As artificial intelligence becomes a bigger part of daily business operations, leaders are running into real hurdles. For some, the hardest part has been setting clear goals and resisting the urge to lean too heavily on automation. Others have struggled with earning trust, protecting creativity, or keeping data accurate.
The good news is these obstacles don’t have to hold teams back. Below, Fast Company Executive Board members share the challenges they faced and the solutions that helped them push forward.
1. DEFINE PURPOSE BEFORE ADOPTING
The biggest challenge has been resisting the urge to adopt AI without a clear purpose. We formed an internal steering committee to guide responsible use and ensure projects deliver measurable value. Externally, we implemented a customer AI advisory board beta to test our products and provide feedback to help us build meaningful solutions. – Haywood Marsh, Cardata
Subscribe to the Daily newsletter.Fast Company’s trending stories delivered to you every day
Privacy Policy
|
Fast Company Newsletters
2. MAINTAIN QUALITY WHILE MOVING FAST
One of the biggest AI challenges in software development is balancing speed with quality. AI can accelerate prototyping and cut repetitive work, but its real value is being a reliable partner. That means reducing errors, limiting context switching, handling complex tasks, and helping developers deliver higher-quality code. In real-world software, intelligence, trust, and code quality matter most. – Kirill Skrygan, JetBrains
3. DRIVE TEAMWIDE ADOPTION
Encouraging AI adoption among employees can be a major hurdle, and I saw that firsthand at a previous company. The team was hesitant about AI. To overcome this, we built a peer-led “AI University,” where colleagues taught colleagues and shared role-specific use cases. It was safe, practical, and fast to try. Engagement rose, confidence grew, and work kept moving smoothly. – Trent Cotton, iCIMS
4. RISE ABOVE GENERIC OUTPUTS
Our AI challenge stood out amid generic outputs. We solved this by using AI to automate back-end tasks, speeding up data access and freeing resources to deepen pricing expertise. By clearly defining AI’s role, we kept workflows efficient and preserved the fundamentals that drive real results. – Avy Punwasee, Revenue Management Labs
5. BUILD RELIABLE MULTI-AGENT SYSTEMS
Our biggest AI challenge? Tackling seamless info exchange in a complex multi-agent setup with APIs and MCP. The magic happened with a human-in-the-loop approach. By iteratively optimizing agent interactions until flawless, our automated workflows kept things running smoothly—it sounds easy, but it’s not. Developing a multi-agent system that is reliable and stable needs a lot of time and patience. – Yusuf Sar, Hardwarewartung 24 GmbH
6. INTEGRATE AUTOMATION SMOOTHLY
A key challenge I’ve faced with AI was integrating new automations into financial software systems without disrupting operations. The solution was to use support tickets as much as possible, and to be very specific with the software support on what did and did not work. In doing so, I was able to integrate automations that could be updated or swapped without affecting operations. – Chalmers Brown, Due
7. ENCOURAGE MINDFUL DAILY USE
The challenge is making sure talents mindfully incorporate AI into their daily workflow. Our solution has been holding weekly office hours, to which around 10% of our talents attend. It is a great, unstructured way to share successes and crowdsource problems, and it also demonstrates that we encourage curiosity, risk-taking, and creativity. – Erin Fuller, MCI
8. BRING YOUR OWN CREATIVITY
A top challenge with AI as a whole is that it can’t replace creativity. The solution to using AI successfully is to view it as a way to elevate your own creative ideas instead of having it come up with them. I think it’s important for leaders to entrust their team members with using AI as a helpful assistant because it promotes greater efficiency, but they shouldn’t allow it to replace creativity. – John Hall, Calendar
9. BUILD TRUST IN PREDICTIONS
One of the biggest challenges we’ve overcome is helping trial sponsors trust behavioral predictions when they don’t match conventional assumptions. Our solution was to ensure every insight generated by our AI solutions is grounded in validated psychological inputs and clearly explainable. Transparency builds trust not just in the technology but in the decisions research teams make based on it. – Krinx Kong, Cognivia
10. FIX AI HALLUCINATIONS
Hallucination is a major problem. We use GPTs for onboarding, but often need to make updates to our processes and tools. GPT consistently uses the initial doc it was built with, instead of reviewing the new one. To solve it, I create new GPTs. They’re so fast and easy that it’s still a time-saver over pre-AI solutions. The key is to remain flexible and continue to test and iterate. – Andrea Lechner-Becker, GNW Consulting
advertisement
11. DEDICATE LEADERSHIP TO AI
Our biggest AI challenge—which we turned into an opportunity—was ensuring dedicated leadership to drive innovation without losing focus on core operations. The solution was to have David transition from CEO to chief AI officer. His deep knowledge of our product, operations, and customer needs made him ideal to prioritize high-impact AI use cases while our new CEO focuses on business growth. – Dan Amzallag, Ivalua
12. CLARIFY OWNERSHIP ROLES
Ownership is a big challenge. AI lives on so many levels—data, tech, process, people, strategy. It’s easy for confusion to develop about ownership rights and accountabilities. And we see symptoms develop as a result, like skepticism around model accuracy and business value. Be clear about ownership stakes, keep it close to the work being performed, and promote collaboration across owners. – Dan Priest, PwC
13. OVERCOME FEAR OF AI
The biggest challenge is fear—people not using AI tools or keeping their usage to themselves, worried it makes them look less capable. We’ve made great strides in evolving that mindset by creating psychological safety: normalizing transparency, encouraging experimentation, and framing AI as a creative multiplier, not a shortcut. When trust is built, confidence follows. – Jessica Shapiro, LiveRamp
14. USE AI WITH INTENTION
The biggest AI challenge I overcame was breaking my own dependence on it. As a technologist, I initially leaned too hard on automation. I stepped back, got intentional, and started using AI as a strategic tool—not a reflex. Now I guide others to do the same. Mastery isn’t about using AI constantly—it’s about using it deliberately. – Alex Goryachev
15. END THE ‘SWIVEL CHAIR EFFECT’
The biggest AI hurdle we tackled was the “swivel chair effect”—IT teams flipping between tools to get full visibility. We solved it by embedding conversational AI into our unified observability platform, eliminating tool-switching and enabling predictive insights to keep operations smooth. – Christina Kosmowski, LogicMonitor
16. FOCUS BEYOND THE HYPE
The biggest AI challenge was moving leaders past the hype to focus on clear business outcomes. The solution was a cross-functional strategy that tied AI directly to prioritized complex problems beyond efficiency measures, addressed bias early, and delivered quick wins to build trust and momentum. – Patti Fletcher, PSDNetwork, LLC
17. PRESERVE HUMAN CONNECTION
We integrated automation into client communications without sacrificing our human-centered approach. Another hurdle was aligning our team’s workflow effectively with AI tools, which we overcame by setting clear boundaries—automating routine tasks while training our team to keep sensitive interactions personalized. This preserved trust, boosted efficiency, and ensured seamless AI adoption. – Muhammed Uzum, Grape Law Firm PLLC
18. TREAT AI AS STRATEGY
The biggest challenge is helping companies stop seeing AI as a magic button. Buying a tool won’t fix everything. Many teams jump in without a plan and get stuck. At our company, we slow down and ask why AI, what problem it solves, and how to prepare people. Treat AI as a strategy, not a quick fix, and you’ll get results that matter. – Debra Andrews, Marketri LLC
19. BALANCE SPEED AND OVERSIGHT
One big AI challenge was balancing automation with human oversight. Over-reliance risked errors slipping through, while too much control slowed output. The solution was a hybrid workflow—AI for speed, humans for judgment. This kept efficiency high while ensuring quality and trust stayed intact. – Boris Dzhingarov, ESBO ltd
20. CENTRALIZE FRAGMENTED DATA
The management of fragmented data across platforms was one of the biggest AI hurdles we conquered. It significantly increased the difficulty of campaign optimization and customization. The solution was creating a centralized, AI-powered insights hub that could extract, clean, and analyze data instantly. This kept campaigns flexible and expedited decision-making. – Christena Garduno, Media culture