IBM Channel Chief Woolley: Our AI Is Not For ‘Writing Poems About Your Dog’
Wade Tyler Millward
‘This is not about trying to be everything to everybody. This is not about writing poems about your dog. This is about curated models that can be tuned and are very targeted for the business use cases that we want the enterprise to use them for,’ Kate Woolley, IBM ecosystem general manager, tells CRN.
As IBM seeks to differentiate itself in the ongoing artificial intelligence gold rush, solution providers will be a key part of its go-to-market, channel chief Kate Woolley told CRN.
AI based on open technologies, trust, accuracy, promising customers control and ownership of their data plus targeted use cases for businesses are ways that IBM and its partners will separate themselves as the Armonk, N.Y.-based vendor increases access to its AI tools.
“This is not about trying to be everything to everybody,” said Woolley, whose formal title is IBM ecosystem general manager. “This is not about writing poems about your dog. This is about curated models that can be tuned and are very targeted for the business use cases that we want the enterprise to use them for.”
IBM AI Partner Opportunity
Bo Gebbie, CEO of IBM solution provider Evolving Solutions, No. 135 on CRN’s 2023 Solution Provider 500, told CRN that he is interested in seeing how IBM’s AI offerings benefit the security and infrastructure business, which have been in demand for Gebbie’s company lately.
Customers have asked Gebbie’s team about using generative AI tools such as ChatGPT—made by Microsoft-backed OpenAI—and automating tasks related to human resources, Gebbie said.
“If IBM can help bridge the gap between how we make meaningful use cases and we can make money at it helping clients, I think that’s the part we’ve got to do some more homework on,” he said. “The reason I’m excited and interested to learn more is we know there’s a skills gap and shortages with skilled employees still. [We need to learn] how does this help make certain tasks less labor-intensive so that we can redeploy the right people for the harder tasks and harder work?”
Two products within the IBM Watsonx AI suite became generally available in July—the Watsonx.ai studio for foundation models, generative AI and machine learning and the Watson.data data store for scaling AI workloads with an open lake house architecture.
The Wastsonx.governance product for directing, managing and monitoring an organization’s AI activities is expected to become generally available in November, according to IBM.
IBM’s Partner Plus program, which was introduced in January to replace PartnerWorld, is a “building block” for how IBM plans to scale AI through partners, Woolley said.
“Whether that is through skilling our partners … but then also providing a very simple and transparent program that’s easy for our partners to understand,” she said. “So as they take our AI technology to market, they know exactly what support sits around that, how they get skilled. Then they have a very clear line of sight to that predictable earnings around, for example, our software stack and what that looks like.”
IBM celebrated a slate of milestones for Partner Plus in a July 27 blog post, including:
*The addition of more than 3,000 new IBM registered partners over the past six months
*New co-marketing benefits to make the best use of funds available at each program tier
*70,000-plus enrollments in one or more IBM badges across partner organizations since October, when IBM introduced new skilling capabilities
*More than 40,000 completed badges since October
*The ability for partners to embed Watson Natural Language Processing (NLP) and other core IBM AI technology to make application experiences more intelligent
*The ability for partners to embed Watson Discovery to automate more core business workflows
*Watsonx proficiency badges, product training and new marketing kits
*A new developer page of free generative AI educational courses and content
And although Wastsonx.ai itself is GA, for now users only have access to its non-generative Slate family of encoder-only models for fast enterprise NLP tasks, according to IBM.
Models coming soon include:
*Granite models based on a decoder-only, GPT-like architecture for generative tasks
*Sandstone models with an encoder-decoder architecture meant for fine-tuning specific tasks
*Obsidian models with a new modular architecture developed by IBM Research that provides high inference efficiency and levels of performance across a variety of tasks
The AI opportunity applies to IBM global systems integrators and smaller partners as well, Woolley said.
“It’s across the board,” she said. “I have not spent time with a partner recently that is not having conversations with their clients about AI no matter what size the partner is. I’ve been with some of our biggest partners who are having these conversations, but also with some of our smallest niche partners in certain areas of the IBM portfolio.”
Here’s what else Woolley had to say about the role of IBM partners in bringing the vendor’s AI offerings to market.