As AI takes over personalised campaigns, customer conversations and creative production, brands are finding new ways to communicate at scale. But recent research suggests that faster responses and smarter recommendations do not always translate into better customer relationships. With consumers still seeking transparency, empathy and human support, marketers face a growing question about how much of the experience should be automated.
A customer receives a promotional email from a favourite fashion brand. The recommendations match their interests, the message arrives at the right moment and the language feels surprisingly personal.
The customer replies with a question about an exchange. Within seconds, another message arrives, addressing the concern in a friendly and reassuring tone.
But was anybody actually involved in that conversation?
As artificial intelligence becomes part of marketing operations, the distinction between communication created by people and communication generated by machines is becoming less obvious.
Brands are using AI to recommend products, develop campaigns, answer queries and personalise customer journeys. Some systems can now manage conversations with limited human involvement.
For consumers, these developments promise convenience. For marketers, they offer the possibility of reaching more people without proportionately increasing resources.
However, an automated response that sounds friendly is not necessarily the same as an experience that feels personal.
The distinction matters because marketing is increasingly expected to function as a conversation rather than a series of promotional messages.
According to Salesforce’s 2026 State of Marketing findings for India, 81% of surveyed marketers had adopted AI, while 92% said customers increasingly expected brands to support two-way conversations.
Yet the expanding use of automation raises another question: when customers expect brands to understand them, how much of that understanding can technology provide?
Recent findings from Salesforce, Canva, Capgemini, Adobe and Kantar indicate that consumers are not rejecting AI outright. Instead, their expectations vary depending on the situation, the information being used and whether human assistance remains available.
When Personalisation Starts Feeling Less Personal
Personalisation has been one of digital marketing’s central promises. The idea is straightforward: relevant communication should be more useful than a generic message sent to everyone.
AI allows marketers to extend that approach across larger audiences and more interactions.
But personalisation depends on what a brand knows about a customer, and that information is not always complete.
Salesforce’s 2026 India research, based on 250 marketing decision-makers within a wider global study of 4,450 respondents, found that 83% needed more personalised content than they could currently produce.
At the same time, 98% reported barriers to personalisation, including disconnected data, privacy concerns and technical limitations. Seventy-one per cent said they struggled to respond promptly because they lacked the necessary customer context.
Consider a hotel sending an automated offer to someone who stayed at its property recently.
The offer may reflect the guest’s previous preferences. However, if that stay involved an unresolved complaint, a cheerful message encouraging another booking could appear insensitive.
Similarly, a retailer might recommend a product that the customer has already purchased offline because its systems have not connected the transaction.
In both cases, the communication is technically personalised but poorly timed.
“The biggest barrier to personalisation today isn’t AI, it’s the quality and connectedness of the data that powers it,” said Nishant Kalra, Vice President, Sales, Salesforce South Asia.
These limitations suggest that AI-generated relevance still requires decisions about data quality, timing and appropriateness.
The concern extends beyond inaccurate recommendations.
Canva’s 2026 State of Marketing and AI study found that 52% of consumers considered an advertisement invasive when it appeared to anticipate a purchase before they had searched for it. Another 58% did not want brands using AI to predict their needs.
The findings illustrate the difference between useful recommendations and experiences that make consumers uncomfortable.
For example, reminding someone about a product left in an online basket may be perceived differently from displaying an advertisement that appears to reveal an interest they never explicitly shared.
AI can identify patterns in both situations. Whether a brand should act on those patterns is a separate decision.
For marketing teams, maintaining a personal connection may therefore involve limiting certain forms of targeting rather than continually increasing precision.
The challenge is not simply how much personalisation technology can deliver, but whether the resulting interaction makes sense from the customer’s perspective.
Do Customers Still Want to Speak to a Person?
AI’s growing role is also changing customer service, an area that increasingly overlaps with marketing.
A brand may attract a consumer through personalised advertising, but the relationship is often shaped by what happens after the purchase.
Customers may accept automated assistance for routine enquiries. They may have different expectations when the situation involves money, urgency or disappointment.
Capgemini Research Institute’s 2026 customer experience study, involving 9,500 consumers across 16 countries and 1,200 executives and front-line employees, found that 66% of consumers ranked front-line staff among their three preferred interaction channels.
By comparison, 56% included generative AI-enabled bots or agents among their top three choices.
The findings do not suggest that consumers universally prefer one channel. Rather, they indicate that human interaction remains important even as automated alternatives become more capable.
The same report found that the proportion of front-line staff using generative AI daily had increased from 42% in 2024 to 78% in 2026.
Among employees using the technology, 76% reported spending less time on mundane activities, while 70% said their responses had improved.
That points towards a different role for AI in customer engagement.
Instead of placing a chatbot between the customer and the company at every stage, organisations can use AI to support the employees handling enquiries.
A service representative could receive an automated summary of previous conversations before speaking to a customer. AI could retrieve policies, translate messages or identify possible solutions.
The representative would still decide how to respond.
This distinction becomes particularly relevant when standard procedures do not adequately address a situation.
A traveller whose flight has been cancelled may appreciate instant information about alternative departures. But someone travelling for a medical emergency might need an employee authorised to consider circumstances outside the usual policy.
In financial services, customers may similarly value automation for routine transactions while preferring personal assistance for complex decisions.
“For personal financial decisions, customers place greater trust in human professionals than AI-driven bots,” said Hitender Sachdeva, Head of Merchant Service, Products and Payments at Scotiabank, in the Capgemini report.
The question for brands is therefore not necessarily whether every enquiry can be automated.
It is which enquiries benefit from automation and which require human judgement.
A system that answers routine questions accurately can reduce waiting times. But an effective customer experience also depends on what happens when the system cannot resolve the problem.
What Happens When the Chatbot Doesn’t Tell You It’s a Chatbot?
The next issue concerns transparency.
As AI-generated conversations become more natural, consumers may find it increasingly difficult to determine whether they are communicating with an employee or an automated system.
That uncertainty can affect trust even when the response itself is accurate.
Adobe’s 2026 AI and Digital Trends research, developed with Oxford Economics, surveyed 3,000 executives and practitioners alongside 4,000 customers across international markets, including India.
It found that 43% of customers were willing to interact with a brand’s AI concierge or agent.
However, 37% said they would disengage if they discovered they had been interacting with AI when they expected a human being.
Customers identified the ability to switch to a person at any time as an important expectation in AI-assisted interactions.
Among organisations, 68% considered clear disclosure of AI interactions important for establishing trust, while 61% identified easy escalation to human support.
These findings suggest that making chatbots sound more human is only one part of designing an acceptable experience.
A customer asking about a delivery date may be satisfied with an automated answer, particularly if the response is immediate and accurate.
But a customer dealing with a missing order may become frustrated if the conversation repeatedly returns to scripted explanations without offering a meaningful resolution.
In such situations, transparency and access to human assistance can be as important as conversational fluency.
Adobe also identified a gap between organisational expectations and consumer preferences.
While 49% of organisations believed customers would eventually want AI agents to become their primary way of interacting with brands, only 19% of customers shared that expectation.
The difference raises questions about how brands measure successful automation.
Resolving a large percentage of enquiries without employee involvement may reduce operating costs, but that figure alone cannot establish whether the experience was satisfactory.
Marketers and customer experience teams may also need to evaluate failed interactions, repeated enquiries, abandoned conversations and the quality of transfers to human staff.
An automated system that recognises its limitations and connects a customer to the right person could deliver a better experience than one designed to prolong every conversation.
Can AI Create Content Without Losing the Brand’s Personality?
The human connection is not limited to customer support. It is equally relevant to advertising, storytelling and creative development.
Generative AI allows marketing teams to produce copy, visuals and campaign variations more quickly. It can also adapt material for different languages, formats and audiences.
However, greater production capacity does not automatically make communication distinctive.
Canva’s 2026 study, conducted with The Harris Poll among 1,415 marketing leaders and 3,547 consumers across seven countries, including India, found that 97% of marketing leaders used AI in their daily creative work.
Consumer attitudes were more complicated.
Although 68% did not mind AI-supported advertisements when they were helpful or relevant, 78% said they preferred advertising created by people, even if AI could produce better results.
Another 87% believed the best advertising still required a human touch, while 70% felt they could usually identify AI-generated advertisements because something appeared to be missing.
“AI has changed how marketing gets made, but not what makes it effective,” said Emma Robinson, Head of B2B Growth Marketing at Canva.
For marketers, this introduces a distinction between generating content and creating communication that audiences recognise as belonging to a particular brand.
An AI tool may produce several versions of a festive campaign. But choosing the right cultural references, understanding regional humour and ensuring the message reflects the brand’s established personality may still require local knowledge and creative judgement.
The same applies to influencer partnerships and customer storytelling.
AI can assist with editing, translation and distribution. It cannot independently establish whether a creator’s relationship with an audience is credible or whether a campaign accurately represents a customer’s experience.
The issue may become more significant as competing brands adopt similar creative tools.
When producing professional-looking content becomes easier, brands could find it more difficult to distinguish themselves through execution alone.
Consumer trust in AI-mediated brand information also remains limited.
Kantar’s Media Reactions 2026 research, covering more than 23,000 consumers across 33 markets and 800 senior marketers, found that 73% of consumers used AI assistants in their personal or professional lives.
Thirty-three per cent already used them to research brands and products, but only 24% trusted brand recommendations from AI assistants.
The findings suggest that growing familiarity with AI does not necessarily translate into confidence in everything it communicates.
For marketers, the implications extend across the customer journey.
AI can help identify relevant audiences, generate messages, support employees and respond to routine questions. Yet the quality of those interactions continues to depend on accurate information, appropriate boundaries and accountable decisions.
The research also indicates that consumers distinguish between situations where automation is convenient and situations where they expect human involvement.
A brand may not need an employee to answer every delivery enquiry. It may, however, need someone available when the delivery fails, the customer is distressed or the standard response is insufficient.
Similarly, an AI-generated advertisement may be effective without imitating a human creator, provided its message is relevant, credible and consistent with the brand.
As marketing becomes more automated, the personal element may increasingly depend on decisions made before and after the technology is used.
The emerging question is not whether brands can make AI sound more human. It is whether customers can still find understanding, choice and accountability behind the automated experience.
Disclaimer: All data points and statistics are attributed to published research studies and verified market research. All quotes are either sourced directly or attributed to public statements.