Early Colleague Involvement – the Key to AI Success?

Across the first two events in the CCMA’s How to…Succeed with AI roadshow series, one message has resonated louder than others: organisations that involve their frontline colleagues early in AI implementation are the ones most frequently declaring success.

While some technology vendors promise transformative results from AI and senior business leaders seek quick-fire efficiency gains, it’s clear that, ultimately, it’s the people using and interacting with the tools daily who will determine whether AI projects thrive or fail.

Real-World Applications

Throughout the roadshow, leaders shared applications that succeeded because frontline colleagues helped shape them. Real-time call transcription ensuring factual accuracy – designed after advisors highlighted compliance concerns. Sentiment analysis tools calibrated based on what frontline colleagues identified as genuine emotional signals. Natural language IVR built by understanding which routing frustrations advisors hear about most frequently.

One organisation found they were losing female candidates for job roles, and used AI to reframe job descriptions, subsequently improving both attrition and absence rates. This insight came directly from recruitment advisors who noticed the pattern.

With each success case, it was evident that when teams feel consulted rather than “done to,” adoption rates soar and any implementation challenges surface early – when they’re easier to fix.

Contact centre leaders understand this challenge – and have emphasised the need to test AI internally first and understand how it can genuinely help their teams, before committing to broader deployment.

The Failure Rate No One Talks About

Kevin McGachy, Head of Solutions at Sabio, provided a sobering statistic during the roadshow: 80% of AI projects fail to move from proof of concept into production. And this failure rate isn’t primarily technical – it’s cultural and operational too.

McGachy emphasised that business cases must demonstrate more than cost savings. They should evidence improvements in customer experience, accessibility and employee satisfaction. Crucially, organisations need to involve frontline colleagues in defining what success looks like and how it should be measured.

He advocated for starting with model offices operating in small numbers to measure against business as usual, accepting that failure isn’t necessarily negative and can often help shape outcomes and secure future buy-in. This approach requires frontline participation from day one – testing, providing feedback and iterating before wider rollout.

Leaders stressed the value of pivoting quickly when solutions aren’t delivering, noting that not everything needs full integration and that small, creative changes suggested by frontline teams can make significant differences without requiring massive investment.

The Path Forward

The How to…Succeed with AI events have heard a number of participants warning against pushing tools onto people without securing hearts and minds. Organisations that disengaged their people by implementing AI without any consultation or active involvement face resistance, workarounds and potential failure.

The consensus has been that successful AI implementation requires collaboration, with technology as the enabler and people as the driving force.

The CCMA’s How To…Succeed With AI events continue through January and February. For attendance details visit: https://www.ccma.org.uk/how-to-events/

Supported By