The disconnect between marketing and operations is now obstructive
Marketing tends to dominate the development of and communication over customer facing channels such as websites, mobile apps and social media. Meanwhile, the Operations team are seen as responsible for Customer Engagement through any additional Contact Centre channels such as voice, chat and email. This has resulted in siloed points of customer contact which make it incredibly difficult for the business to implement new concepts such as ‘end to end’ customer journey analytics and orchestration.
In truth, both the marketing and operations sides of the business can learn something from each other. Marketeers have been using struggle analysis and customer effort metrics for years to boost sales engagement and conversions, with an abandoned shopping cart often resulting in a follow-up email nudging the customer to complete a purchase.
Those same customer journey orchestration concepts and new real-time data technologies can now be used by operations to add additional context to augment the customer experience (CX) so that when a customer drops out of an on-line self-service process the business can re-engage more effectively with them. From a Sales and Marketing perspective, the information on customer intent can be made available to the agent together with relevant ‘next best action’ information when they call into the contact centre, enabling that agent to incentivise the customer to complete a purchase.
AI needs to be ‘on brand’ and ‘optimised for CX’ in 2026
AI voice and chatbots are currently available in a limited range of personas. The challenge being that these default personas don’t allow the business to differentiate nor do they align with the brand and its values which risks the business compromising or losing its tone of voice and its identity.
A further challenge is that although the AI can be configured and set up to only reference validated data to gain a deeper understanding of the domain and customer interaction business use cases it is automating and supporting, there is still an issue with how AI responds. This needs to be performed in a consistent ‘on-brand’ way using phraseology that is optimised to influence the desired customer behaviours and outcomes. What’s happening today is that LLMs are influenced by the customer’s own phraseology in their interaction with the AI, and this then influences the AI and its responses not just to that customer but in future interactions too.
Of course, this also represents an opportunity for the evolution of specific CX Language Models. What we’d like to see are LLMs designed around optimising the ‘CX’ whilst reflecting the company’s Brand and Tone of Voice. This is a key requirement to improve AI driven customer engagement, speedier resolutions and reduced handling costs by influencing the right user behaviour. How the AI engages can then be designed from a CX perspective to elicit certain responses.
Moreover, as we better understand current customer journeys and analyse the volumetrics that reveal why people are calling we can look at how effectively those are handled. Voice and Chat bots will be able to contextualise customer intent in real-time to determine changes in sentiment and whether they need to be referred to self-service or a human agent with the requisite skills. Those referrals will then be analysed to determine how successful they were. Where a bot is incorrectly correlating intent with outcome, humans can be brought back into the loop to educate the bot. We’re already seeing the beginnings of this today with the analysis of transcriptions capturing both voice and chat bot, as well as the customer-agent interaction.
Humans will become a precious resource
Human-to-human interactions will always add more value than a bot interaction. For the business this creates a dilemma. When should you tone down your AI and refer to a human? To know that you need real-time awareness and to leverage all of the known validated information that is within the organisation about that customer, their intent and why are they contacting you.
In certain situations, it makes sense to have some form of agent interaction because there’s an upsell opportunity or simply an opportunity to engage and build brand loyalty that could circumvent the business losing that customer. But in other scenarios, AI becomes the first form of response.
If the contact centre comes under pressure from high call volumes, for instance, the time to answer may reach the point where calls are likely to drop. In this scenario, servicing those calls through an AI bot could lead to the better outcome. So, it’s a matter of using AI dynamically to address calls and human agents to handle perhaps more complex calls. The strategy will vary from business to business. But there’s currently insufficient thought being given to how these resources are used. There’s no reason why response should not be more dynamic, with the depth of AI engagement ramped up or down according to the needs of the business.