Are Your Internal AI Apps Falling Short?
It seems like poorly designed internal AI apps are missing the mark when it comes to providing the experiences that employees really need to thrive. This gap is only fueling the rise of shadow AI within organizations.
Despite a whopping 92% of companies planning to increase their AI investments, only 21% of office workers believe that AI apps significantly boost their productivity. This discrepancy is causing businesses to struggle with bridging the 71% gap between expectations and reality. It’s high time for organizations to step up and enhance the employee experiences delivered by their in-house apps.
“The biggest challenge in enterprise AI adoption is that companies are investing heavily, but employees aren’t reaping the benefits,” shared Vineet Arora, CTO at WinWire, in a recent interview with VentureBeat. “It’s not about the algorithms; it’s about usability. When AI tools lack the intuitiveness of tools employees are already comfortable with, adoption suffers, and shadow AI steps in to fill the void.”
Most employees resorting to creating shadow AI apps aren’t doing so with ill intentions. Rather, they are struggling to cope with increasingly complex workloads, time constraints, and tighter deadlines.
“Every day, we encounter around 50 new AI apps, and the count has already crossed 12,000,” revealed Itamar Golan, CEO and co-founder of Prompt Security, now a part of SentinelOne, during a recent discussion with VentureBeat. “Approximately 40% of these apps default to training on any data provided, which means your sensitive information could become embedded in their models.”
Golan likened this scenario to doping in the Tour de France, emphasizing how individuals seek an edge without fully grasping the long-term repercussions.
Bridge the Expectations Gap to Curb Shadow AI
A recent study by Ivanti highlights the stark contrast between employee expectations regarding AI apps and the actual benefits delivered. Through confidential interviews conducted via Signal, VentureBeat has uncovered innovative methods employees across various sectors, including consulting, finance, and marketing, are employing to leverage AI for enhanced efficiency. However, this trend comes with the risk of confidential data seeping into unauthorized AI models.
Legacy UI Practices: A Catalyst for Shadow AI
“In my recent interactions with customers, I’ve noticed that enterprises often underestimate the significance of UI and UX when deploying AI tools and solutions,” noted Arora. “Employees compare every enterprise app with the ease of use found in ChatGPT and other AI applications they use in their personal lives. Most enterprise AI solutions fail to match the natural and effective feel of consumer-grade apps, leading to poor adoption rates.”
Constructing AI tools based on outdated usability standards only serves to encourage shadow AI. IT teams must step out of their comfort zones of developing internal apps in traditional ways to deliver exceptional new employee experiences.
The aftermath is becoming foreseeable as shadow AI gains ground. VentureBeat continues to uncover the proliferation of shadow AI financial analysis apps integrated with APIs from top AI firms like OpenAI, Perplexity, and Google. Consulting firms are leading the pack in embracing these tools, viewing them as a safeguard against potential layoffs. By the year-end, an estimated 115,000 shadow AI apps will be seamlessly integrated into client delivery workflows, with mobile apps witnessing the most rapid growth.
Shadow AI poses a $670,000 problem for most organizations, unbeknownst to them. Breaches stemming from employees’ unauthorized use of AI tools incur an average cost of $4.63 million, nearly 16% higher than the global average of $4.44 million.
The $4 Million Productivity Paradox
While most IT teams have a roadmap for their current and future AI app initiatives, there exists a significant variance in how AI usability is defined. Relying on UI and employee experience strategies that worked well for past generations of internal apps, when applied to new AI apps capable of delivering profound insights, inadvertently creates more hindrances than productivity.
Ivanti’s 2025 Digital Employee Experience Report highlights that companies are losing an average of $4 million annually due to employees abandoning apps owing to poor UI design and the resulting friction. Not surprisingly, 27% of employees are going rogue, shifting 73.8% of workplace AI to personal ChatGPT accounts beyond the purview of security teams.
Digital friction stands out as a key factor contributing to lost employee productivity. On average, employees face 3.6 tech interruptions and 2.7 security update disruptions per month. The cumulative loss in productivity and time within a typical 2,000-person organization can effortlessly reach the $4 million mark identified by Ivanti’s research team.
The Greater the Frustration, the Deeper the Shadow
Most organizations lack visibility into the efficiency and value of their internal AI apps. Only 67% monitor Digital Employee Experience (DEX), which furnishes data on how employees truly interact with technology. Mid-sized companies fare better in tracking DEX performance, with 81% actively involved in enhancing app productivity. Absence of DEX metrics renders IT teams clueless as to why their AI investments aren’t yielding productivity gains or why employees resort to crafting and sharing shadow AI apps.
Source: Ivanti 2025 Digital Employee Experience Report
Unveiling the Reality Behind the Curtain
Inadvertently subpar employee experiences are fostering the development of productivity boosters that employees leverage to accomplish more in less time and gain a competitive edge at work.
The more severe the time crunch and tighter the deadlines, the more prevalent shadow AI becomes, especially in consulting. Entire departments are equipped with shadow AI apps to amplify productivity within compressed timeframes. “I see this unfolding every week,” noted Arora. “Departments latch onto unsanctioned AI solutions because the immediate benefits are too enticing to overlook.”
“Shadow AI is today’s version of shadow IT, albeit with much higher stakes,” Arora cautioned. “Employees aren’t acting maliciously; they’re responding out of sheer frustration. Security teams face a lose-lose situation: either they block shadow AI and lose out or they craft enterprise-grade experiences that mirror consumer-grade ones and emerge victorious.”
“Most conventional IT management tools and processes lack comprehensive visibility and control over AI apps,” Arora observed, elucidating why enterprises struggle to curb shadow AI. His assessment strikes at the core of the issue. Companies must transcend legacy processes and acknowledge that adaptability, agility, and speed are pivotal for optimal AI app performance. Processes and workflows that sufficed for a homegrown CRM, ERP, or order management system fall short for AI apps.
Arora pointed out that entire business units are leveraging AI-powered SaaS tools under the radar. With independent budget authority granted to multiple line-of-business teams, units are swiftly deploying AI sans security endorsements. “Suddenly, you’re left with numerous unknown AI apps processing corporate data without a single compliance or risk assessment,” Arora highlighted.
“The most astute CISOs and CIOs I collaborate with aren’t drafting new policy manuals or devising fresh security protocols,” Arora continued. “They’re immersing themselves in the AI realm, constructing guardrails that facilitate secure experimentation while delivering user experiences that rival public AI tools. Their emphasis is on UX facets that ensure seamless usage and foster heightened adoption—enterprises must prioritize the innovation impulse over thwarting it.”
Perfecting User Experience and AI Accessibility
“Organizations must establish strategies that blend robust security with enabling employees to effectively utilize AI technologies. Outright bans often drive AI activities underground, amplifying risks,” Arora recommended. CISOs and security leaders are confronted with a quandary: granting AI access to employees, a known productivity enhancer, while safeguarding invaluable intellectual property.
For many CISOs, security leaders, and tech honchos today, employee experience stands at the crux of this dilemma.
Sam Evans, CISO of Clearwater Analytics, faced a pivotal juncture in October 2023. Addressing Clearwater Analytics’ board, he had to address apprehensions about employees inadvertently exposing data that could jeopardize the firm’s $8.8 trillion assets under management. “The most dreadful scenario would involve an employee extracting customer data and feeding it into an AI engine beyond our purview,” Evans shared with VentureBeat. “The employee, unaware or attempting to address a customer issue…that data aids in training the model.”
A Seven-Point Strategy to Halt Shadow AI Before It Imperils Your Organization
Pooling insights from Arora, Golan, and Ivanti’s latest findings unveils a precise roadmap to combat the proliferation of shadow AI while offering the employee experiences that deter its inception:
1. Audit Everything: Map shadow AI to digital friction. Avoid guesswork on the whereabouts of shadow AI—identify it. Employ comprehensive network monitoring and proxy analysis to establish a baseline for unauthorized AI usage and the digital experience gaps propelling it. Organizations triumphing in this battle religiously track DEX metrics akin to monitoring security logs.
2. Centralize AI Governance Under One Roof. Arora’s argument holds weight: fragmented AI oversight spells doom. Erect an Office of Responsible AI with tangible authority—one that oversees both security protocols and user experience enhancements. Half-hearted measures create blind spots that shadow AI readily exploits.
3. Monitor User Pain Points, Not Just Security Threats. Traditional DLP mechanisms weren’t designed to address AI risks. Deploy monitoring tools that capture both text-based AI breaches and the user grievances precipitating workarounds. If employees battle your tools daily, they’re already crafting alternatives out of sight.
4. Build a Living Catalog of Approved AI Tools. Cease playing catch-up with in-house solutions that take months to roll out. Maintain a vetted AI catalog that updates based on real user performance data, not IT’s comfort zone. If your approved tools fail to outshine shadow alternatives in speed and usability, you’ve already lost.
5. Train for Reality, Not Compliance Theater. Generic AI awareness training falls flat. Educate employees on genuine shadow AI risks while furnishing clear pathways to request superior tools. Frustrated users flout policies, but empowered users emerge as your staunchest defense.
6. Make User Experience a Board-Level Risk Metric. Boards are awakening to this reality: poor digital experiences directly correlate with shadow AI adoption. Embed DEX metrics into your GRC dashboards. When user satisfaction dwindles, shadow AI usage typically surges.
7. Deploy Enterprise AI That Truly Works. Cease attempting to replicate what already exists. Enterprise AI solutions evolve more rapidly than any internal team can match. Seek expert assistance for due diligence, opt for peer-vetted tools, and concentrate on solutions that employees genuinely desire to use. Security alone won’t propel adoption.
The Bottom Line: Shadow AI isn’t merely a security quandary; it signifies a failure in user experience. Rectify the employee experience, and you eliminate the impetus for shadow AI before it materializes. Persist in frustrating your users with subpar tools, and they’ll continue crafting superior ones unbeknownst to you. “Every enterprise should treat UI and UX design as a security control,” Arora concluded. “Intuitive AI application design doesn’t solely enhance productivity; it’s the most potent defense against employees resorting to tools that IT can’t discern or safeguard against.”
