When you think of leading companies in AI, IBM might not be the first name that comes to mind because, so far, it hasn’t been caught up in the consumer-facing AI frenzy driven by OpenAI, Microsoft, Google and others. However, over the years, IBM has significantly contributed to the field, most famously with the cognitive computing platform IBM Watson. For a detailed analysis of IBM’s prowess in this field, look at a prior Moor Insights & Strategy analysis here.

The company’s recent annual Think conference allowed IBM to showcase its hybrid cloud and AI innovations, particularly enterprise-grade generative AI (GAI). This article dives into the key highlights and my observations from the event.

What is generative AI?

Let’s get grounded amidst all the hype. Generative AI (GAI) is a form of AI that employs deep learning models to generate content based on user input. GAI uses machine learning and deep learning algorithms to create various types of content, such as text, images, video, audio, music and code.

The now famous ChatGPT, developed by OpenAI, is an example of GAI that can provide detailed written responses to user queries and engage in ongoing conversations. Other companies like Google and Facebook have also developed generative AI tools that produce authentic-looking text, images or code.

Generative AI works by training on a large dataset. Researchers input massive amounts of data, such as words, pictures, music or other content, into a deep learning system. For example, ChatGPT’s original training data, the OpenAI Codex, consists of over 700GB of data collected from various sources like books, websites and technical manuals. The system learns to identify and understand complex relationships through a supervised neural network by rewarding successes and penalizing errors. Over time, with human oversight, the system becomes proficient in generating new content.

Forbes Daily: Get our best stories, exclusive reporting and essential analysis of the day’s news in your inbox every weekday.

ChatGPT is a tool heavily used by consumers; enterprises must apply generative AI with several key considerations in mind. Accuracy is crucial, as incorrect answers from AI systems can be costly. Scalability is another consideration because enterprise-grade AI models must be deployed across hundreds or thousands of endpoints while maintaining accuracy. Strong governance is also essential to ensure transparency and accountability—and to minimize biases in AI models.

In his keynote, Arvind Krishna, chairman and CEO of IBM, focused on hybrid cloud and AI as two transformative technologies that can bring significant business value. I agree with him.

Hybrid cloud has become the predominant choice for enterprises

Krishna mentioned an IBM Institute for Business Value (IBV) study that found that over 75% of IBM customers plan to leverage a hybrid model. This finding is consistent with my long held prediction that the hybrid cloud architecture will be the predominant choice for most organizations.

Key to a successful hybrid cloud is a common platform across all clouds, on-premises and edge environments, leading to a single skill set, built once and managed from a single pane of glass. Core to IBM’s strategy is Red Hat OpenShift. IBM’s Red Hat OpenShift meets the challenge by creating a logical container layer that makes data and workloads portable across clouds.

As the hybrid cloud approach gains momentum and organizations recognize the value of leveraging both public and private clouds, Red Hat OpenShift will become crucial in achieving consistency across multiple clouds and driving IBM’s growth in the cloud market.

AI took center stage at the event

While the hybrid cloud is crucial to IBM’s strategy, AI took center stage at this event. Krishna discussed how AI could enhance various areas such as human resources, configuring price quotes, pricing, supply chain and inventory management. AI can also significantly impact IT operations, improving code development and increasing productivity by 40–80%. Customer care can be automated to handle many queries, providing 24/7 availability and scalability. In some cases, up to 90% of customer care volume is dealt with by AI agents.

AI is also crucial in cybersecurity, as it can help triage and address the large volume of attacks that organizations face. Other types of digital labor, including promotions, employee movement, onboarding and procurement tasks, can also be automated with AI, enabling enterprises to scale operations efficiently.

Krishna also highlighted IBM’s quantum computers and the potential ability to tackle complex problems in battery technology, carbon sequestration, traffic routing and other areas. He predicted that quantum computing would be mainstream within three to five years.

A coming-out party for generative AI and foundational models

While IBM’s story around hybrid cloud is well-established, its focus on AI has remained unclear. IBM Think 2023 was the coming-out party for IBM’s generative AI and foundational models. For more information on the importance of foundational models, look here.

The watsonx platform, developed over the past three years, combines AI capabilities, including machine learning, deep learning and foundation models. Krishna touted its ability to empower businesses to harness the power of generative AI. The platform is built on Red Hat OpenShift, allowing flexible deployment options.

watsonx has three key components:

1. watsonx.ai is a new studio that provides a comprehensive set of tools to create new foundation models, generative AI and machine learning. It offers increased productivity compared to traditional approaches, enabling users to work with AI models efficiently.
2. The watsonx.data lake house combines a data lake’s flexibility and a data warehouse’s performance. This component is a centralized repository to store various data types, including structured, unstructured, semi-structured and multimodal data.
3. The watsonx.governance toolkit enables AI workflows with responsibility and transparency. It helps users track the data used to train models, understand model lineage, identify biases and monitor model drift. Enterprises can ensure reliable and accountable AI deployments by consolidating governance processes into a single platform.

Enterprises must adopt an AI-first approach

The age of generative AI and foundational models will force enterprises to approach AI differently. Indeed, the transition from viewing AI as an add-on or a nice-to-have to becoming an AI-first company will be crucial for continued success in the marketplace. Every company now faces a decision: become AI-first or let the competition take the lead.

As one example of approaching AI differently, integrating AI interfaces into applications like SAP allows users to access information through AI-assisted natural language queries easily. I am already seeing foundational models that leverage enterprise data to generate accurate predictions and that improve over time by combining a company’s proprietary data with broader industry information.

Whether you are part of a 500-person or 50,000-person company, I believe the focus should be on digital labor (work performed by robotic process automation (RPA) systems.) and leveraging AI for tasks that can benefit from automation and improved efficiency.

Wrapping up

The overall sentiment emerging from Think 2023 is that IBM is renewed and reinvigorated. I have been following IBM for decades, and the excitement surrounding the company’s recent developments is unprecedented.

We will look back at Think 2023 as a pivotal moment for IBM, highlighting leadership in two major growth industries, hybrid cloud and AI, just as Arvind Krishna promised when he joined IBM.

While the hybrid cloud part made more sense, thanks to the Red Hat acquisition, I still had questions about the AI part. Think 2023 has now shed light on that with its spotlight on generative AI and foundational models for the enterprise.

IBM wants to be your go-to provider for enterprise generative AI. The company offers the necessary tools, data layer, governance and fabric to build the needed workflows, but its speed to market remains a critical question.

In conclusion, this event was a significant milestone for IBM. I will closely monitor progress, especially regarding the company’s speed and ability to deliver on the productivity enhancements offered by the hybrid cloud and AI.