Artificial Intelligence (AI) headlines often warn of a future where humans outsource cognition to algorithms, eroding our intellectual capabilities.

The reality, however, appears to be just the opposite. Emerging research and real-world implementations show that when properly integrated into workflows, AI doesn’t diminish human cognitive capabilities — it enhances them. By handling routine tasks and information processing, AI allows knowledge workers to engage with higher-order thinking, develop new meta-skills, and focus on uniquely human contributions. 

With AI, the cognitive challenge for humans has shifted from accessing information to effectively filtering, evaluating, and directing technology while preserving human judgment and critical thinking. Rather than diminishing our capabilities, AI-powered tools might even help extend our creativity into entirely new domains previously beyond our reach.

Today, you can push a button and the data just comes at you.

_{Leah McGowen-Hare, SVP of Trailhead Community, Salesforce}

This cognitive shift represents a fundamental change in how we process information, according to Leah McGowen-Hare, SVP of Trailhead Community at Salesforce. While earlier generations had to learn systems like the Dewey decimal classification just to access information, the issue now is having too much. 

“Today, you can push a button and the data just comes at you,” she said. “The skills that my kids are going to need are filtering the data, critically thinking, and looking at that data.”

Research backs up this collaborative relationship. A 2024 study on collaborative intelligence from Slack, which is owned by Salesforce, found AI-powered tools save users 97 minutes per week‌ on average — ‌time that can be redirected to deeper thinking and creative problem-solving. 

“Artificial intelligence doesn’t have to be a threat to employees; it can be an opportunity,” the research report noted. “AI tools pick up the more mundane, repetitive tasks, freeing up team members to focus on higher-value jobs.” 

McGowen-Hare believes that companies will see the greatest performance improvements when they view AI agents as teammates with complementary strengths rather than as a replacement for human judgment.

“Working with an agent is like a skilled tennis player and their racket — when they’re in the zone, the racket becomes an extension of their arm,” she said. “It’s no longer a separate tool; it’s part of their capability. In the same way, AI agents can become seamless extensions of our professional abilities, helping us move faster and work smarter.”

AI as a partner in strategic decision-making

The positive effect of AI on critical thinking extends beyond basic information processing to more complex cognitive tasks like strategic decision-making. Some worry this represents a dangerous new frontier, but history suggests otherwise.

“AI is not that new,” said Felipe Csaszar, Professor of Strategy at the University of Michigan’s Ross School of Business. “Everything that a computer can do, at some point, was called AI. We have a lot of experience with outsourcing parts of our thinking to external devices.”

And, just like these tools that we once called AI became helpful to us but didn’t replace us, today’s AI will, too. Csaszar sees it as the next logical step in our technological progression. “I’m hopeful that this will generate opportunities for us to do more thinking at the higher level,” he said.

His research puts this hypothesis to the test. In one experiment, AI-generated business plans competed against human-created plans in a business competition. The AI plans ranked 7% higher, according to human judges. In another study, AI evaluations of business plans closely matched those of human experts.

“What will happen is not that people will just let AIs build their company’s business plan and forget about it,” Csaszar said. “You will ask AI to generate multiple plans, and you will pick out of those, or you will come up with an idea and the AI will help you improve it.”

Agentforce, Salesforce’s solution for deploying AI agents in business workflows, is an example of this at work. A key feature of Agentforce is its capacity for autonomous operation within specifically defined guardrails. This means that while ‌AI agents can operate independently, making decisions and taking actions, they do so within boundaries established by human teams, ensuring alignment with business objectives and policies.

For example, Agentforce can autonomously segment customers based on various criteria that human teams develop, enabling businesses to define their target market more precisely. Then, it can present a plan with multiple marketing strategies, from which humans can choose the best using their judgment.

This combination of autonomous operation and customizable guardrails provides a framework for AI to contribute to problem-solving and process execution while remaining within acceptable limits, allowing humans to focus on strategic direction and higher-level innovation rather than micromanaging every AI action.

This approach creates a powerful synergy: Humans bring judgment, domain expertise, and contextual understanding, while AI contributes computational power, pattern recognition, and the ability to rapidly generate alternatives. The resulting collaboration unlocks efficiency and effectiveness beyond what either could achieve alone.

“Our day will continue to have 24 hours, but we will be able to do more things in those 24 hours,” Csaszar said. “Those who take advantage of AI will be the ones that will be more successful.”

This partnership approach is already transforming industries, like healthcare company Precina, which is revolutionizing diabetes management with an AI-powered approach to patient care. 

Agentforce lets us identify the most important interventions and bring them forward to patients.

_{John Oberg, CEO, Precina}

“Agentforce lets us identify the most important interventions and bring them forward to patients,” said Dr. Oberg, CEO of Precina. “This ensures patients receive the right care at the right time to improve outcomes and enhance the patient experience.”

By integrating patient data and automating routine tasks, Precina’s team can make more strategic decisions about patient care, focusing their critical thinking on complex cases rather than administrative procedures. The company estimates Agentforce will save between $13,000-$25,000 in training costs per clinical provider annually while simultaneously improving patient outcomes.

“In my lifetime, I want to help a billion people improve their health and have a better life because of the work we’re doing,” said Oberg. “With Agentforce, that dream of helping a billion people is a reality.”

Enhancing human creativity

The creative process offers another window into how AI can amplify rather than reduce human cognitive capabilities.

“Humans don’t do creativity out of nothing,” said Zorana Pringle, Senior Research Scientist at Yale and author of the forthcoming book, The Creativity Choice. “You are using your experiences. You are recombining them. You are connecting things that otherwise would not be connected.”

This perspective challenges fears about AI replacing human creativity. “When ChatGPT came on board, there was this flurry of activity like, ‘Is it more creative than humans?’ Which I think is a little bit of a wrong question,” Pringle said. “It’s a different thing.”

The fear that AI will diminish human creativity parallels historical anxieties about other technological innovations. “When photography was first invented, it was talked about as this is going to be the end of painting,” Pringle said. “It wasn’t the end of painting, so it’s not going to be the end of anything else.”

When researchers and professionals collaborate with AI systems, they often find their own thinking enhanced. Pringle describes using AI to analyze open-ended research responses: “We have sped up the process we have to do as scientists. What would take us a month to do becomes a week.” But ‌critical thinking remains with ‌humans, who evaluate the AI’s output, identify what might be missing, and add their insights.

“Sometimes humans hit the creative block,” Pringle said. “In those times, now we have a new tool. Sometimes it’s that interplay. We are responding to the output, and we merge human critical thinking and the machine output.”

Even when creators don’t directly use what the AI produces, the interaction itself can be transformative, Pringle said. 

“There have been really cool cases with artists and writers using these models, where they’re essentially having a conversation with a machine, and they don’t end up using a single word of what it produces, but it’s still inspirational,” she said.

Salesforce’s ​​Draft with Einstein, a Salesforce AI feature that accelerates email creation by giving reps a strong foundation for every email, is an example of this in action. 

It helps our reps be more creative. The AI generates multiple ways to phrase repetitive emails, enhancing our outreach.

_{Marissa Scalercio, VP of Sales Operations, Carnegie Learning}

“[It] helps our reps be more creative,” said Marissa Scalercio, VP of Sales Operations, Carnegie Learning, in a recent case study on the technology. “The AI generates multiple ways to phrase repetitive emails, enhancing our outreach.”

Salesforce’s developer hackathons are another study in creative thinking with agentic AI. At recent events, including World Tour NYC and TDX 2025, Salesforce asked developers to think creatively about problems that needed solving — and how Agentforce could fix them.

Using the low-code, AI-powered Agent Builder, hundreds of developers crafted custom agents and actions, directly addressing defined business challenges and showcasing the platform’s ability to rapidly deploy AI-powered workflows — all in just 36 hours.

Developers built agents that could help people with retirement, connect first responders in emergency situations, event planning, and more, showcasing the creative thinking that AI can unlock.

The jagged frontier: Navigating AI’s uneven capabilities

As AI capabilities continue to rapidly evolve, our understanding of how these tools affect critical thinking will need to evolve alongside them. What seems impossible for AI today might become routine tomorrow, while new limitations may emerge in unexpected areas. Staying ahead of these changes requires both vigilance and engagement.

One crucial insight from Harvard Business School (HBS) research that helps us navigate this shifting terrain is what experts call the “jagged frontier” of AI capabilities. This frontier isn’t smooth or predictable — AI excels at some complex tasks while struggling with seemingly simpler ones.

In one striking HBS experiment, consultants using AI for creative tasks and strategic analysis saw dramatic performance improvements. However, when the same professionals used AI for tasks involving nuanced data interpretation, requiring integration of quantitative data with qualitative interviews, AI actually decreased performance by 19 percentage points.

To that end, the company has rolled out several Trailhead learning modules, new Agentblazer Statuses, and the Agentforce Specialist certification to give Agentblazers practical tools and hands-on opportunities to build and deploy agents themselves. 

“Tomorrow’s jobs belong to today’s learners,” McGowen-Hare said. “It’s not AI that’s going to replace jobs — it’s the people who learn how to use these AI tools.”

Blazing a new trail

The evidence suggests that AI isn’t reducing our cognitive abilities — it’s changing the definition of what it means to be smart. The cognitive skills that matter most are evolving, shifting from information retention to information evaluation, from isolated problem-solving to collaborative intelligence.

“Thirty years ago, we could have never predicted humans would create some of the technology we rely on today,” said McGowen-Hare. “With humans and agents working together, each using their skills to advance society, the possibilities are endless.”

Go deeper:

• Find out if you’re an Agentblazer, then join the Agentblazer community
• Pursue Salesforce Agentblazer Status to validate your skills

Blog Article: Here