Why AI will transform how customers self-serve online

AI agents are already answering simple customer queries across voice and web service channels, often directing them to self-service portals. Anything that can’t be solved immediately gets escalated to a person. However, AI is going to quickly transform how customers self-serve online and businesses who don’t quickly respond will get left behind.

Having built many self-service portals for use in B2C and B2B settings, and more recently customer-facing agents, it’s a trend we’re playing close attention to.

After he spoke on the topic at the Digital Transformation Conference 2026 in London, UNRVLD’s CTO Darren Webb gave us his thoughts on why the change occurring in customer service is more fundamental than earlier stages of evolution, and where businesses should start with readying themselves for this market disruption.

Why do you think the introduction of AI is going to lead to more significant transformation in the way customer service is delivered?

Let's start with a bit of customer service history first, because I think it helps explain why what’s happening with AI is different. From 1876 when the telephone was invented, through switchboards, IVR and call centres we were trying to handle more volume of enquiries at lower cost. Then the 90s happened. The web arrived with email and live chat promising to reduce call volumes. Then, from 2007, social media emerged.

None of these actually replaced what came before. The advancements in each era just created more channels, complexity and cost. But AI isn't just another channel. For the first time, it has the genuine potential to change everything by restructuring how customer service actually works. That's what makes this different.

Why do we need to invest in AI for customer service now?

Customers are already trying to self-serve whether you support that journey or not. 81% of customers try to resolve their issue before they pick up the phone. The question is whether your portal is good enough to let them.

61% prefer digital for service interactions, that's up from what was 45% a year ago. Plus, self-service resolves routine issues three times faster than traditional channels.

Change in the experience it is possible to deliver comes from AI and the data layer you build on.

In the old reactive customer-service model with static FAQs, rules-based chatbots walk customers down a decision tree until it runs out of branches. It’s generic content for everyone and the customer has to define the problem in a way that matches the organisation’s rules. Data required to solve the problem or answer the question is often siloed across systems. So, there are high hand-off rates to agents.

With an AI model, the intent of the query can be understood, not just keywords. Journeys can adapt to who the customer is, and predictive capability anticipates needs before an ask. A unified data layer underneath it all enables this.

The shift from reactive to predictive changes the commercial logic of a portal for customer service. It stops being a cost centre and starts being a competitive asset.

What capabilities do we need to enable truly intelligent self-service solutions?

There are four that we advise clients to focus on:

Intelligent Search - this is probably the highest-impact, most underrated improvement available. When a customer can type a question in their own words and get the right answer, search abandonment drops by around 40%.

Conversational AI - proper conversational AI holds context across a multi-turn interaction. The customer can change their mind, add details or rephrase their ask. The system keeps up. Around 35% of queries can be fully resolved without any human handoff.

Personalisation - the portal knows this customer's history, their segment, their likely next need. Content and customer journeys adapt in real time.

Predictive Support - the system spots that something's probably about to go wrong and reaches out before the customer even notices. That's a substantial shift in experience delivered.

What’s UNRVLD’s ambition for where AI can take customer service?

Today's portals answer questions, and the ones we build do a great job of that. The next generation completes tasks. Agent orchestration means AI can receive customer intent and coordinate across your backend systems - billing, CRM, scheduling, fulfilment - to resolve it end to end.

Classic example: "Cancel my subscription and refund the last month." On today's portal, that's probably three screens and a phone call. With an agentic layer, the AI authenticates you, checks eligibility, processes the refund, cancels the subscription, sends confirmation — all in one go.

This can be done now. The question is whether your architecture will support it when you're ready?

Where should a business start if they want to enable their AI ambition for customer service?

Start with why customers are contacting you, do not start with the technology. Map the top contact reasons first and solve them. Then measure success by the reduction in calls and emails.

Build on clean, connected data. If your customer data is fragmented, your AI will be confidently wrong, and trust collapses fast.

Design for adoption. Regardless of how technically impressive your solution is, if it’s not designed with customers in mind, it will fail.

Plan for the agentic layer architecturally now. Composable, API-first platforms can absorb AI orchestration. Monolithic ones need a rebuild to get there.

What does this mean for customer service in the next three years?

Right now, AI is deployable with proven technology: AI-assisted search, conversational chatbots, personalised dashboards.

In 12 to 18 months proactive, predictive services will launch with AI agents reaching out before the customer realises they have a problem. Multi-modal interfaces, voice plus chat. Sentiment-aware escalation. And within three years, we can expect fully agentic self-resolution as described above.

For B2B companies we’ll see AI-to-AI buying. Gartner is forecasting that 90% of B2B purchasing will be AI-agent intermediated by 2028. Your portal won't just be serving human customers. It'll be serving other organisations' procurement agents. The window to build the right foundations before that arrives is shorter than it looks.

What is you key takeaway on the AI evolution within digital self-service?

Self-service is the strategy; AI is the enabler. The data foundation determines what AI can actually do, so the organisations who will see the strongest returns will have done the unglamorous work of connecting their data first.

The organisations best positioned for what's coming are fixing their foundations with clean data, connected systems and a composable architecture. This is the groundwork intelligent customer service will be built upon.

Drop us a message if you’d like to speak to Darren or one of our other experts about how UNRVLD can support your business to build the foundations for the AI service revolution.