According to McKinsey & Company, more than 70% of companies are using AI in at least one business function. International Data Corporation forecasts global spending on AI technologies will exceed US$500 billion by 2027. Across sectors including finance, logistics, agriculture and healthcare, AI is increasingly being integrated into operational workflows.

Staffs discussing on using AI. Photo courtesy of Groove Technology

However, as adoption accelerates, many organizations face challenges in translating AI capabilities into measurable business outcomes. While the technology continues to advance, effective implementation requires more than deploying tools or adding standalone AI features.

Market pressure is a key driver. Companies are seeking to demonstrate technological readiness, respond to competitors and signal innovation to stakeholders, leading some to adopt AI without clearly defined use cases.

Matt Long, CEO of Groove Technology, said businesses often approach AI from a technology-first perspective rather than focusing on operational needs. Without a clear problem to solve, AI initiatives may become disconnected from day-to-day workflows and fail to deliver tangible results.

Matt Long (2nd, R), CEO of Groove Technology, said businesses often approach AI from a technology-first perspective. Photo courtesy of Groove Technology

In response, organizations are increasingly focusing on identifying specific operational bottlenecks where AI can improve efficiency or decision-making. Aligning AI with clearly defined workflows allows companies to better measure its impact.

Demand for AI integration is also evolving. Businesses are moving beyond basic applications such as chatbots toward more advanced use cases, including workflow automation, continuous data analysis and decision support systems.

Hung Do, business development manager at Groove Technology, said companies are seeking solutions that support real operational tasks rather than standalone features.

However, implementation often exposes gaps between expectations and operational readiness.

In many cases, organizations can define desired outcomes, such as automated reporting or predictive insights, but lack structured processes to support them. Incomplete workflows, unclear responsibilities and fragmented data systems can limit the effectiveness of AI applications.

Data readiness remains a critical issue. AI systems depend on structured, consistent and high-quality data, yet many organizations still manage data across multiple platforms and formats. Preparing workflows and data infrastructure often requires more effort than deploying the AI tools themselves.

Industry experts note that the main challenges in AI adoption are not purely technical but operational. The effectiveness of AI depends on how well it is integrated into existing processes, decision-making structures and data flows.

Mai Nguyen, general director at Groove Technology, said organizations can build AI-enabled systems relatively quickly, but ensuring long-term reliability, security and scalability requires sustained effort. In practice, teams often need to continuously adjust processes, refine inputs and monitor system performance after deployment.

Mai Nguyen, general director at Groove Technology, in a meeting with staffs. Photo courtesy of Groove Technology

Unlike traditional software, AI applications typically deliver value gradually through ongoing testing and refinement rather than immediate results, which can affect expectations and slow adoption.

As a result, many companies are adopting a "pilot-first" approach, testing AI in controlled environments before scaling. This allows organizations to evaluate performance using real data and workflows, identify limitations early and build the operational foundation needed for broader implementation.

At Groove Technology, this approach is often applied through structured pilot projects focused on specific use cases such as workflow automation or task-level optimization, helping businesses assess whether AI can deliver measurable value before committing to larger deployments. As adoption matures, companies are increasingly recognizing that success depends on aligning technology with operational needs, including clearly defined workflows, reliable data systems and a gradual approach to integration.

Rather than prioritizing speed of deployment, organizations are focusing on how AI can be applied effectively within existing operations. In this context, AI is becoming part of a broader operational transformation, requiring adjustments in processes, data management and decision-making.

The long-term success of AI adoption is likely to depend less on how quickly companies implement new technologies and more on how effectively they integrate them into practical business use.