For many organizations, AI training has become synonymous with productivity. Employees learn how to write better prompts, automate routine tasks and generate content faster than ever before. But as AI becomes embedded in everyday business decisions, a more important question is emerging: Are organizations teaching people how to use AI, or how to use it responsibly?
AI can generate recommendations, summarize information and accelerate workflows, but it cannot assume accountability for outcomes. That responsibility still belongs to people. Yet many training programs spend far more time on tools than on judgment, ethics, governance and critical thinking.
This concern is reflected in Deloitte’s “The State of Generative AI in the Enterprise” research, which found that regulatory compliance concerns, risk management challenges and the lack of governance models rank among the leading barriers to scaling AI initiatives. As organizations move beyond experimentation, the challenge is no longer simply getting employees to use AI—it is ensuring they can use it responsibly.
To explore what modern AI fluency should look like, we turned to members of the Senior Executive AI Think Tank, a curated community of experts in machine learning, generative AI and enterprise transformation. Their perspectives offer a roadmap for moving beyond AI tool proficiency and building the judgment, oversight and responsible-use practices that enable organizations to create lasting value from AI.
Prompt Engineering Is Only the Beginning
Punit Bhatia, Founder of Grow Skills Store, believes organizations still need practical AI skills training, but those efforts should begin with realistic expectations about what large language models can and cannot do.
Grow Skills Store helps organizations build practical skills in privacy, artificial intelligence, information security, risk management and compliance, giving Bhatia a firsthand view of how enterprises are approaching workforce readiness.
“Modern AI fluency needs to rely on skills and training in prompt engineering,” Bhatia says. “Staff need to learn how to use large language models, what to expect from them and what not to expect.”
He notes that this capability is becoming increasingly important among clients seeking to operationalize AI across business functions. Without foundational understanding, employees can easily overestimate the reliability of AI-generated outputs or misunderstand where human review remains essential.
The observation highlights an important distinction in AI education. Teaching people how to generate content is relatively straightforward. Teaching them how to evaluate the quality, reliability and limitations of that content requires a more nuanced approach.
“When people let AI do the thinking first, their brains disengage faster and they lose ownership of their work.”
Keeping Human Thinking in the Loop
Daria Rudnik, Team Architect and Executive Leadership Coach at Daria Rudnik Coaching & Consulting, argues that organizations should be paying much closer attention to the cognitive effects of AI adoption.
Rudnik, author of CLICKING and co-author of The AI Revolution: Thriving Within Civilization’s Next Big Disruption, has spent years helping organizations redesign teams and leadership approaches for an AI-enabled future. Her concern is that employees may unintentionally outsource their thinking to AI systems.
“Modern AI fluency should include judgment, critical thinking and awareness of how AI shapes decisions, language and team dynamics,” Rudnik says.
She points to emerging research suggesting that how people engage with AI matters just as much as whether they use it.
“When people let AI do the thinking first, their brains disengage faster and they lose ownership of their work. But when they think first and use AI to challenge or expand their ideas, judgment stays active.”
The implications extend beyond individual performance. Rudnik notes that AI increasingly influences how teams communicate and solve problems together.
“AI doesn’t just influence individual thinking,” she says. “It also shapes how teams think together.”
As organizations increasingly standardize workflows around AI-generated outputs, leaders face a new challenge: preserving cognitive diversity. Teams may become more efficient by adopting AI-generated language and frameworks, but they also risk converging around similar assumptions and solutions. For Rudnik, effective AI training should therefore emphasize active engagement rather than passive acceptance.
“Leaders need to teach teams not only how to use AI, but how to stay cognitively engaged while using it,” she says. “AI should be a tool that helps us think better, not a replacement for thinking.”
The Competitive Advantage of Responsible Decision-Making
Richie Adetimehin, AI Advisory and Transformation Delivery Consultant at Visani America, sees a troubling imbalance in how organizations define AI readiness.
His work focuses on helping enterprises translate AI investments into measurable business outcomes while maintaining governance, security and compliance. Through large-scale transformation programs, he has observed that many organizations emphasize technical execution while underinvesting in decision quality.
“AI fluency is no longer about knowing how to use the tools,” Adetimehin says. “It’s about knowing when to trust them, when to challenge them and when to override them.”
That distinction becomes increasingly important as AI influences customer experiences, operational decisions and regulatory outcomes.
“Too many organizations focus on prompts and productivity while overlooking judgment, accountability, bias awareness, risk management and decision governance,” he says.
The growing regulatory landscape reinforces his point. New frameworks such as the European Union’s AI Act and emerging governance requirements around the world are placing greater emphasis on transparency, accountability and oversight.
Adetimehin argues that future leaders must understand the broader consequences of AI deployment.
“Modern AI leaders need to understand not just how AI works, but how AI impacts customers, employees, regulations and business outcomes,” he says.
Ultimately, he believes the organizations that succeed with AI will be those that develop stronger human decision-makers rather than simply stronger technical users.
“AI capability creates potential,” Adetimehin says. “Human judgment determines value.”
Building Fluency Around Governance and Accountability
Venkata Kondepati, Manager of Data Architecture and Engineering at Ascentt, believes AI literacy must expand far beyond operational efficiency.
Ascentt helps enterprises create value from data and AI investments through advanced analytics, AI products and enterprise-scale technology solutions. Kondepati’s decades of experience leading cloud, engineering and data modernization initiatives give him a broad perspective on the challenges organizations face as AI becomes more autonomous.
“AI fluency should go far beyond prompt engineering and productivity tools,” Kondepati says.
He argues that responsible AI use requires leaders to understand the factors that influence the quality and trustworthiness of AI outputs.
“Leaders need to develop judgment—the ability to know when to trust AI, when to challenge it and when human expertise must prevail,” he says.
In addition, organizations should educate employees on the operational and ethical foundations that support AI systems.
“It should also include understanding bias, data quality, privacy, security, governance and accountability for AI-driven decisions,” Kondepati says.
He further stresses the importance of systems thinking, particularly as organizations deploy AI across interconnected business processes.
“Equally important is recognizing how AI impacts customers, employees, operations and society,” he says.
For Kondepati, successful organizations will balance speed with responsibility.
“As AI becomes more autonomous, the competitive advantage will not come from using AI faster, but from using it responsibly, transparently and in ways that align with business goals and human values.”
Governance Fluency Is the Missing Skill
Egbert von Frankenberg, CEO of Knightfox App Design Ltd., believes many organizations are solving the wrong problem.
Knightfox helps startups and enterprises implement AI-powered automation, intelligent customer experiences and digital transformation initiatives. Through that work, von Frankenberg has seen a recurring pattern emerge.
“Most organizations treat AI training as a tools problem—learn the interface, boost productivity,” he says. “But the real gap is governance fluency.”
He notes that discussions among AI leaders increasingly point to governance failures rather than technical failures as the primary source of implementation challenges.
“AI failures in production rarely come from technical incompetence,” von Frankenberg says. “They come from misalignment: teams deploying models without understanding what the model can’t do, without audit trails and without clear accountability when outputs go wrong.”
To address those risks, he believes organizations should educate employees on several critical competencies.
“Modern AI fluency must include understanding hallucination risk and when not to trust outputs, knowing your compliance obligations, building human-in-the-loop checkpoints and recognizing that AI cannot fix organizational misalignment—it amplifies it,” he says.
These governance capabilities can determine whether promising AI initiatives successfully scale or stall under scrutiny from legal, compliance and executive stakeholders.
“Leaders who skip this end up with pilots that never reach production because legal, compliance or the board pulls the plug,” von Frankenberg says.
Perhaps most importantly, he rejects the notion that ethics slows innovation.
“Ethics isn’t a constraint on AI adoption,” he says. “It’s the prerequisite for it.”
“The ultimate goal of artificial intelligence is to augment human intelligence—never to entirely replace strategic executive judgment.”
AI Should Augment Executive Judgment, Not Replace It
Uttam Kumar, Engineering Manager at American Eagle Outfitters, believes many organizations remain overly focused on operational efficiency while underestimating the strategic implications of AI-driven decision-making.
Having led large-scale retail technology initiatives across global markets, Kumar has seen firsthand how automation can improve performance while also introducing new risks when business context is overlooked.
“Organizations heavily prioritize tool efficiency, yet they frequently overlook the strategic judgment required to guide autonomous agents safely,” Kumar says.
He argues that AI literacy should include a strong understanding of organizational goals and stakeholder impact, not simply technical proficiency.
“True AI fluency must combine technical capability with sharp business context, enabling teams to evaluate whether an automated recommendation aligns with long-term brand goals,” he says.
That perspective is particularly relevant as organizations deploy increasingly autonomous systems capable of making recommendations that influence pricing, customer engagement, inventory management and resource allocation.
“Without a strong ethical foundation, predictive models may optimize for short-term profit at the direct expense of margin health or merchant relationships,” Kumar says.
For leaders, the lesson is straightforward: AI should support strategic thinking rather than replace it.
“The ultimate goal of artificial intelligence is to augment human intelligence—never to entirely replace strategic executive judgment,” he says. “Fluency means knowing when to override an algorithmic prediction to safeguard your wider enterprise ecosystem.”
Critical Thinking Is the First Line of Defense
Pradeep Kumar Muthukamatchi, Principal Cloud Architect at Microsoft, believes many organizations are investing heavily in AI skills while overlooking the capabilities that matter most when systems fail.
As a cloud and AI strategist who advises startups and enterprises on secure, scalable AI adoption, Muthukamatchi regularly helps organizations navigate the risks and opportunities that accompany rapid deployment.
“Most leaders are missing the bigger picture,” he says. “They train staff on how to type prompts and save hours, but they forget the risky parts.”
For Muthukamatchi, true AI literacy begins with critical thinking.
“AI can make things up or share biased data, so people need to know when to question the machine,” he says.
That skepticism is increasingly important as generative AI becomes integrated into workflows that influence customer interactions, operational decisions and strategic planning.
He also believes organizations must prioritize ethical awareness and information security as core components of AI education.
“Teams have to understand data privacy and how to use these tools responsibly without leaking company secrets,” Muthukamatchi says.
The reason is simple. Productivity gains mean little if organizations scale poor decisions.
“A fast worker who copies bad data just makes mistakes quicker,” he says. “Real AI fluency means knowing how the technology works, seeing its flaws and making smart choices about when to trust it.”
“While healthcare organizations often emphasize AI training for productivity and tool proficiency, leaders must also prioritize judgment, ethics and responsible use.”
Trust Depends on Responsible Use
Will Conaway, President of Tuxedo Cat Consulting, approaches AI fluency through the lens of healthcare, an industry where trust, transparency and ethical decision-making carry especially high stakes.
Conaway, an award-winning healthcare executive, educator and AI strategist, believes organizations cannot separate technical AI training from broader responsibilities around accountability and human oversight.
“While healthcare organizations often emphasize AI training for productivity and tool proficiency, leaders must also prioritize judgment, ethics and responsible use,” he says.
The importance of those capabilities is reinforced by growing concerns about bias and transparency in AI-enabled healthcare systems. The World Health Organization’s guidance on artificial intelligence for health emphasizes the need for transparency, accountability and human oversight as AI becomes more integrated into clinical environments.
Conaway argues that AI literacy should prepare employees to evaluate outputs rather than simply accept them.
“Modern AI fluency should include understanding data bias, patient privacy, transparency in decision-making and the ability to question AI recommendations,” he says.
Those competencies are critical not only for healthcare but for any industry where AI influences outcomes affecting people, trust and reputation.
“This broader skill set ensures that AI supports, not overrides, human expertise,” Conaway says. “Ultimately, well-rounded AI fluency empowers clinicians to use these technologies safely and thoughtfully while minimizing unintended harm.”
Judgment Has Become a Governance Capability
Aditya Vikram Kashyap, Vice President of Firmwide Innovation at Morgan Stanley, believes organizations are defining AI literacy too narrowly.
A recognized innovation leader whose work spans AI governance, enterprise transformation and responsible technology adoption, Kashyap argues that technical proficiency is rapidly becoming table stakes.
“We’ve made AI literacy synonymous with tool proficiency when it should be synonymous with decision quality,” he says.
As AI becomes embedded in decisions affecting customers, employees, financial markets and public trust, organizations must focus on strengthening the human capabilities that govern those decisions.
“Knowing how to use AI is rapidly becoming a baseline skill,” Kashyap says. “Knowing when not to trust it will be the competitive advantage.”
He identifies several competencies that should sit alongside technical training.
“Modern AI fluency must extend beyond prompts and productivity to include critical thinking, bias recognition, uncertainty assessment, data provenance and accountability for outcomes,” he says.
For Kashyap, successful organizations will institutionalize challenge and oversight rather than blindly pursuing speed.
“The organizations that create lasting value from AI will not be those that deploy it the fastest, but those that build the strongest culture of challenge, oversight and responsible use around it,” he says. “In the age of AI, judgment is not a soft skill; it is a governance capability.”
Create Deliberate Pause Points
Divya Parekh, Founder of executive coaching brand DivyaParekh.com, believes organizations are overlooking one of the most important aspects of AI readiness: teaching people when to stop and think.
As a Thinkers50-recognized leadership coach and AI adoption advisor, Parekh works with executives to build AI-ready performance systems while preserving the human capabilities that drive sound decision-making.
“Leaders are not paying enough attention to the judgment layer of AI yet,” she says.
While prompt engineering and productivity training remain important, Parekh believes they represent only the starting point.
“Modern AI fluency should teach people how to question outputs, protect confidential information, recognize bias, understand context, check the quality of the data behind the answer and know when human judgment must lead,” she says.
Perhaps most importantly, she advocates building structured checkpoints into AI-enabled workflows.
“Teams need clear pause points where they stop, verify, challenge or override AI before a decision moves forward,” Parekh says.
The rationale is straightforward. AI amplifies existing strengths and weaknesses alike.
“AI does not just speed up work,” she says. “It speeds up the thinking behind the work.”
That reality places even greater importance on developing critical thinking and accountability across the workforce.
“If that thinking is clear, AI can strengthen execution,” Parekh says. “If it is flawed, AI can scale the blind spots faster than leaders realize.”
Building Responsible AI Fluency
- Teach employees what AI can and cannot do. Practical prompt engineering should be paired with realistic expectations about model limitations and appropriate use cases.
- Keep human thinking active. Encourage employees to develop ideas before consulting AI so the technology expands thinking rather than replacing it.
- Train decision-makers, not just tool users. Build capabilities around accountability, bias recognition, governance and risk management.
- Expand AI literacy beyond productivity. Include data quality, privacy, security and systems thinking in training programs.
- Develop governance fluency. Teach teams about hallucinations, compliance obligations, auditability and human-in-the-loop controls.
- Connect AI outputs to business context. Ensure employees understand how recommendations align with long-term organizational goals and stakeholder interests.
- Normalize healthy skepticism. Train staff to question outputs, identify inaccuracies and protect sensitive information.
- Protect trust through responsible use. Emphasize transparency, privacy and oversight in every AI-enabled process.
- Treat judgment as a governance capability. Reward employees who challenge assumptions and evaluate uncertainty rather than simply accepting AI recommendations.
- Build deliberate pause points. Create structured opportunities to verify, challenge or override AI outputs before important decisions move forward.
Why Responsible AI Starts With People
The conversation around AI skills has largely focused on what the technology can do. But the experts featured here suggest the more important question is whether people know how to use AI responsibly. As AI becomes embedded in decisions affecting customers, employees, patients and businesses, technical proficiency alone is no longer enough. Prompt engineering and productivity gains matter, but so do critical thinking, accountability, ethical reasoning and the ability to recognize when AI is wrong.
What stood out across these conversations was the consistency of the message: the organizations that create lasting value from AI will be those that develop stronger decision-makers, not just more efficient AI users. As artificial intelligence becomes more capable, human judgment is becoming more valuable, not less. In the end, AI may accelerate decisions, but people remain responsible for making the right ones.
