Enterprise AI 2.0: The Next Wave of B2B AI Innovation

The Evolution of Enterprise AI: From Hype to Real Results

After two decades of AI implementation, it's reasonable to suggest that the anticipated improvements in efficiency and profitability haven't fully materialized for many businesses, despite the initial excitement.

Superficial analysis of current data seems to validate the concerns of AI skeptics. A substantial 90% of data science initiatives fail to reach full deployment. Furthermore, only a fifth of analytical findings generated by 2022 are projected to translate into tangible business results.

Even organizations with comprehensive, company-wide AI strategies are experiencing failure rates as high as 50%. However, this should be viewed in context.

Understanding Enterprise AI 1.0 and 2.0

The last quarter-century represents merely the initial stage in the development of AI within enterprises – a period we can define as Enterprise AI 1.0. Many organizations currently operate within this framework.

However, forward-thinking companies are already progressing to the next iteration, which will shape the future of big data, analytics, and automation – Enterprise AI 2.0.

The distinction between these two generations of enterprise AI is significant, not merely theoretical.

Implications for Business Leaders

For executives across diverse sectors – including healthcare, retail, media, and finance – the transition from 1.0 to 2.0 presents an opportunity to learn from previous setbacks.

It allows for the establishment of realistic expectations for future applications and provides justification for the increasing investments in AI observed across various industries.

Companies that successfully achieve Enterprise AI 2.0 will likely emerge as leaders in the economy.

The Path to Future Success

These organizations will have distinguished their offerings, gained a larger portion of the market, and established a foundation for continuous innovation.

Conceptualizing future digital transformations as an evolution from Enterprise AI 1.0 to 2.0 offers a useful framework for business leaders.

This model aids in developing strategies to effectively compete in an era defined by automation and advanced analytics.

Key Takeaways:

• Enterprise AI is evolving through distinct phases.
• Learning from past failures is crucial for future success.
• Strategic adaptation is essential for competitive advantage.

Enterprise AI 1.0 (the Status Quo)

Beginning in the mid-1990s, the field of Artificial Intelligence was characterized by speculative experimentation and exploratory research. These endeavors were largely confined to the realm of data science professionals.

As noted in a recent Gartner report, these initial attempts resembled “alchemy,” performed by specialists whose skills were difficult to replicate throughout an organization.

However, the limitation posed by the data science bottleneck – the necessity for all work to pass through a limited group of experts – wasn't the sole impediment to widespread adoption. The effectiveness of AI is fundamentally tied to the data systems it utilizes.

Many organizations experimenting with AI during this period possessed data fragmented across isolated systems, coupled with insufficient data infrastructure and processes for optimizing the technology.

Furthermore, initial B2B AI solutions centered on intricate, horizontal “machine learning” platforms geared towards model creation. Successfully implementing these custom-built models presented a significant challenge, requiring substantial customization and integration with existing enterprise applications and workflows.

These Enterprise 1.0 solutions were often unwieldy and difficult to manage, despite demanding significant investment for deployment.

The majority of projects originated from grassroots efforts. Data scientists initiated them as exploratory investigations, focusing on hypothetical applications often disconnected from core business goals.

Consequently, many projects remained purely scientific in nature, and the overall failure rate was remarkably high.

In 2017, Gartner’s Nick Heudecker estimated the AI project failure rate to be as high as 85%. Heudecker emphasized the need for a clear path to production, stating that many organizations treated big data as “technology retail therapy” without proper planning.

Even those projects showing initial promise encountered further obstacles during the production phase, stemming from issues related to transparency, trust, potential biases, ethical concerns, and broader governance considerations.

Due to the limited scope of the few projects that reached production, the overall impact fell short of the strategic expectations surrounding accelerated innovation, enhanced competitiveness, improved customer satisfaction, increased profitability, and greater productivity.

This situation is poised for a shift, however.

Enterprise AI 2.0 (2021 to ~2030)

The COVID-19 pandemic served as a catalyst, significantly accelerating the pace of digital transformation across various enterprises. Businesses spanning banking, retail, and entertainment sectors transitioned their operations and customer interactions to online platforms, becoming entirely reliant on digital tools.

This widespread adoption subsequently established the groundwork for Enterprise AI 2.0. This term characterizes a novel generation of automation, analytical techniques, and operational practices designed to enhance efficiency and foster sustained profitability.

Unlike the initial phase of Enterprise AI 1.0, which heavily depended on costly and extensive model training using shallow learning methods, the subsequent generation of AI possesses the capability for more complex analysis. It incorporates advanced unsupervised learning and utilizes models pre-populated with semantic intelligence, minimizing or eliminating the need for training.

A notable shift is occurring within many organizations, where AI initiatives are now receiving direct sponsorship from C-level executives. These leaders fully recognize the critical need for digital transformation.

Rather than remaining isolated data science endeavors, Enterprise AI 2.0 is now a fundamental driver of business model transformation, systematically applicable across diverse operations and industries.

The importance of this executive support cannot be overstated. A recent McKinsey survey on analytics adoption revealed that robust commitment from all management levels was a key differentiator for the top 8% of companies in terms of AI performance.

While many AI projects undertaken in the past two decades yielded limited results, these initial explorations into big data were essential. They cultivated the current environment conducive to successful AI implementation.

Most companies have now invested in data lakes, warehouses, and feature stores – systems designed to process, standardize, and analyze incoming data streams.

Enterprise AI 2.0 leverages this existing infrastructure to overcome the obstacles that hindered the operationalization of models in Enterprise AI 1.0. These solutions are equipped with pre-built, industry-specific knowledge and capabilities.

This allows for efficient, streamlined, and widespread deployment, unlocking the potential of AI for all employees. The most effective outcomes are achieved through hybrid intelligence – a synergy between human expertise and cutting-edge machine learning algorithms.

As machines continue to learn from human interactions and emulate our cognitive processes, these solutions will only become more potent. In Enterprise 2.0, technology will progress beyond “machine learning” to “machine reasoning.”

This advancement will enable AI to semantically comprehend user activities, contexts, and events, interpret findings, explain results, pinpoint root causes, suggest decisions, or execute optimal actions.

In contrast to the narrowly focused tactical tasks of Enterprise AI 1.0 projects, 2.0 solutions will expand their scope to encompass complete end-to-end processes.

These processes will feature integrated systems of intelligence, delivering insights and discovery to end-users, augmenting decision-making, or enabling autonomous operations across various business functions.

The result of this innovation is the emergence of a “self-driving enterprise.” This transition necessitates a heightened level of trust in these systems.

Consequently, we will witness the development of robust governance mechanisms to oversee AI systems – envisioning AI oversight bodies monitoring data flows and supervising AI operations.

We are currently in the nascent stages of realizing this vision. Many organizations remain constrained by challenges related to their strategy, culture, personnel, technology, or processes.

However, the self-driving enterprise represents the clear trajectory forward, and Enterprise AI 2.0 is the crucial next step for leaders aiming to achieve this future.