3 days ago
Worth talking about: the future of conversational AI in business
Cognigy is a Business Reporter client
Stunning advances in AI technology over the past couple of years are creating new ways for organisations to conduct more meaningful and natural conversations with their customers.
One such technology is conversational agentic AI.
Now launching into the UK, Cognigy is a specialist in enterprise conversational AI. Its flagship solution, combines generative and conversational AI to deliver hyper-personalized, multilingual service across channels, empowering enterprises with scalable Voice and Chat AI Agents, Agent Copilot tools, and real-time support. With proven success in industries like banking, travel, and utilities, and trusted by major brands including Nestlé, Lufthansa, and Mercedes-Benz, Cognigy is setting a new benchmark in intelligent automation for contact centres worldwide.
We sat down with Sebastian Glock, Technology Evangelist at Cognigy, to ask him how changes in conversational agentic AI are unfolding, where it's all headed and what organisations need to keep in mind as they explore the new possibilities.
How capable is conversational AI today and what could it offer in the future?
The chatbots and voice bots of a few years ago often disappointed users with bad answers or by the need to respond in unnatural ways, such as giving single-word requests like 'refund' instead of using full sentences. They were also slow. Above all, they fell short of the hype and expectations that had been built up. People expected science fiction but instead they got 'Sorry, I didn't get that' over and over.
The rise of large language models has transformed conversational AI. Yet that's not the full story. LLMs are impressive but can also be unpredictable. Where things really click is when you combine that raw power with structure, purpose and guardrails that contain tight controls, so responses stay relevant, safe and on-brand.
AI agents can then conduct interactions with humans in a way that feels natural. Human-to-machine communication becomes almost indistinguishable from human-to-human communication. The ability for a machine to have a smart and contextual conversation with a human is something that was impossible even as recently as two years ago. And even though it's not yet widely implemented, the technology to have effortless, natural and productive interactions is here and it works.
Looking ahead, there will soon come a time when humans will prefer to talk to the AI rather than a human contact centre agent. Interactions will become so good that nobody will want to spend the time or effort trawling through a website to find the information they want. Imagine being able to simply talk to a website and it instantly responds with exactly the information you wanted.
How can AI customer service agents meet the varying needs of different organisations?
AI agents allow companies to combine all the benefits of automation with a greatly improved customer experience that offers less waiting time, better answers and more empathetic communication. At the same time, organisations can decrease service costs by automating their customer conversations.
The technology is flexible, allowing organisations to blend human and AI interactions to suit their needs. Intelligent conversation design ensures that if a customer makes a difficult request – for example, asking for a discount that AI cannot authorise – a human will take over. Workflows can be tailored so the AI might say, 'Let me check with my supervisor,' and then follow up with a human-style email for a personal touch, even if the response is AI-generated.
For premium brands, an AI can verify identity and route calls, yet every interaction ultimately connects the customer with a human expert. Conversely, some companies may limit human involvement; here, if the AI is unsure, it will call you back with an answer after consulting a human.
It's also critical to enable AI agents to access existing data and tools like a CRM or ERP system. This lets the AI understand a caller's recent orders, preferences or past issues. That context allows for a much more personalised and efficient exchange, which makes things smoother for both sides.
We also work with our clients to make sure the AI assistant speaks in a way that reflects their brand identity, whether that's professional and formal or more casual and conversational. It's not just about words. It's about pacing, empathy and how solutions are delivered. Done right, it can feel like a natural extension of the brand.
What is a good example of how a company is using your AI solutions to benefit their business?
The best results come from industries with a high volume of incoming customer enquiries that are typically repetitive. A good example is insurance or finance, where most interactions often involve similar requests.
In one case, we work with a large European bank to improve how it confirms appointments for credit requests that customers make online. Previously, skilled loan advisors manually dialled applicants' numbers, but 80 per cent of these calls failed due to no answer, hang-ups or confusion. Now, Cognigy's voice AI agent automatically calls each number, verifies the loan application and asks if the customer is ready to speak with an advisor to complete the process. The AI agent even offers flexible scheduling or records if a customer is no longer interested, with all data fed into the system.
About 80 per cent of calls still don't result in a successful loan application, which is the same as before. However, the huge difference here is that it dramatically decreases the workload on loan advisors, saving them countless hours and thousands of unnecessary calls. Most importantly, of the 20 per cent which succeed, about 85 per cent are directly transferred to a human loan advisor which accelerates the process, boosts conversion rates and ultimately generates more revenue for the bank at a much lower cost.
How do you answer an organisation's concerns about an AI agent's reliability and compliance with data protection and privacy rules?
We often get people requesting an AI solution that is easy, fast, all-knowing and transactional. They also want to know the solution is safe, with guard rails to prevent it operating outside of the intended scope.
Transparency and control are critical to meet such requirements and to comply with regulatory demands. We give companies a clear view into how their systems operate and make sure the data stays within their environment. For industries with strict rules, like healthcare or finance, we offer deployment options that meet even the most rigorous requirements, including on-premises setups if needed. Every case is different and we use different setups, different configurations, and different cloud vendors for different requirements. The key point for enterprises is that proven, production-ready solutions already exist. Success or failure rather lies in getting the implementation right which is why experience plays a big role, something that is in high demand but low supply right now.
What advice would you offer companies that are just beginning to explore this technology?
Don't try to do everything at once. Look for a use case that can demonstrate success quickly and then expand from there into more use cases. Instead of an 18-month overhaul, consider a four-week proof of concept to quickly deliver results and expand from there. Your customers, agents and your project will benefit faster and those quick results will get you more internal buy-in, and perhaps budget, to continue expanding.
AI agents will soon outnumber human employees, with each person managing multiple digital assistants. In this new landscape, it's crucial to implement future-proof, scalable solutions rather than isolated point-to-point systems that lack seamless integrations.
Any final thoughts?
Large language models are improving rapidly while costs plummet – new advances emerge every month. I'm excited for a future where personal AI assistants handle everyday tasks like booking restaurants and scheduling appointments through our mobile devices or wearables. By learning our habits and accessing our calendars, these assistants will simplify our lives.
However, this shift also brings challenges for businesses. Imagine a customer telling their AI agent to call their insurance company, which is using its own AI to answer. As human-to-AI and AI-to-AI interactions become common, companies must adapt quickly to evolving customer experiences and rethink processes for the AI first era.
