A few months back, I sat down with the team to plan our yearly budgets. Like most businesses, we set aside a portion for R&D, including the need-of-the-hour requirement, “AI”. Just as we started the financial year, the team came back saying we needed to rework our AI budgets. The AI token cost had multiplied 3x over the past two months.
That moment stayed with me. I know that as a CEO, “ambiguity tolerance” is required. As much as I had to drill down to course-correct, it made me realize how much harder AI adoption is for SMEs compared to large enterprises.
In enterprises with dedicated R&D teams, there is room to experiment. They plan for failures and invest in long R&D cycles without expecting immediate returns. AI adoption in SMEs is not just a technology shift; it’s a business survival decision where every investment must prove value quickly. It is about staying relevant in an ever-evolving technology landscape.
We don’t have the luxury of saying, “Let’s try this and see how it goes over the next year.” The expectation is clear: value must be created, ROI must be visible, and it must be visible quickly. Anything long-term becomes difficult to justify. This isn’t unique to us; many SME organizations think the same way.
Enterprises have dedicated IT teams that track technologies, evaluate tools, and drive adoption. SMEs don’t. In many cases, there isn’t even a single person focused purely on emerging technology. There is a constant need for teams handling day-to-day operations to also keep up with technological change.
We tend to operate within our own world, and looking beyond that can initially feel irrelevant. Not everyone has the ability or willingness to step outside their daily responsibilities. There’s a natural tendency to stay within comfort zones and believe that doing what we already do well is enough to sustain growth.
How did we overcome the adoption challenges?
- Mandated upskilling for the leadership team
Ensured decision-makers had a baseline understanding of AI to drive informed choices. - Created a mindset where “failure is an option”
Encouraged experimentation while consciously managing risk. - Made Proof of Concepts (POCs) mandatory
Tested ideas early before committing larger investments. - Instituted weekly/bi-weekly POC reviews
Enabled faster course correction and reduced the cost of prolonged mistakes. - Shifted to an outcome-based approach
Focused on solving business problems instead of building solutions around specific AI technologies.
Our failure points
- Assuming everyone could upskill and perform at the same pace
Not all team members could adapt equally, leading to uneven progress. - Trying to explore every AI technology
The noise around AI led us to experiment with too many tools, increasing costs beyond budget. - Misunderstanding tokenization
Lack of clarity on input vs output token pricing resulted in unexpected cost escalations. - Choosing the wrong pricing models initially
Uncertainty between pay-per-use, subscription, and credit-based models impacted cost efficiency. - Not implementing revenue control mechanisms early
Missing controls like model access restrictions, batch processing, and rate limits led to avoidable cost increases.
Way forward
Our journey made one thing very clear: AI adoption in SMEs is not just a technical challenge; it’s an operational and financial balancing act. Unlike enterprises, we don’t have the buffer to absorb prolonged experimentation or inefficiencies. Every decision must be intentional, measured, and aligned to outcomes.
What helped us move forward was not chasing AI as a trend, but grounding it in business value. By focusing on controlled experimentation, continuous learning, and outcome-driven implementation, we were better able to navigate uncertainty.
AI is powerful, but for SMEs, the real challenge is not access to technology. It is adopting it in a way that is sustainable, cost-effective, and aligned with how we actually operate.
