• March 29, 2024
  • By R "Ray" Wang, founder, chairman, and principal analyst, Constellation Research

Achieving Trust in the Age of AI

Article Featured Image

Back in 2016, the World Economic Forum, running out of ideas to inspire, dubbed the move to digitization and digital transformation “the Fourth Industrial Revolution,” but in reality, that was instead a necessary half-step to the move toward artificial intelligence. While some will now profess that the Fifth Industrial Revolution is upon us, AI is more exponential than the internet or the fake Fourth Industrial Revolution of eight years ago. The dawn of the real Fourth Industrial Revolution begins with AI, and this cognitive era will be with us for decades to come. This new world will place great value on “data to decisions”—data that is accurate and reliable enough to guide trustworthy AI.


Digital transformation has made more data available by digitizing the physical world. Companies that led the way in digital transformation built a foundation to take advantage of AI. In 2023, companies rushed to AI for an exponential advantage. Unfortunately, many organizations soon realized they did not have enough data, or computing power, to achieve the accuracy needed for precision decisions. In fact, Constellation estimates that 91 percent of companies will determine they lack sufficient data to achieve a level of precision that their stakeholders would trust. Some will find that their internal data is good but not good enough for many use cases.

For example, would 85 percent accuracy be sufficient for customer service and support? Probably yes. Organizations could hire one less person on the front line taking orders, or one less support person. Would 85 percent accuracy cut it for procurement? No. Operations would fail without the critical supply chains at the right time. Would 85 percent accuracy make a CFO feel confident in finance? Definitely not. People would go to jail. Would 85 percent accuracy be OK for patient visits in healthcare? Absolutely not! Lives would be put at risk.

Sadly, hundreds of millions, even billions, will be wasted because organizations did not develop a data strategy. Without enough precision data, AI cannot be trusted. Hallucinations in generative AI lend well to creativity but become dangerous when users expect precision and trust. As skepticism grows around AI, mistrust will increase. The timeline over the next five years will look like this (see the figure below).


Trust and transparency require constant training and a voracious appetite for ever-growing amounts of data. Yet the half-life of data makes much of your data worthless in seconds. The battle for more data and increasing number of signals requires data mastery. Data mastery leads to not only better AI use cases but also a new class of organizations known as “Data Inc.” companies.

Deep data resident knowledge of the flat sides of data provides rich understanding. Unique datasets and partner models will augment synthetic data approaches to feed the beast. Despite best efforts, no organization will have enough data.

For this reason, hybrid models for data will persist, as a shift back to on-premises storage for mission-critical datasets will accelerate because of security concerns and for efficiency reasons. Data collectives will bring richer and higher-precision inputs. Small language models will become as valuable as large language models as data collectives orchestrate marketplaces of insight. Data collectives will provide insight amid the “dark ages” of post-AI, where publicly available information is rare and useless. Why? The only data left on public sites will be misinformation, credit card offers, and bad promotions.


Winners will emerge to reap exponential advantages. This new class of Data Inc. companies will be valued according to their data as well as their revenue. Moreover, these companies will emerge as their own asset class valued by the market for their ability to create flywheels of monetization. Five types of Data Inc companies will emerge:

  1. Companies with unique datasets
  2. Companies with data-driven digital networks
    (network + data)
  3. Companies with longitudinal datasets
  4. Companies with derived data advantage
  5. Companies with new classes of data


In discussions and workshops with more than 50 leading enterprises, seven rules have emerged to create trusted AI entities and Data Inc companies:

Rule 1: Achieve data mastery.

Rule 2: Understand how to partner for data sources
and signals.

Rule 3: Generate new derivatives.

Rule 4: Monetize outcomes.

Rule 5: Engage stakeholders.

Rule 6: Nourish networks.

Rule 7: Trust but verify.

In the post-AI world, the quest for more data sources will result in the dark ages of publicly available data. Most entities will not want their data to be publicly available for ingestion by other AI models. Hence, high-quality public data will be scarce in the future.

To source high-quality data, data collectives will emerge to create trusted AI models in private networks. Data collectives will reward small language models for providing the last mile in precision. Solid data strategy will drive the valuation of companies in the future and provide stakeholders with trusted AI.

As organizations build AI for machine scale, humanizing AI will require different design aesthetics. Organizations will have to determine when they will insert a human in the process and at what cost-benefit to the organization.

In the age of AI, trust will emerge as the North Star to guide Data Inc companies. What will your organization do to achieve it? 

R “Ray” Wang is the author of the new book Everybody Wants to Rule the World: Surviving and Thriving in a World of Digital Giants (HarperCollins Leadership) and founder of Constellation Research.

CRM Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues

Related Articles

The 12 Major Risks of AI

Opportunities abound and optimism is warranted, but so is a degree of caution.

CX Leaders Must Prepare for Exponential Efficiency

Paradigm shifts await customer experience organizations.

Generative AI’s Use Cases Are Already Emerging

ChatGPT has lit up imaginations, but the underlying technology is more than a shiny object.