.svg)
AI adoption in enterprises is facing heightened scrutiny as early excitement fades, with boardrooms demanding measurable ROI from pilots and forcing executives to confront the gap between hype and real business value.
Prafulla Patil, CEO of OneSolve, shared his problem-first approach, focusing on user pain points and emphasizing the importance of starting AI projects with clear goals and measurable outcomes.
Leadership engagement and a security-first approach are crucial for successful enterprise AI implementation.
Patil advised leveraging existing talent and starting small to achieve measurable impact without waiting for the perfect team of AI experts.
AI adoption in enterprises is no longer about novelty. Stakeholders demand results. Initial pilots dazzled teams, but a large portion of the initial enthusiasm has given way to intense ROI scrutiny as boardrooms worldwide are confronting the costly reality of AI pilots that fail to deliver measurable value. The gap between expectations and reality has also handicapped the 'all gas, no brakes' mentality often exhibited by enterprise AI adoption, as harder questions like "What's the business value?" and "What outcomes can we prove?" start to become the prevailing voices of reason.
We spoke with Prafulla Patil, the Founder and CEO of OneSolve, and former CIO/CISO at Mapbox, with leadership roles at LinkedIn, Atlassian, and Flexport. Having overseen enterprise AI and automation at multiple Fortune 500 firms, Patil knows where projects stall and how to get them unstuck. Patil argued that as the initial AI fatigue sets in from failed pilots, a structured playbook is emerging for enterprises ready to turn pilots into real results, starting with one simple principle: define your goal.
Start with the goal: Patil said that every AI initiative should begin with the end in mind. "You can start backwards: define your goal and your KPIs, and exactly what you are trying to do." Rather than focusing on the latest tools or technologies, organizations should orient their efforts around measurable outcomes, ensuring that every AI investment drives real business value.
Focus on the problem: According to Patil, successful AI projects start with understanding the problem, not the solution or tool. “Talk to your users as often as you can: shadow them, survey them, and ask what tasks they perform mechanically or repetitively,” he said. Leaders and frontline staff often see workflows differently, so gathering input from both groups ensures that automation addresses meaningful pain points rather than perceived priorities.
Patil warned that without a problem-first approach, enterprises risk investing in flashy pilots that impress on the surface but fail to move the needle. Focusing on real challenges ensures AI adoption delivers tangible outcomes rather than empty buzz. He brought this point to life with a clear example of the approach in practice.
A playbook in practice: Patil illustrated his approach with a real-world example: One customer’s CSM team, which handled roughly 2,000 inquiries per month, aimed to automate responses and reduce the manual workload. In his words, "By starting with a clear goal and auditing their data and processes, they achieved a 54% deflection rate—more than half of the questions were resolved automatically." This case underscores the value of aligning objectives, foundational processes, and data before scaling AI initiatives.
While a clear, measurable goal is the starting point, Patil insisted that the real differentiator between the enterprises that are winning with AI and those stuck in pilot purgatory is a combination of leadership and a security-first mindset.
The leadership mandate: Patil emphasized that successful AI adoption demands executive engagement, alongside technical implementation requirements. "Leaders need to get involved; it can’t just be a mandate handed down to a team," he explained, "Leadership must flow both ways—executives set direction from the top down, while teams provide input from the bottom up. When these perspectives align, AI initiatives gain momentum and deliver meaningful results."
Practicing what you preach: Patil said that shaping an organization’s AI culture begins at the top. "It’s the executive mindset that matters," he said. In one example, a CEO actively participated in Slack channels, helping guide strategy, evaluate approaches, and discuss tool options with the team. "Leadership," Patil explained, "can't be limited to issuing mandates—it requires daily engagement and hands-on involvement."
A security reckoning: "AI adoption without security is a ticking time bomb," Patil said. He called the current enterprise AI landscape a 'wild west' with sensitive data often flowing unchecked into AI systems. Organizations must adopt zero-trust principles, enforce role-based access, and maintain compliance to safeguard against breaches and protect both data and trust.
Patil's advice for enterprise executives is simple: Start small, but start now and allow measurable impact—not hype—guide broader adoption. He encourages enterprises to focus on one use case, leverage existing talent, and experiment without waiting for over-qualified AI experts. "There are no AI strategists with 10 years of experience today. You can start with the team you have, clearly define your workflows, build a data foundation, and iterate. The important advice is just get started today."
