Just 11% of CIOs Have Fully Implemented AI as Data and Security Concerns Hinder Adoption
As CIOs develop their data infrastructure and AI expertise in real time, the need for trusted solutions — like Agentforce — over DIY approaches is clear
Eighty-four percent of enterprise CIOs believe Artificial Intelligence (AI) will be as significant to their businesses as the rise of the internet. However, only 11% say they’ve fully implemented the technology, citing an array of technical and organizational challenges, led by security and data infrastructure, that must be overcome first.
The data, which comes from a new Salesforce survey of 150 verified CIOs of companies with 1,000 or more employees, offers a snapshot of the state of enterprise AI, along with the hurdles ahead that must be addressed as companies pursue their AI strategies and ramp up their adoption of agents.
Key findings include:
- CIOs feel pressure to be AI experts. Sixty-one percent of CIOs feel they’re expected to know more about AI than they do, and their peers at other companies are their top sources of information.
- CIOs agree that AI is a game changer, but are cautious. Eighty-four percent of CIOs believe AI will be as significant to businesses as the internet, but 67% are taking a more cautious approach compared to other technologies.
- IT is focusing on data initiatives before leaning into AI. CIOs report spending a median of 20% of their budgets on data infrastructure and management, versus 5% on AI. Security or privacy threats and a lack of trusted data rank as CIOs’ biggest AI fears.
- Business partners must examine their AI timelines. Sixty-six percent of CIOs believe they’ll see return on investment (ROI) from AI investments, but 68% believe their line-of-business stakeholders have unreasonable expectations for when that ROI will occur.
- CIOs see a mismatch between departments when it comes to AI. While functions like customer service are seen as having the most AI use cases, they may be perceived as being the least prepared for the technology.
“Generative AI is one of the most transformative technologies of this century,” said Juan Perez, CIO, Salesforce. “This research gives a glimpse at the foundations that CIOs across industries and geographies are laying in preparation for the rollout of truly transformational AI.”
AI implementation lags behind stakeholder expectations, introducing risk
AI use is rapidly increasing in the workplace as business stakeholders see opportunities for increased efficiency. In fact, 77% of CIOs say they have good or excellent executive buy-in on AI’s value. Yet CIOs are concerned about their business partners’ urgency, with 68% believing there are unreasonable expectations of when they’ll see ROI from the technology, for example.
Large shares of business leaders across sales, marketing, service, and ecommerce view themselves as having “fully implemented” the technology into their workflows. There’s evidence that much of this adoption is unsanctioned, introducing significant security risks as workers send sensitive data through unsecured LLMs.
“We don’t yet know how to use AI among the larger employee base and how to control its governance.”
— CIO, professional services industry, United States
On the other hand, just 11% of CIOs — with their greater technical expertise and broader view of the organization — say they’ve fully implemented AI — 18 to 38 percentage points less than their line of business counterparts.
“The adoption of mass market generative AI tools by workers is ushering a new era of “shadow AI” that highlights the urgency of implementing trusted tools,” said Perez.
CIOs address data and security as a precursor AI
The slow pace of sanctioned, enterprise-wide AI strategies and tools can be attributed to a focus on preparatory work that CIOs must tackle first.
Given the transformative nature of AI, 67% of CIOs say they’re taking a more calculated approach to its implementation compared to other technologies. CIOs have no shortage of hurdles to consider when implementing AI, but those related to security and data are by far the most common.
While safety concerns can be addressed by selecting vendors with robust security infrastructure and ethics and credibility safeguards, ensuring that the data underpinning AI is recent, high-quality, and accessible can be a particularly daunting challenge that was underscored by several interviewees across industries.
“The biggest barrier to AI implementation is data accessibility. Additionally, legacy systems and outdated infrastructure can hinder the effectiveness of AI.”
— CIO, manufacturing industry, United States
“We have lots of old data centers and legacy applications that don’t natively support data exchange.”
— CIO, financial Services industry, Italy
“Our data lacks a single source of truth, and it’s not timely.”
— CIO, retail industry, India
Recognizing that they must get their data in order before they can fully embrace AI, CIOs are currently allocating, on average, four times more budget toward data initiatives.
“Data initiatives aren’t new, but they’re at an unprecedented level of urgency and prioritization,” said Perez. “Nearly every company I speak with is shifting resources to ensure their data is integrated, accessible, and relevant.”
Still, CIOs feel uncertainty over how much to allocate toward AI in these early days of the technology. Only 47% are confident they’ve allocated the correct amount of budget to AI initiatives.
CIOs struggle to identify where to prioritize AI
Beyond the challenges of shoring up their data foundations, many CIOs are finding it difficult to define where and how AI should show up in their broader organizations. This is especially true given the novelty of AI and a lack of understanding, or even fear, of the technology across departments.
“There’s a lack of understanding of how AI works and consequently a lack of trust in it. So far we don’t have many credible use cases or studies to instill confidence.”
— CIO, architecture, engineering, and construction industry, United Kingdom
And while AI use cases may be more apparent in certain areas of the business, those departments may not be the most receptive or capable of harnessing the technology. For example, CIOs view customer service as having the most use cases, but also as the least enthusiastic. On the other hand, marketing is viewed as eager to use AI, but least prepared across factors like skill sets.
Some CIOs are finding that this mismatch between AI business value, enthusiasm, and readiness requires a more surgical approach to the technology’s implementation across the enterprise. “Leaders have a unique opportunity to showcase AI across their enterprises and to demonstrate to skeptical employees that AI can help — not hinder — their work. Effective enablement is critical to empower those employees with the necessary skills, tools, and guidelines to drive tangible value so their organizations can begin to trust AI and believe in the power it delivers,” noted Jeff Amann, EVP & GM of Salesforce Industries.
“AI cannot be a one-size-fits-all top down approach. Our intent is that departments and geographies find opportunities and value from AI that fit their needs while, whenever possible, leveraging common technologies and platforms.”
— CIO, architecture, engineering, and construction industry, France
As such, some CIOs are finding it more effective to launch pilot projects that showcase AI’s power in order to make the case for broader implementation. In fact, 75% of respondents describe their organizations as being in the experimental stage of AI adoption.
“We are experimenting by implementing in select functions/departments and monitoring their usage, adoption and benefits.”
— CIO, technology industry, India
CIOs turn to each other for AI education
As their bosses focus on AI’s business implications, CIOs are under what is arguably the most pressure of their careers to help define and execute strategies. AI is not only an unfamiliar technology, but one that is evolving at a remarkable pace. As a result, CIOs must come up to speed, and fast.
“There’s a lack of sufficient AI expertise at all levels of the company, including me.”
— CIO, financial services industry, the Netherlands
However, CIOs know they aren’t alone when it comes to an AI knowledge deficit. While more than three out of five CIOs think stakeholder expectations for their AI expertise are unrealistic, only 9% think their peers are more knowledgeable.
As they seek to develop expertise in a rapidly advancing technology, CIOs are turning to familiar sources of information like analyst firms, technology vendors, and technology media outlets. Yet CIOs are most likely to trust their peers at other companies on this topic, a trend that Salesforce’s own CIO welcomes.
“This is the perfect opportunity for knowledge sharing,” said Juan Perez, CIO, Salesforce. “Today’s CIOs haven’t experienced many technology shifts this consequential in their careers until now. We’re all navigating this revolution in real time, and it’s critical that we share our learnings with each other so we build not just successful AI enterprises, but a successful AI economy.”
More Information
- Explore Agentforce
- Learn more about Sales Agents
- Learn more about Service Agents
- Learn more about Industries AI
- Salesforce’s AI Use Case Library
- Read the Trends in AI Report
- Watch the CIO Corner Series
Methodology
Salesforce commissioned NewtonX to administer a double-anonymous survey of 150 verified enterprise CIOs worldwide. The survey was administered between July 23 and July 30, 2024 and yielded responses across 18 countries and 17 industries. An enterprise is defined as a company with at least 1,000 employees.
