Executive Summary. Vasili Triant explains why AI is not replacing enterprise systems but eliminating redundant CRM layers as the stack shifts toward real-time orchestration and unified agent workflows.
Enterprise customer experience is entering a structural transition as AI moves from front-end automation to real-time orchestration across systems. The question is no longer whether AI will replace existing software, but which layers of the enterprise stack remain necessary.
In this conversation, UJET CEO Vasili Triant outlines how agentic AI is reshaping the experience stack by shifting the center of gravity away from traditional CRM systems toward real-time data and orchestration layers. Drawing on experience leading contact center platforms at scale, he explains why static ticketing systems are becoming redundant, how enterprises are consolidating fragmented AI pilots, and why the future of customer experience depends on simplifying architecture while strengthening human-led interactions.
Vasili, there’s a growing narrative that AI will replace enterprise software — particularly CRM systems. You’ve argued that’s both true and misunderstood. What do you mean by that?
AI is absolutely going to reshape enterprise software. What’s misunderstood is how that happens. AI won’t suddenly replace SaaS tomorrow. It’s that AI changes which layers are still necessary.
Parts of the CRM, like case management and ticketing systems, were built in an era when the application was a static system of record. Data lived inside the tool. Customer service interactions were in an entirely different silo from ticketing and case management - and the tools did not connect seamlessly. This is the main reason why customer service is so slow and fragmented.
To attempt to solve this customer service problem, most enterprises centralized their data in lakes or customer data platforms to better understand customer needs in real time. That shifts the center of gravity away from the CRM.
When AI can leverage customer conversations as real-time context, read and write directly to modern data environments, and orchestrate actions across systems, legacy CRM workflow layers start to look redundant.
So yes, AI will reduce reliance on certain enterprise software layers. But the real transformation isn’t about ripping and replacing. It’s about simplifying the stack and moving from static systems of record to real-time agentic AI experience orchestration.
If agentic AI can now reason, route, summarize, and act in real time, what specific layers of traditional enterprise software become redundant?
The layers that potentially become redundant are the static ticketing and case management platforms sitting at the center of the customer service agent desktop. Agents toggle between four to ten tools per interaction, manually updating tickets, copying context, and stitching systems together.
When agents put customers on hold, it’s not because they are trying to be difficult, but that their legacy systems are challenging to manage. They might first go to the CRM to understand customer context from a previous interaction, but that does not have the real-time context. Customers will have already been triaged by AI self service, again lacking context or personalization from previous interactions. Agents must then reauthenticate customer identity and ask what they are calling for again, because their tools do not provide the real time information and guidance. This leaves customers frustrated and agents scrambling to clean up the mess from new AI tools bolted onto their legacy systems.
Agentic AI can help unify those systems into a single workspace, automating procedural tasks, and executing actions across tools, potentially making disconnected legacy tech layers unnecessary. .
What parts of the enterprise stack are fundamentally irreplaceable — even in an AI-native future?
The foundation of the enterprise remains. You still need governed data environments. You still need real-time communications systems, billing systems, fulfillment systems, underwriting systems — the engines that actually run the business. And you absolutely still need humans because humans are the ones that build trust, loyalty, relationships, and ultimately lifetime value with your consumers. The companies that invested in AI to cut costs by replacing humans are starting to see that the AI as a human replacement strategy is a failure. Companies either haven’t reduced headcount, haven’t properly accounted for the TCO of AI solutions, and in some cases, are starting to hire back human contact center agents.
AI should be focused on automating low-value tasks and providing contextual guidance in real time. But when loyalty, trust, empathy, and revenue are at stake, human connection is still irreplaceable.
The goal isn’t automation for automation’s sake. It’s using AI to strengthen human-led relationships — not replace them. AI should be doing the chores so humans can focus on building relationships.
AI should sit between customer conversations and enterprise systems, providing context and executing tasks while humans focus on relationships.
Vasili Triant
As AI agents become autonomous actors inside customer workflows, how should enterprise teams rethink their architecture?
They need to move from an automation mindset to an orchestration mindset. For the past few years, the focus has been on front-end virtual agents deflecting customers away from humans and attempting to solve problems autonomously. This has failed. Layering automation on top of fragmented systems doesn’t fix the underlying bottleneck; it accelerates bad interactions at scale.
The shift now is architectural: unify data, streamline the human agent experience, automate cross-system workflows, and keep humans in the loop where judgment and oversight matters. AI should sit between customer conversations and enterprise systems — providing context at every step of the journey and executing tasks in the background so human agents can focus on solving problems - not fumbling with 10+ applications on their desktop.
That’s a fundamentally different approach than just adding another bot.
As a preferred Google CX partner, what shifts are you seeing inside large enterprises as they evaluate AI-driven customer experience platforms?
The biggest shift is discipline. Enterprises are moving from experimentation to consolidation. Finance and legal are in the room. Leaders are asking: what redundant systems can we eliminate? Where is the measurable ROI? Can we scale operations and revenue without expanding headcount?
Many organizations are running multiple AI pilots today. But over the next year, most will consolidate down to fewer platforms that deliver real architectural simplification and actual ROI from legacy system elimination.
There’s also growing recognition that deploying front end virtual agents to solve every problem doesn't create positive customer experiences. If agents are still buried in multiple tools, the experience remains fragmented. The conversation is shifting from hype to outcomes.
Five years from now, what does the modern “experience stack” look like — and which layers survive this transition?
Five years from now, the experience stack is simpler and more unified. At the foundation is a governed data layer — centralized, secure, and AI-ready.
Above that sits an orchestration layer that leverages customer conversations as real-time context, coordinates workflows across systems, and enriches human-led interactions. Specialized systems that execute the business remain. What shrinks are the redundant ticketing and workflow layers that primarily exist to manage records and manual processes.
And at the center are supercharged agents — working from a single workspace, equipped with context, and empowered by AI to drive resolution, loyalty, and revenue. AI doesn’t win by replacing people. It wins by removing friction around them.
Are investors beginning to distinguish between shallow workflow SaaS and core operational infrastructure — and what does that mean for the future of CX technology?
The industry is finally waking up from a decade-long hypnosis. For the last five years, investors were throwing money at anything with an ".ai" suffix. We saw a massive wave of shallow workflow SaaS—basically, pretty digital paint jobs on top of the same old, broken foundations. But people are starting to ask, "If we’ve spent millions on these tools, why are my agents still toggling between 10 tabs and why is my CSAT still in the gutter?"
The future of CX isn't about firing all your people to save a buck; it’s about using technology as to provide more contextual, personalized, empathetic, and human responses. It’s about agentic workflows that handle the back-office clutter so the 85% of people who still want a human connection actually get a good one.
How should leaders think about data ownership, privacy, and system design when AI operates across multiple SaaS platforms in real time?
When AI operates in real time across platforms, you need Privacy by Design—utilizing decentralized data models so that sensitive information is processed locally or ephemeralized, rather than stored in every connected cloud.
Seamlessness cannot come at the cost of transparency. If the AI’s cross-platform journey is a black box, you aren't building a solution; you're building a liability. The goal is Zero-Friction Trust with observability for every step of the AI process.
If you were advising a CIO today, what is the first architectural question they should ask before layering AI into their existing stack?
What does this AI actually do to eliminate costly, complex back office systems and remove tools from the agent desktop? Or, are we just chasing human replacement and forcing our agents to use yet another application and act as the bridge between our siloed systems?
