Want Employees To Adopt AI? Give Them Time To Learn
Neal Bradbury is Chief Product Officer at Barracuda.
gettyThe biggest obstacle to AI adoption isn’t access to the technology. It’s finding the time to learn how to use it effectively. Give an engineer an AI coding assistant without changing anything else about their workday, and experimentation has to compete with sprint commitments, meetings and deadlines. Organizations should create dedicated time and space for employees to explore AI in a practical, low-pressure environment. Whether through hackathons, innovation sprints, workshops or focused learning days, these initiatives help employees build confidence and discover how AI can enhance their daily work. Of course, setting aside time doesn’t automatically translate into meaningful AI adoption. Real adoption happens when experimentation is structured around meaningful business problems, supported by leadership and designed to turn individual discoveries into organizational capabilities. Based on lessons learned from a two-day AI hackathon, here are several insights organizations can use to accelerate AI adoption and turn curiosity into impact.
Leaders shouldn’t just offer an open-ended mandate to “build something with AI.” The first step is giving employees the right problems to solve. Experimentation should be focused on a small number of business-relevant themes. Some of the highest-value AI applications are also surprisingly mundane. They are often repetitive processes people already understand but dislike doing manually. That familiarity is an advantage because employees already know the baseline and can quickly determine whether AI improves it.For our hackathon, teams focused on one of two challenges: delivering measurable business impact in the near term or radically increasing velocity by compressing the entire path from idea to customer value. One track centered on solutions that could drive revenue, retention or cross-product value within the year and the other challenged teams to rethink software delivery itself, with a goal of getting innovations from an engineer’s hands to a customer’s hands in less than an hour.The velocity theme wasn’t just about writing code faster. It encouraged teams to examine every stage of the delivery lifecycle—from ideation and design to testing, deployment and operational handoff—and identify where AI could eliminate friction and accelerate outcomes.Those themes provided direction, but teams still needed to decide which specific problems were worth spending their limited time on. That’s why it helps for teams to submit concepts and leadership to evaluate ideas and select the most meaningful projects before the event. Protected time is valuable, so make sure people spend it solving problems that matter.
Nothing drains enthusiasm from a hackathon faster than spending the first several hours requesting permissions and configuring tools.After setting up our first hackathon, we found that friction was the most common complaint. We made sure the infrastructure was ready before the next one began. This time, API access, MCP servers, integrations, licenses, tokens and model capacity were addressed before the build clock started.Standardizing access had another unexpected benefit: Teams naturally began converging around common toolchains. That made it easier for participants to help each other during the event and to support successful approaches afterward.The same preparation should extend to security as well. Teams were given sandbox environments with synthetic rather than customer data, least-privilege access to cross-domain resources and clear rules for reviewing AI-generated work. Sensitive information stayed out of prompts and demos, and humans remained responsible for approving AI-generated changes.
The most important factor for success was protecting participants’ time during the hackathon. We gave teams two full build days and formally took normal sprint work off their plates. If experimentation has to compete with sprint commitments and deadlines, employees will naturally fall back on the workflows they already know. But blocking two days off on their calendar isn’t enough. Executive sponsorship, centrally issued calendar blocks and manager buy-in make it clear that participants aren’t expected to keep one eye on their normal workload. Mentor office hours and dedicated help channels also prevented protected time from being lost to troubleshooting.
If protected time creates the opportunity to learn, collaboration determines how quickly that learning spreads. We found mob programming particularly effective because participants could see how their colleagues worked with AI in real time. That means a technique discovered by one person could spread across a team within minutes.There’s also value in making those teams cross-functional. Engineering doesn’t always see the same problems as support or sales, so bringing those perspectives together surfaces issues that a single team might overlook. Those relationships and shared techniques can then outlast the event, turning individual experimentation into organizational knowledge.
If you reward flashy demos, you’ll get flashy demos. If you reward measurable business outcomes, teams will build toward them.For our hackathon, impact represented 35% of the total score, alongside criteria such as delivery, differentiation and theme-specific metrics. High-impact scores required a measured before-and-after result. “We think this could save time” was not enough.That produced concrete operational results. One team reduced the time from completed code to customer delivery to under an hour, demonstrating AI’s potential to dramatically accelerate innovation and time to value. Another cut detection-efficacy rule creation from three weeks to less than 24 hours, while a third reduced an eight-hour video production workflow down to 30 minutes.There is no question that teams optimize for what leaders measure.
A hackathon shouldn’t end when the demos do. Useful agents, integrations, skills and prompt patterns should go into a shared playbook so the next team doesn’t have to rediscover them. Reusable contributions can even become part of the scoring criteria.The best projects also need a path forward. Nothing undermines enthusiasm for the next hackathon faster than watching the best ideas from the last one disappear into a backlog. But the interesting projects aren’t the only win. The bigger measure of success is whether people return to their regular work with the confidence and habits to solve problems differently.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
