LLMs like ChatGPT are raising the bar for support interactions. Here’s what to do about it

If you’ve used ChatGPT today, you’re in good company. Nearly 700 million people now rely on it each week, according to a recent paper from the National Bureau of Economic Research (NBER).

And while you may be asking ChatGPT to tailor your weekly meal plan, others questions are falling into 3 distinct categories:

1. Asking: Asking is when a user seeks information or clarification to inform a decision. Let’s say your vacuum broke. You may ask ChatGPT to look up warranty information or tips on how to fix it before leaning on it to suggest different styles for a home with two dogs.
2. Doing: Doing is when a user wants to produce some output or perform a particular task. If you’re at work and you need to analyze a data set, you may ask ChatGPT to review the data and pull out key takeaways so you have a starting point for your research.
3. Expressing: Expressing is when a user expresses views or feelings but isn’t seeking any information or action. Maybe it’s the end of your day and you’re looking to release some stress but don’t want to bog down your partner. You tell ChatGPT about a conversation with your friend that’s been weighing on your mind.

When we look at these categories (especially “asking”) through the lens of B2C customer service, it’s clear that interactions with customers will evolve because of LLMs like ChatGPT. Support teams need to match the quality, speed, and empathy these LLMs are already training customers to expect.

Here, we cover how ChatGPT and other AI tools are shifting customers’ service expectations, how your brand can keep up, and what you need to do to better tackle these new demands.

1. Show up where users are asking questions

According to Similarweb, ChatGPT’s web traffic surpassed Bing’s in volume in the US in early 2024, making it one of the top search tools on the web.

In the past, when your customers needed help, they would search your brand’s website and go to your knowledge center for quick answers. Now, they’re often heading to AI for answers first: 49% of ChatGPT requests are categorized as “asking,” according to the NBER study.

ChatGPT and other LLMs draw from your brand’s public help content to generate their responses, which means adapting your online presence for AI requires investing in content that’s accurate, useful, and easily scannable.

Publish and highlight structured web content that LLMs can ingest and reference confidently, like:

• Detailed product descriptions that answer cross-platform questions, provide comparative context, and include specific use cases and benefits
• Exhaustive FAQs and policy summaries that address real customer questions and are phrased conversationally
• Comparison and product care guides that help AI tools understand where your products fit in the market
• Long-form educational articles with headers and subheads that mirror customer questions
• User-generated content, like product reviews and testimonials

This approach makes your brand discoverable when customers ask ChatGPT about “how to return a product” or “how to cancel a subscription.”

2. Provide guided, trustworthy experiences once customers arrive

Once customers ask ChatGPT all their initial questions and are ready to purchase or make a return, they’ll arrive on your site with more information and more primed to take action.

This is where an AI customer agent comes in. According to Klaviyo’s 2025 AI shopping index, 40% of consumers believe AI assistants build brand trust.

An AI agent provides brand-trained, always-on, conversational support across channels like email, web chat, text message, and WhatsApp. It works around the clock to make sure your customers get the trusted answers they need, when they need them, and can even answer product questions on the spot to help prevent cart or check-out abandonment.

Here are a few ways to make sure every conversation with your AI agent is both personalized and accurate:

• Make sure your AI agent has access to real-time customer data from your CRM and any customer-facing systems, like shipping platforms, loyalty programs, and marketing software.
• Train your AI customer service agent on your brand voice, policies, and past service logs.
• Enrich your AI agent’s responses by uploading custom answers for brand-specific questions and PDFs, docs, or guides.
• Implement an AI agent that can escalate to a human agent when necessary, with complete customer context preserved.

When home fragrance brand Happy Wax added an AI customer agent to their team, 50% of conversations handled by the agent were fully resolved without any service team involvement. This gave customers instant answers, and gave the team time back for high-touch moments.

3. Follow up with purpose

Just because a customer completed a purchase or had a positive support interaction doesn’t mean your job is done. You need to be proactive with your customer communication, and follow up with personalized responses so you can take advantage of the momentum.

A CRM collects, consolidates, and analyzes real-time and historical customer data from across the customer journey, then stores it in unified customer profiles and makes it available for activation in personalized marketing. That means every team member across your organization, from marketing to customer service, has visibility into the latest, most up-to-date information on any given customer, whenever they need it.

When your CRM is autonomous and AI-powered, you can easily keep up with customers during and after key moments, nurturing them over time as they build their relationship with your brand by:

• Automatically generating ongoing flows and campaigns tailored to your brand voice with help from an AI marketing agent
• Using signals from customer service conversations to create new segments for marketing campaigns and flows
• Following up with timely reminders to complete purchases on customers’ preferred channels
• Creating new automated flows triggered by customers interacting with specific products or categories

For example, when a human or AI agent resolves a customer’s issue, you can trigger a pre-planned flow to re-engage that customer with a post-resolution message that says something like, “We’re sorry you had that issue. Here’s a coupon for 10% off your next purchase, on us.”

With the right tools in place, your brand can transform routine follow-ups into moments that strengthen customer loyalty and drive revenue. Korean beauty store Adilsons, for example, uses an AI marketing agent to automatically create content, select the right audience, and optimize each send. So far, they’ve seen 50% higher open rates, 51% higher click rates, and 40% higher attributed revenue per campaign.

4. Meet consumers’ rising expectations

In their research, NBER looked for an expression of satisfaction or dissatisfaction in users’ ChatGPT conversations and categorized them as “good,” “bad,” or “unknown.” NBER found that “good” ChatGPT interactions outnumber “bad” ones 4 to 1, up from 3 to 1 in 2024.

As people get used to AI that seems emotionally intelligent and is quick, helpful, and polite, their expectations for how brands show up are rising accordingly. An AI-powered helpdesk helps your customer service team respond faster and more empathetically to customers by:

• Operating on the same set of data, so customers never have to repeat themselves
• Bringing customer conversations across channels into one unified inbox, so your team doesn’t have to waste time hunting down information
• Adding the right context to a ticket the moment it arrives and routing it to the right person, so your team can dive in fast
• Turning closed tickets and satisfying interactions into marketing triggers and repeat revenue, or pausing marketing campaigns when a customer has an open ticket
• Making every interaction actionable, with reporting that shows trends, agent impact, and opportunities to grow

By combining the power of an AI customer agent with an AI-powered helpdesk, brands are better equipped to deliver the kind of immediate help LLMs are training consumers to expect. With an AI-powered helpdesk that’s embedded in your CRM, your customer journey could look something like this:

1. A customer visits your site to track their order, and opens a web chat with your AI agent about the status.
2. Your AI agent instantly recognizes them, surfaces their recent purchases, sends them the tracking information, suggests some related products, and mentions that the customer has loyalty points they can redeem.
3. The customer has a few tricky questions about the product they ordered, which prompts your AI agent to transfer the chat to a human agent. Meanwhile, the customer has switched from web chat to text messaging.
4. Your human agent sees the full conversation with this customer across web chat and text message in one unified inbox, plus all of their order and loyalty details in their profile, and quickly gets them the answers they need.
5. Your marketing team picks up where your service team left off, following up with that same customer to see if they’ve got everything they need, and offering some recommendations for complementary items.

For London-based clothing brand Folk Clothing, implementing an AI agent and helpdesk reshaped how they support customers and scale their service operations.

Affectionately known as their “robot colleague,” their AI agent works by ingesting the brand’s entire storefront to inform responses, 24/7. Their helpdesk, meanwhile, brings messages from every channel into one organized workspace so that no enquiry slips through the cracks.

The results: 53% of support conversations resolved by the AI agent in 90 days, a 75% period-over-period reduction in average ticket resolution time in 30 days with the helpdesk, and a more reliable, transparent process for both staff and customers.

Bring your support org into the AI era

Customers now expect service that feels human, acts instantly, and knows them by name. That’s a challenge for traditional ticketing systems and customer service operations.

Klaviyo Service brings service and marketing together onto one platform where resolutions fuel relationships and revenue. It helps brands deliver service that’s fast, empathetic, and predictive, and it’s how support teams keep the relationship when AI starts the conversation.

Klaviyo Service includes:

• Klaviyo Customer Hub: a personalized destination where customers can manage orders, redeem offers, discover products, and get support, all in one place
• K:AI Customer Agent: a 24/7 autonomous service agent that’s trained on your storefront and customer data to answer questions, recommend products, and resolve issues instantly, across web chat, email, text messaging, and WhatsApp
• Klaviyo Helpdesk: an AI-powered helpdesk that brings AI and human agents into a unified workspace across email, chat, SMS, WhatsApp, and social, providing full customer context for faster response times and more personalized interactions

With Klaviyo, the results are faster responses, happier customers, and stronger loyalty. Because when support becomes intelligent and empathetic, it stops being a cost center—and becomes your most powerful retention engine.