AI Agents: Going Beyond the Automation Narrative

New BCG research shows 67% of companies are planning for autonomous AI systems, but the real challenge lies in turning potential into business value.

Every technology wave creates a moment where its impact becomes impossible to ignore. For artificial intelligence (AI), that moment is now. AI is too different, too disruptive and too close to human capabilities for any business to sit on the sidelines. That’s why it isn’t surprising that 67% of companies are considering AI agents as part of their transformation strategy, says the BCG AI Radar report. The nature and potential impact of AI is rewriting the rules of what’s possible in business. And unlike recent technological trends like artificial reality, virtual reality and blockchain where use cases remained abstract, generative AI’s ability to mimic human conversation brings its potential startlingly close to home. People can see their own work reflected in what AI produces, making its impact immediately tangible and relevant.

Agents vs. automation

However, while everyone’s considering AI, businesses are all asking the same thing: where and how should we deploy it? In our own experience at Nutun implementing AI platforms, the early signs are incredibly promising. We’ve found that approximately 30% of our workloads could be handled by AI agents, leaving humans free to tackle more sophisticated challenges. (Keeping in mind that this isn’t about replacement, it’s enhancement.) What makes these emerging ‘AI agents’ fundamentally different from current automation tools is their intuitive interaction model. Traditional automation tools rely on prescriptive programming – click here, press there, take information from this spreadsheet. AI agents, on the other hand, work through natural conversation. That said, we haven’t yet reached the point where AI can observe a human process end-to-end and replicate it. The technology excels at discrete components rather than complete process automation.

From a business process outsourcing perspective, we see three main opportunities. First, fully autonomous AI voice and text agents can handle customer interactions while seamlessly integrating with organisational processes. Second, AI can serve as a powerful assistant to human agents, eliminating those frustrating moments when customers wait while agents search through systems. Third, and perhaps most transformative, AI excels at quality assurance and more interestingly, insights. Consider the traditional approach to customer insights like net promoter scores, satisfaction surveys and mystery shopping. These mechanisms are always one step removed from actual customer conversations. AI can analyse millions of interactions, understanding not just what customers say, but what it means for your business. For example, you can ask the AI about trends in sales decline, and it will analyse countless conversations to identify patterns in pricing concerns, feature requests, or usability issues. Brilliant.

Getting AI right

However, there are several key challenges to getting this right. Many organisations struggle with use case identification amid an overwhelming choice of technology platforms. This abundance of choice often leads to analysis paralysis. My advice? Pick a platform that reasonably does what you need and has relevant industry expertise built in. The technology decisions will become clearer once you’ve identified your specific use case. Integration is another issue. AI agents must become part of a broader customer journey. Companies trying to replace entire systems or processes at once will struggle. The key is taking it step by step, focusing on deployments that add clear value and enhance the customer journey. Take lessons from the chatbot era – when poorly implemented, these tools frustrate customers and drive them away. AI agents offer superior capabilities, but the principle remains the same: integration must enhance, not hinder, the customer experience.

Surprisingly, the skills needed to work alongside AI agents are less technical than you might expect. AI is actually easier to adopt because it mimics how people naturally communicate. Unlike the digital era which required huge new teams of UX, CX and a host of other new titles to make things simple, AI just needs people who are willing to try it. From what I’ve experienced, the real challenges lie in organisational readiness – managing policies, controls and risk frameworks for AI deployment. For business leaders, the message is clear: AI agents represent a transformation that cannot be ignored, but neither should it be rushed. Success will come to those who take an incremental approach, focusing on specific use cases where AI can genuinely enhance customer experience and operational efficiency. The future isn’t about replacing humans with AI agents; it’s about creating intelligent partnerships that leverage the strengths of both. The organisations that thrive won’t be those that use AI merely for cost reduction, but those that harness it to augment their workforce’s capabilities and create new value. In the end, customers will vote with their feet. Make the AI-enhanced journey seamless and valuable, and they’ll embrace it. Create friction, and they’ll go elsewhere.