When faced with a problem or customer service query, most of us would rather speak to a real person than a chatbot.
It’s quicker, easier and less likely to result in your phone being thrown at the wall.
But that may not be the case for much longer.
As machine learning and AI advance, chatbots are becoming increasingly sophisticated – capable of holding natural conversations, solving problems and personalising interactions.
Could this redefine how brands connect with consumers?
To find out, we sat down with Simone Blakers, Customer Service and Digital Transformation Consultant, to track the evolution of chatbot technology – and explore how marketers can harness its potential. From robotic scripts to real conversations
Everyone can remember a frustrating encounter with an interactive voice recording.
You know - the ones that have you shouting ‘yes’ and ‘no’ down the phone, only to wind up in a maddening loop of ‘Did you mean…?’ and ‘Sorry, I didn’t quite catch that.’
These early chatbots were little more than glorified decision trees developed by customer service teams to follow rigid scripts. They couldn’t understand context or solve issues on their own.
But today’s chatbots? An entirely different breed.
‘Thanks to recent developments in AI, today’s bots are much more advanced,’ says Simone Blakers, a customer experience and automation specialist who helps businesses navigate digital change. ‘They’re more like digital assistants than chatbots as we’re used to thinking of them.’
Using a blend of robotic process automation (RPA), large language models (LLMs) and conversational AI, these digital assistants can process information dynamically and generate responses on the fly.
‘This means they can execute more complex workflows,’ explains Simone. ‘Everything from capturing a new lead to populating a database with their information, extracting key details – even generating a personalised PDF to send back to them. It’s a big step forward from simple scripted responses.’
And the technology is evolving fast.
‘These digital assistants aren’t yet fully autonomous, but we’re heading in that direction. At the most advanced end of the spectrum, we’re seeing the rise of full-scale AI agents. These won’t just assist humans; they’ll work entirely independently – reasoning, making decisions and taking actions based on data, much like a human employee would.’
Naturally, this technology is transforming how businesses operate.
‘Digital assistants can handle significantly more interactions than human teams, making engagements faster and more efficient,’ Simone explains. ‘Plenty of companies – across a range of industries – already have digital workers handling tasks behind the scenes, helping with everything from data entry to processing payments.’
In healthcare, for example, they’re used to schedule appointments and provide basic follow-up care information – reducing the administrative burden on staff.
Retailers, meanwhile, are leveraging them for customer enquiries, product recommendations and purchase assistance.
‘Expedia is a great example in travel,’ Simone adds. ‘Expedia’s generative AI assistant, Romie, can join WhatsApp chats, listen in on holiday plans and help with booking and itinerary suggestions.’
Though the benefits are substantial, deploying digital assistants in customer-facing roles does pose challenges.
‘When you put these more autonomous bots in front of consumers, there’s a lot more risk involved,’ says Simone. ‘Businesses need to be mindful of that.’
Security, in particular, is paramount.
‘If you’re connecting a bot to transactional data – order numbers, shipping addresses, payment information etc. – your infrastructure must be secure,’ Simone warns. ‘If that system gets breached, you could face serious legal and reputational consequences.’
A major concern is jailbreaking, where a chatbot is manipulated to bypass its built-in safeguards.
‘If an AI assistant is compromised, it can be tricked into providing false information, sharing restricted data or responding in ways that could damage a brand’s reputation,’ Simone explains. ‘That’s why many consultants don’t recommend deploying fully autonomous AI agents in public-facing roles.’
When it comes to chatbots, the best – and safest – outcomes are achieved when AI supports human agents, enhancing their capabilities rather than replacing them entirely.
‘Research shows AI often performs best when working alongside people, not instead of them,’ Simone notes. ‘A human agent supported by AI can handle more queries, make better decisions and deliver higher-quality customer service. While AI will replace certain tasks – and unfortunately, in some cases, entire roles – the priority right now is identifying where human expertise adds the most value and where automation can drive greater efficiency.’
Making the most of AI in customer engagement
With keen oversight, robust training in your business’s policies and branding, and a thorough overview of the backend technical solutions required, today’s AI-driven chatbots can do far more than answer basic queries, including:
Delivering comprehensive support: Handling FAQs and providing real-time support in a natural, conversational way.
Triaging leads: Identifying high-value prospects, collecting relevant information and routing them to the right teams.
Personalising interactions: Tailoring responses based on user data, past interactions and preferences to create a more engaging experience.
Automating workflows: Handling administrative tasks like appointment scheduling and database management to reduce staff workload.
Generating content and recommendations: Creating personalised proposals or product suggestions.
Supporting human employees to boost customer engagement: Retrieving data, summarising customer interactions and flagging complex issues that need human input.
Equipped with these advanced capabilities, AI-driven chatbots do more than enhance customer experience. They free up human agents to focus on more complex tasks.
Now that we know what the next generation of chatbots can do, let’s explore how you can implement them effectively into your business.
Below are Simone’s top tips.
Before deploying a digital assistant in a customer-facing position, identify what you want it to achieve. Are you aiming to qualify leads? Reduce customer service workload? Personalise marketing?
Understanding the problem you're solving will help align the assistant’s function with your business objectives.
AI chatbots require careful technical implementation to ensure reliability, accuracy and cost-effectiveness. Work closely with your technical teams and vendors to:
Test for security vulnerabilities to prevent data breaches
Ensure consistency and accuracy in chatbot responses
Monitor for hallucinations (AI-generated false or misleading responses)
Optimise cost management – sometimes a simple database query is more cost-efficient than using a complex LLM
Even the most advanced AI needs boundaries to function effectively and ethically. Establish rules that define:
What topics the chatbot can and cannot address
When and how it should escalate to a human
How it should respond in sensitive situations (e.g. customer complaints)
A chatbot should reflect your brand’s voice.
‘Design the experience to reflect how your brand would speak and act,’ says Simone. ‘In healthcare, for example, chatbots are designed with built-in empathy, using phrases like ‘I’m so sorry to hear that’. It’s all about personality design.’
Even small details – like using dots to indicate the bot is ‘thinking’ – can make interactions feel more natural.
Before developing a chatbot, leverage the data you already have.
Your existing customer interactions, FAQs and support tickets hold valuable insights that can help you identify where automation will have the biggest impact and train your chatbot to handle common queries effectively.
Unstructured data – such as customer feedback and video responses – can also be a goldmine for refining chatbot capabilities.
Modern chatbots can do far more than simple text-based interactions. They can now analyse voice, images and video.
‘With chatbots no longer limited to pre-set options, there are all sorts of new possibilities for engagement,’ says Simone. ‘Let’s say a customer purchases a new piece of software. They could record a video or upload a document explaining how they plan to use it – and the digital assistant could then generate a personalised onboarding experience tailored to their needs.’
Make the most of these new capabilities to deliver richer, more personalised interactions.
Transparency isn’t just ethical. It means a better user experience and stronger brand credibility.
‘Customers need to know when they’re interacting with an AI assistant, and brands must be upfront about its limitations,’ says Simone.
Take inspiration from retail, where new employees often wear ‘Hi, I’m new!’ badges to set expectations. The same approach works for AI. Let users know when a bot is still learning and offer opportunities for feedback.
‘And always provide an easy way for users to escalate to a real person if needed,’ Simone advises. ‘Complex issues still require human judgment.’
Treat your chatbot like an employee. Like a human, it needs ongoing training and supervision.
‘Never set and forget,’ warns Simone. ‘If something goes wrong, it’s going to damage your brand. It’s the same as having a bad sales assistant – if they’re rude to your customer that customer won’t come back.’
To avoid costly missteps, make sure you:
Regularly review interactions to identify issues and fine-tune responses
Test for biases to ensure fair and accurate replies
Gather user feedback to improve the chatbot’s effectiveness
By continuously monitoring and optimising, you can ensure your bot is a valuable asset – not a liability!
Marketing is changing fast. Stay on top of the latest trends by following us on LinkedIn or exploring our latest insights in the ADMA Resource Centre.