Fragmented tech hinders Australia's AI agent gains

Twilio has published a poll suggesting fragmented technology stacks are hindering Australian marketing and customer experience leaders' use of AI agents. According to the survey, 85% of respondents linked tech fragmentation to difficulties managing AI agent productivity.

The poll surveyed 93 contact centre, customer experience and marketing practitioners attending Customer Contact Week Australia and New Zealand in Sydney. It points to a broader concern among organisations that have begun deploying AI tools but are still struggling to turn them into measurable operational gains.

AI agent productivity emerged as one of the strongest themes. Almost all respondents, 94%, said their organisation would achieve greater financial gains if productivity improved, while nearly half said the potential benefit could be 50% or more.

Respondents also identified broader business impacts from better performance. Two-thirds said stronger AI agent productivity would support faster innovation, and just over half said it would improve customer satisfaction scores.

At the same time, confidence remained mixed in some areas. Most respondents did not expect security to improve, indicating ongoing concern about the effect of AI agents on data privacy and protection.

Time burden

The results also suggest AI tools are still consuming significant staff time rather than reducing it. A quarter of respondents said AI agents take up half or more of their teams' time, while 41% said they account for a quarter of team time.

For many organisations, that points to a substantial operational overhead in managing AI systems. Rather than delivering straightforward efficiency gains, AI deployment appears in many cases to add work through oversight, maintenance and integration.

Howard Fyffe, Director of Australia and New Zealand at Twilio, said the industry is still in an early and unsettled phase of adoption.

"We are in this transient period right now, where most organisations are deploying AI Agents, but not yet using them strategically, or seeing the full ROI. Many teams are also still spending as much or more time managing these AI Agents than they are saving," Fyffe said.

He linked the issue to the complexity of existing software environments.

"This is being compounded as marketing leaders have too many disparate technology tools. There was a consensus among many delegates that the issue of tech fragmentation, and contact centres not built for the AI Age, was making it harder to use and manage AI Agents as effectively as they'd like," Fyffe said.

Patchy adoption

The poll also found that strategic use of AI agents remains limited. Nearly a quarter of respondents said their organisations use AI agents on an ad hoc basis without a central strategy, while 31% said they were still running pilots and had not yet scaled deployment.

A further 13% said they remained sceptical and had not deployed AI agents at all. Together, those figures suggest that despite strong interest in AI across customer-facing functions, many businesses are still testing the technology rather than embedding it in core workflows.

Only about one-third of respondents said their organisations had multiple agents in production linked to core business goals. That points to a gap between experimentation and broader operational use, particularly in contact centre and customer experience settings where AI has been widely discussed as a path to efficiency and service improvement.

For Australian marketing and CX leaders, the findings reflect a familiar technology management problem. Many companies already operate across multiple software systems for communications, customer data, analytics and service functions. Adding AI agents to that mix can increase complexity if integration is weak or governance is unclear.

The survey was based on a relatively small sample of event attendees rather than a broad market study, but it offers a snapshot of sentiment among practitioners directly involved in customer operations and marketing. The results indicate that while AI agents are firmly on the agenda, practical barriers around fragmented tools, staff workload and limited strategic coordination are still shaping how quickly businesses can generate returns.

"Only around one-third of the industry people we spoke to have multiple agents in production, who are driving core business goals in their organisations. This, along with conversations we are having every day within the industry, shows us that locally in Australia we are only touching the surface of what's possible with Agentic AI, and the real gains are yet to come. There are exciting times ahead," Fyffe said.