AI adoption in Singapore firms slows at execution stage as data gaps persist

AI adoption among businesses in Singapore has moved into daily use, but progress into more advanced applications remains limited, according to new research from HubSpot.

The study, based on responses from more than 700 business leaders, found that 64% of organisations are already applying AI consistently across daily workflows. However, only 18% have implemented fully autonomous AI agents capable of executing tasks and making decisions end to end.

The findings point to a plateau in adoption, where early experimentation has transitioned into operational use, but scaling beyond that introduces new constraints rather than simplifying deployment.

Scaling AI exposes underlying system constraints

The research indicates that organisations with more advanced AI use are encountering deeper structural challenges. Rather than reducing complexity, increased adoption brings greater pressure on data systems, integration layers, and internal capabilities.

Data quality and integration issues were cited by 37% of respondents as a key barrier, second only to trust and reliability concerns at 43%. Among organisations already deploying autonomous AI agents, these constraints become more pronounced. Data integration challenges rise to 41%, legacy system limitations to 42%, and skills gaps to 39%.

This progression suggests that the limiting factor has shifted from access to AI tools to the ability to operationalise them across fragmented systems. AI systems require consistent, connected data inputs to function reliably, and gaps in this foundation restrict their effectiveness as use cases expand.

Megan Hughes, Managing Director and Vice President for JAPAC at HubSpot, said the issue is no longer centred on adoption itself.

“The key challenge among Singapore businesses is no longer whether they are using AI. It is whether they have the knowledge of customers, market trends, and operations needed to scale the business reliably,” she said.

Demand for reliable AI output shapes investment decisions

Despite these constraints, appetite for AI adoption remains strong. More than two in five respondents, or 43%, expect AI agents to become highly important to their operations within the next 12 to 24 months. Only 2% indicated no intention to invest in AI agents.

However, most organisations are still in a transitional phase. While 28% are already investing in AI agents, others remain cautious, waiting for clearer evidence of business impact. Around 30% of respondents said demonstrated results would be the primary factor influencing further investment.

Leaders identified accuracy and reliability as the most critical requirement for AI agents, cited by 66% of respondents. This was followed by system integration at 56%, governance at 53%, and access to relevant business context and data at 48%.

These priorities reinforce the role of data infrastructure in determining whether AI deployments deliver measurable outcomes or remain limited to isolated use cases.

Product updates focus on integrating data and workflows

In response to these constraints, HubSpot is developing an agentic customer platform designed to unify customer data and business knowledge across systems.

At its Spring Spotlight 2026 event, the company introduced a set of updates across marketing, sales, and customer support functions, aimed at embedding AI into operational workflows with access to contextual data.

For marketing teams, tools such as Breeze Assistant and HubSpot AEO are positioned to support campaign development and visibility in AI-driven search environments. Breeze Assistant uses internal data such as customer records, campaign performance, and website analytics to generate campaign briefs and define target profiles. HubSpot AEO focuses on how businesses appear in answer engines, including large language model platforms.

In sales, the Prospecting Agent uses CRM data and intent signals to prioritise leads and automate outreach. Early users reported response rates reaching twice the industry benchmark. Smart Deal Progression complements this by recommending follow-up actions and CRM updates based on customer interactions and historical data.

For customer support, the Customer Agent is designed to handle routine queries while integrating with existing help desk workflows. HubSpot reports that teams using the tool alongside its support platform are resolving 50% more tickets and achieving 29% faster resolution times. The system can resolve up to 70% of conversations on average, with higher rates reported among more advanced users.

These updates reflect a focus on embedding AI within existing operational systems rather than adding standalone tools, with performance tied to how effectively data is integrated and applied across the organisation.