UJET\u2019s latest report reveals that agents\u00a0remain\u00a0wary of\u00a0AI\u2019s\u00a0accuracy<\/a>, context, and real\u2011world usefulness,\u00a0as 93% do not fully trust AI outputs at face value.\u00a0despite it now being embedded in\u00a0nearly every\u00a0customer interaction.\u00a0<\/p>\nAs a result,\u00a0AI has been deployed faster than the data foundations, workflows, and system architectures\u00a0required\u00a0to make it truly reliable at the frontline.\u00a0<\/p>\n
Speaking with CX Today,\u00a0Vasili\u00a0Triant, CEO of UJET,\u00a0argues that\u00a0agents\u00a0don\u2019t distrust\u00a0AI because they resist change, but because poorly designed, data\u2011fragmented systems\u00a0haven\u2019t\u00a0earned their trust.\u00a0<\/p>\n
\u201cThe 93% verification rate reflects a system design problem, not a\u00a0behavior\u00a0problem,\u201d he explained.\u00a0<\/p>\n
\u201cWhen AI is layered onto fragmented data environments, the outputs reflect poor AI implementation.\u201d<\/p>\n
Why Accuracy Breaks Down at the Frontline<\/p>\n
In enterprise environments, AI tools are\u00a0frequently\u00a0layered onto fragmented data sources, where customer records, interaction histories, and real-time signals are split across multiple systems.\u00a0\u00a0<\/p>\n
As a result, this creates an incomplete operational picture,\u00a0likely leading\u00a0to\u00a0structurally\u00a0inaccurate\u00a0outputs,\u00a0particularly\u00a0when models generate responses without synchronized context.\u00a0\u00a0<\/p>\n
From here, the architecture issue can become more pronounced when AI systems lack access to real-time customer context,\u00a0resulting in\u00a0models\u00a0that\u00a0produce\u00a0partially outdated or inconsistent\u00a0responses that\u00a0don\u2019t\u00a0align\u00a0with the present situation.\u00a0\u00a0<\/p>\n
That gap\u00a0increases the\u00a0likelihood\u00a0of hallucination\u00a0and\u00a0misalignment between recommendations\u00a0and customer needs, as\u00a015% of agents\u00a0agree that\u00a0real-time AI recommendations\u00a0are\u00a0unreliable or inaccurate,\u00a0as well as\u00a054% saying AI is helpful but lacks sufficient context and depth.\u00a0<\/p>\n
\u201cWhen AI does not have access to the latest real time data on the customer\u2019s latest touchpoint, the risks for AI hallucinating responses skyrockets,\u201d\u00a0Triant\u00a0continued.\u00a0<\/p>\n
\u201cAgents have seen enough incorrect answers and context-free suggestions to know that blind trust creates risk for the customer experience.\u201d\u00a0<\/p>\n
As a result, many organizations interpret this as a\u00a0behavioral\u00a0issue requiring more user training or compliance when the underlying constraint is architectural friction.\u00a0\u00a0<\/p>\n
When verification requires excessive time or effort, users may typically default to\u00a0skepticism\u00a0because the cost of error is high in live customer interactions.\u00a0\u00a0<\/p>\n
\u201cThe goal\u00a0shouldn\u2019t\u00a0be\u00a0eliminating\u00a0verification,\u201d explained\u00a0Triant.\u00a0<\/p>\n
\u201cIt should be making it effortless. When an agent can validate an AI recommendation in two seconds instead of twenty, quality stays high and efficiency finally becomes real.\u201d<\/p>\n
How Early Assumptions Shaped Today\u2019s Friction<\/p>\n
Secondly, the agent-AI trust gap is also shaped by how the technology\u00a0was initially positioned and deployed inside enterprise systems, with\u00a0the common\u00a0early industry narrative focusing\u00a0on automation, deflection, and cost reduction.\u00a0\u00a0<\/p>\n
From here, that framing\u00a0encouraged organizations to treat AI as a replacement layer placed on top of existing operations,\u00a0meaning\u00a0that\u00a0AI tools were introduced into environments that were already operationally fragmented.\u00a0\u00a0<\/p>\n
Instead of\u00a0redesigning the underlying workflow, the\u00a0result was limited improvement in day-to-day effectiveness, even when the models themselves performed well in isolation.\u00a0<\/p>\n
In fact, around 78% of agents reported that their AI tools are not transformative, while 81% are\u00a0required\u00a0to manage more than four tools during a single interaction, and\u00a0nearly 20%\u00a0handle seven or more tools at once.\u00a0\u00a0<\/p>\n
Despite these challenging conditions, 93% agreed they could still do their job without AI,\u00a0implying that\u00a0the tools are not yet integrated into a coherent workflow that changes core execution.\u00a0\u00a0<\/p>\n
However, success metrics\u00a0are still\u00a0frequently\u00a0defined in terms of reduced staffing rather than improved workflow\u00a0performance,\u00a0and\u00a0do not address the structural inefficiencies agents face.\u00a0<\/p>\n
\u201cFor years, the industry framed AI\u2019s value around deflection and headcount reduction,\u201d he explained.\u00a0<\/p>\n
\u201cThat pushed investment toward the wrong outcomes and set unrealistic expectations for ROI.\u201d\u00a0<\/p>\n
With AI\u00a0originally\u00a0frequently\u00a0layered onto workflows that were never designed for real-time intelligence or cross-system coordination, many agents today still move between disconnected tools, re-entering or reconciling information across systems that\u00a0lack\u00a0a unified context.\u00a0\u00a0<\/p>\n
Even\u00a0accurate\u00a0AI outputs lose value when they cannot be applied directly within the flow of work.\u00a0<\/p>\n
\u201cEnterprises also share responsibility. AI is often deployed on workflows that were designed decades ago,\u201d\u00a0Triant\u00a0continued.\u00a0<\/p>\n
\u201cWhen agents are still juggling five or six disconnected systems, even good AI struggles to deliver value.\u201d\u00a0<\/p>\n
This enables enterprises to\u00a0shift the ROI discussion toward\u00a0a meaningful measure of AI performance, evaluating how effectively it reduces friction for existing teams\u00a0for\u00a0fewer context switches, faster resolution paths, and more consistent decision support during live interactions.\u00a0<\/p>\n
The Emotional Cost of Failed Self\u2011Service<\/p>\n
As a result,\u00a0the breakdown in trust around AI-powered self-service is\u00a0largely a\u00a0result of how these systems are structured,\u00a0as many organizations focused on\u00a0containment rather than resolution.\u00a0\u00a0<\/p>\n
When many implementations\u00a0were\u00a0designed to keep customers within automated flows for as long as possible, this technique\u00a0optimized\u00a0for deflection rates over\u00a0successful outcomes, creating a structural gap.\u00a0\u00a0<\/p>\n
When a tool\u00a0is\u00a0capable of\u00a0answering\u00a0questions\u00a0but struggles to\u00a0execute\u00a0changes, it pushes unresolved cases forward with increasing frustration attached to them, with\u00a0roughly\u00a065%\u00a0of customers reporting\u00a0frustration when\u00a0having to\u00a0repeat information after moving from AI to a human agent.\u00a0\u00a0<\/p>\n
In fact,\u00a014% of agents say they now handle more emotionally charged interactions as a direct result of failed self-service, as\u00a0AI is already present in more than three-quarters of customer interactions for 75% of agents.\u00a0<\/p>\n
\u201cMost self\u2011service is designed around containment \u2013 keeping customers away from humans,\u201d said\u00a0Triant.\u00a0<\/p>\n
\u201cBut customers aren\u2019t trying to avoid agents. They\u2019re trying to solve a problem, and they feel that disconnect immediately.\u201d\u00a0<\/p>\n
Moreover, when self-service systems cannot take meaningful action, they\u00a0tend to pass unresolved issues downstream, often after the customer has already repeated steps or provided information multiple\u00a0times.\u00a0<\/p>\n
\u201cWhen self-service can only answer questions and can\u2019t execute actions, it escalates the most emotionally charged cases,\u201d he explained.\u00a0<\/p>\n
\u201cCustomers arrive at agents frustrated and forced to start over.\u201d\u00a0<\/p>\n
However, a\u00a0more effective architecture treats self-service as part of a continuous, stateful journey\u00a0where\u00a0context is preserved across\u00a0transitions\u00a0so agents receive the full history of what has already been\u00a0attempted.\u00a0\u00a0<\/p>\n
For the customer, this\u00a0reduces repetition\u00a0friction\u00a0and allows the human interaction to begin at the point where automation left off.\u00a0<\/p>\n
\u201cSelf-service should be the first chapter of a continuous journey, not a dead end,\u201d emphasized\u00a0Triant.\u00a0\u00a0<\/p>\n
\u201cIf escalation happens, the agent should inherit full context so the experience feels like a continuation, not a restart.\u201d<\/p>\n
Ultimately, the\u00a0disconnect between AI adoption and frontline trust signals that enterprise architecture has been incapable of keeping up\u00a0with CX ambition.\u00a0\u00a0<\/p>\n
As AI becomes embedded across customer journeys, its\u00a0customer\u00a0impact\u00a0will be\u00a0dependant\u00a0on\u00a0how effectively it is grounded in real\u2011time data, integrated into coherent workflows, and aligned with how agents\u00a0work.\u00a0\u00a0<\/p>\n
If an AI\u00a0is able to\u00a0reduce cognitive load, preserve context, and remove operational friction,\u00a0the customer\u00a0trust\u00a0will\u00a0naturally\u00a0follow.\u00a0\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"The growing\u00a0disconnect\u00a0between AI\u00a0investment\u00a0and frontline trust is becoming more\u00a0evident\u00a0across customer service organizations, despite rapid adoption and daily use.\u00a0 UJET\u2019s…\n","protected":false},"author":2,"featured_media":48160,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[179,509,405,7537,513,28729,16544,28730,28288,28731],"class_list":["post-48159","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-agentic-ai","tag-agentic-ai-in-customer-service","tag-ai-agents","tag-artificial-intelligence-agents","tag-autonomous-agents","tag-customer-self-service-software","tag-data-management-software","tag-identity-verification-software","tag-self-service","tag-trust-safety"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/48159","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=48159"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/48159\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/48160"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=48159"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=48159"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=48159"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}