than 1,000 IBM Consulting professionals supporting client deployment of Watsonx and other AI, according to IBM.