The contact center stories enterprise IT leaders actually read.
Volume 343, Issue 3 · 8 articles from 2 issues
Personal AI agents from Meta, OpenAI and SpaceX are already contacting businesses through channels built for humans.
Personal AI agents such as Meta's Muse, OpenAI's Dots and shopping features in SpaceX's Grok Bot arrived in the last month, and retailers like Amazon have already blocked some of them.
People are using these agents to reach businesses through voice, web chat and websites designed for humans, so contact centers and digital commerce teams now have to decide how to identify an AI agent, determine what it is authorized to do and respond to it.
A Genesys executive says customers report agents arriving through human channels, though volume is still hard to measure.
The article outlines the IT stack questions this raises for CIOs and CX leaders.
Bots that misunderstand customers drive demands for live agents; experts explain the causes and fixes for CX leaders.
CX expert Brad Cleveland says a bot can understand every word a customer says and still miss the point, because customers conflate issues, use sarcasm or omit context.
Outdated knowledge bases make the problem worse, so organizations must keep source content current and accurate.
He recommends testing with real, frustrated customers in noisy settings rather than polite, scripted ones.
CX leaders should also measure failures and make escalation to a human easy, which can increase trust in AI self-service.
Context lost between bots and human agents reflects integration gaps and siloed service teams, analysts say.
Generative and agentic AI handles basic support well but struggles with complex issues, often because of call flow and operational design.
When channels are not integrated, context is lost and customers must repeat themselves after reaching a human, so Valoir's Rebecca Wettemann recommends enabling session IDs to reconcile conversations across email, chat and voice.
Many service organizations have separate teams per channel, which complicates handoffs.
Experts urge a review of how the service function is structured and a move to an omnichannel approach.
CX is shifting from reactive inbound service to proactive outreach as AI blurs the lines between CRM and contact center platforms.
The author recounts the long fight between CCaaS and CRM vendors over who owns the agent desktop.
That line is now blurring as Salesforce launches Agentforce Contact Center and AI makes SaaS seats easier to replace.
A bigger change is the move from inbound reactive service to proactive engagement by phone, text, reminders and in-app notifications.
Because customers compare every company against the best experience they've had anywhere, the article says, businesses must take the initiative to reach out first if they want to keep strong relationships.
Disconnected back-end systems still undermine customer service and AI agent rollouts, and only 5% of firms have fully integrated customer functions.
Many companies still have not connected their back-end systems, making it harder to deliver seamless service, understand customers and roll out AI agents.
Signs of siloed data include customers being passed between agents, getting different answers to the same question, or receiving marketing after they cancel, and if the service platform isn't integrated with billing, shipping, returns and CRM, neither human nor AI agents can see what the customer just did.
A KPMG survey found only 5% of companies have fully integrated marketing, sales and service.
Acxiom research shows brands want AI to orchestrate interactions but struggle with accountability and consistency.
Using AI only to cut service costs breaks trust; mass-market customers want low effort plus a clear handoff to a human.
The author argues that framing customer service AI around efficiency misses both the risk and the opportunity for brands with large, price-sensitive customer bases.
Mass-market customers do not want less service; they want less effort, and AI excels at fast self-service for simple requests, which Zendesk says 51% of consumers prefer for immediate help.
Problems start when automation becomes an endpoint that follows rules instead of producing good outcomes, given that 95% of consumers say service affects loyalty.
Humans must own outcomes, and AI should recognize its limits and hand complex, emotional or high-stakes cases to a person.
Aspirational CX strategies have relied on employee judgment, and AI agents need explicit trade-off rules that expose those gaps.
Most CX strategies use inspiring statements, such as delighting every customer, rather than guidance for consistent decisions, and experienced employees have quietly filled the gaps by balancing satisfaction against policy and fairness against revenue.
AI agents cannot infer meaning from aspirational language and need explicit instructions on trade-offs, such as what delight means, which customers get priority, and when customer value outweighs margin.
The analyst urges firms to separate values from executable strategy and to test that strategy across people, processes and machines.
Agent deployment should be treated as a chance to tighten governance and accountability.
ICMI co-founder Brad Cleveland says contact center AI helps, but most leaders report workloads rising, not falling.
Ahead of ICMI's Contact Center Expo, ICMI founding partner Brad Cleveland says contact centers sit somewhere between vendor hype and reality.
AI delivers real value in summarization, agent assistance, knowledge retrieval, translation and routine interactions, but most organizations are still working out integration, quality safeguards and staff comfort.
When he polls audiences, most leaders say workloads are increasing.
He cites unmet demand surfacing as access improves, growing product complexity, more channels and more sophisticated fraud, plus a self-service paradox: the more routine work automation absorbs, the more complex the work left for humans.
More from Call Centers · All IT section newsletters · This week