Editorial
Research Report
The State of Enterprise Site Search: 2026 Benchmark Report
The Gist
- AI and CX separation creates customer friction. Organizations measuring AI deployment and customer experience independently often create fragmented journeys that weaken trust and retention.
- Workflow redesign beats AI bolt-ons. Research throughout the article points toward process redesign and unified ownership models as key differentiators for enterprise AI performance.
- Winning organizations govern AI inside customer outcomes. Enterprise leaders producing measurable value increasingly treat AI as part of journey ownership rather than a standalone technology program.
Here is something I keep seeing inside large enterprises, and it does not get said directly enough.
The CX leader is tracking NPS or any other metric and journey maps. The AI lead is tracking model deployments and use case velocity. Nobody is tracking what the actual customer experiences when both programs touch them at once. The gap between those two rooms is where revenue quietly walks out.
This is not a technology problem. It is an organizational fiction — the idea that you can run AI and CX as parallel strategies, hand them to different owners and still produce a coherent customer experience. You cannot. What you produce is two internal success stories and one confused customer.
I have been working in this space long enough to say what the data now confirms: the separation of CX and AI strategy is not a governance choice. It is a competitive disadvantage with a measurable price tag.AI Governance and Customer Experience FAQ
Editor's note: Key questions surrounding enterprise AI governance, customer experience ownership and organizational design.
Table of Contents
- 1. The Gap Is Not Between Adoption and Maturity. It Is Between Ownership and Outcome.
- 2. What 'Separation' Actually Looks Like in Practice
- 3. The IBM Watson Case: What $5 Billion of Separation Costs
- 4. The Companies Building It Right Are Not Separating Anything
- 5. Where the Human Remains Non-Negotiable
- The Decision in Front of You Right Now: Bind Together AI and CX
1. The Gap Is Not Between Adoption and Maturity. It Is Between Ownership and Outcome.
The numbers have been consistent for two years. McKinsey's November 2025 State of AI report, surveying 1,993 respondents across 105 countries, found that 88% of organizations use AI in at least one function — yet only 39% report any EBIT impact at the enterprise level, and for most, it sits below 5% of EBIT.
BCG's Widening AI Value Gap, published September 2025 after surveying 1,250 firms globally, is sharper still: 60% of companies generate no material value from AI despite sustained investment, and only 5% achieve substantial value at scale.
These are not technology failure numbers. The models work. The vendors deliver. The failure is structural — organizations deploying AI into functions without redesigning how those functions serve the customer.
McKinsey identifies the single highest-leverage fix: workflow redesign. High performers are 2.8 times more likely to have rebuilt workflows around AI than their peers. They are not asking AI to assist the old process. They are building a new one. That distinction — bolt-on versus built-in — is exactly the CX-AI separation problem described from a different angle.
Bain's 2025 B2B Commercial Excellence survey of 1,263 executives found that B2B revenue growth winners in 2024 deployed an average of 4.5 AI use cases versus 3.3 for laggards — and realized almost 2x the cost efficiencies per use case. The gap was not which AI tools they bought. It was how deeply those tools were integrated into customer-facing work.
2. What 'Separation' Actually Looks Like in Practice
I want to be specific, because the problem is usually invisible until it is expensive.
Separation looks like this: the AI program reports into the CDO or CTO. The CX program reports into the CMO or chief customer officer. Both have quarterly OKRs. Both show green dashboards. The B2B customer — a procurement director, a plant manager, a VP of Operations — contacts your company and hits three different interaction models depending on which channel they use. None of them know what the others have already told the customer.
Forrester's 2025 CX Index, measuring 275,000 customer perceptions across 469 brands, found that 25% of US brands declined significantly in CX quality for the second consecutive year, while only 7% improved. Forrester explicitly named "disappointing implementations of potentially game-changing technology, including AI" as a primary cause. They actually stated also: one in three companies will erode customer trust in 2026 through premature AI self-service — deployed under cost pressure before the journey design could support it.
And Gartner's March 2026 prediction — which most boards have not yet absorbed — states that over 50% of customer service organizations will double their technology spend by 2028 without an equivalent reduction in headcount. They will spend more on AI and keep the same number of humans, because they deployed AI on top of broken processes rather than inside redesigned ones. That is the financial signature of separation.
Related Article: Before You Scale AI in Customer Experience, Fix These 5 Things
3. The IBM Watson Case: What $5 Billion of Separation Costs
There is no better-documented B2B enterprise AI failure than IBM Watson Health. IBM spent over $5 billion in acquisitions and development building Watson into a platform positioned as the future of clinical decision-making for hospital and healthcare enterprise clients. The AI had its own product team, its own roadmap, its own marketing. It was sold as something that would sit alongside clinical workflows.
The results: Watson was trained on hypothetical cases, not real patient data. Its recommendations frequently contradicted local clinical guidelines. Adoption at enterprise clients fell far below projections. By 2022, IBM sold the Watson Health assets to Francisco Partners for approximately $1 billion — a retreat from a program that had consumed nearly a decade and multiple billions in capital.
The failure was not computational. Watson processed information. What it could not do was interpret that information in the context of an actual patient relationship, an actual clinical workflow, an actual human decision in real time. The AI was separate from the experience it was supposed to improve. That is the architecture of failure — and it is exactly what many enterprise organizations are quietly repeating today, at smaller scale, in their CX programs.
CallMiner's 2025 research found that 67% of organizations — including those with formal AI governance — implement AI without adequate structures for B2B account risk. For a mid-market SaaS provider with customer acquisition costs above $700, one account lost to an automated failure means years of renewal, expansion, and referral revenue gone.
4. The Companies Building It Right Are Not Separating Anything
The clearest evidence for convergence is not theoretical. It is operational, happening now across the regions and industries where the real work gets done:
- Germany / Manufacturing: Siemens deployed its Industrial Copilot — built on Microsoft Azure OpenAI Service — into the engineering workflows of over 100 enterprise customers including Schaeffler and thyssenkrupp Automation Engineering, which began a global rollout in early 2025. Engineers now create panel visualizations in 30 seconds and generate PLC code requiring only 20% manual adaptation. The AI is not adjacent to the customer experience. It is the work itself.
- Israel / Enterprise CX Technology: NICE reported a 400% increase in CXone Mpower Autopilot interactions in 2024. TD Bank Group eliminated 88 million minutes of customer wait time in a single year while contact volume grew 11%. That came from AI embedded as the primary operating layer — not a pilot.
- UAE / Enterprise Telecom: e& launched AI-powered human-digital advisors in October 2024, multilingual and built on 160+ machine learning models in production, designed for enterprise and government customer profiles. The AI does not support the customer strategy. It delivers it.
- Japan / Automotive: Toyota restructured its US Brand Engagement Center on Amazon Connect with generative AI handling call summarization and real-time agent assistance, reducing call transfers by 13%. The AI sits inside the agent workflow, not beside it.
- South Korea / Enterprise Telecom: SK Telecom deployed a Telco-LLM-powered customer service system that collapses post-interaction documentation previously consuming hours of agent time per shift. Speed, quality, and human retention all improved together.
- USA / Enterprise SaaS: Salesforce's own Agentforce deployment handles approximately 32,000 conversations per week with an 83% resolution rate and less than 1% requiring human escalation. The reason it works: Agentforce sits inside Service Cloud — unified ownership, one journey.
- USA / Enterprise Telecom: Verizon uses generative AI to accurately predict the reason behind 80% of incoming service calls, routing callers to the right human agent. This retained an estimated 100,000 customers in 2024 alone.
The pattern across every region and every industry is identical. AI is inside the journey. The people who own the customer outcome govern the AI. There is no separation.
5. Where the Human Remains Non-Negotiable
The argument for convergence is not an argument for replacement. Let me be direct about that too.
Gartner's June 2025 research found that 95% of customer service leaders plan to retain human agents to strategically define AI's role — and by early 2026, Gartner confirmed that only 20% of organizations had actually reduced agent headcount despite heavy AI investment. The prediction that AI would eliminate service roles has not materialized.
What has materialized is a shift in what human agents do: less documentation, less routing, more judgment, more relationship management on complex accounts. AI handles speed, context retrieval, and pattern recognition. Humans carry the relationship, the trust and the edge cases no model should handle alone.
Webinar
May
20
AI With Intent: How CX Leaders Decide Where Automation Belongs (And Where It Doesn't)
Senior CX leaders from REI & Hyatt reveal their framework for using AI to boost retention.
Register
Webinar
May
21
Content Strategy Leaders Live: Managing Scale, Safety & AI in Manufacturing
How manufacturing leaders update content systems to handle product complexity and scale while integrating AI safely.
Register
Webinar
May
27
Agents like AI! But Why? Find Out How to Implement, Build Trust and Deliver Results
Register
Webinar
Jun
3
From AI Investment to CX Results: What Enterprise Leaders Need to Know
Move beyond experiments. See how top enterprises scale AI for CX results.
Register
Webinar
Jun
4
Content Strategy Leaders Live: Managing Risk, Compliance & AI in Financial Services
Learn how financial services leaders are modernizing content systems without disrupting trust, compliance or experience.
Register
Webinar
Jun
16
From Legacy to Launch-Ready: How Gainbridge Made Its Website a Marketing-Led Growth Engine
Join in to learn how a D2C annuity brand gave marketing full website ownership — without slowing down or risking compliance.
Register
Gartner's survey of 5,728 customers found that 64% would prefer companies did not use AI in customer service at all — and 53% would consider switching to a competitor if they learned a company planned to implement it. They are not rejecting AI. They are rejecting AI deployed without accountability for the outcome it produces.
In B2B, where a single account can represent years of contracted value, the human layer is not a legacy cost. It is a risk management function. The companies treating human agents as a cost to eliminate are not running a more efficient CX program. They are running an untested liability.
Related Article: Forget Handle Time: Customer Satisfaction Is Now the Top AI Agent KPI
AI-CX Separation Patterns and Enterprise Outcomes
How organizational structure decisions influence AI effectiveness and customer experience performance.
| Pattern | What Happens | Business Impact |
|---|---|---|
| Separate AI and CX ownership | Technology teams and customer teams optimize independently. | Customers experience disconnected journeys. |
| AI layered onto old processes | Organizations automate existing workflows without redesign. | Technology spending rises without proportional value creation. |
| Journey redesign with AI embedded | Organizations rebuild operating processes around AI capabilities. | Higher efficiency and stronger measurable outcomes. |
| Unified accountability | Customer outcomes and AI governance share ownership. | Better coordination and reduced customer friction. |
| Human-plus-AI operating model | AI handles speed and pattern recognition while humans manage complexity. | Stronger trust and reduced operational risk. |
The Decision in Front of You Right Now: Bind Together AI and CX
Here is the question I ask every executive team, from day one.
Who in your organization is accountable for both the customer outcome and the AI that shapes it?
If the answer is two different people with two different scorecards, you have the separation problem. If one person owns the journey and a different person owns the model, something will fall between them — and your customer will notice before you do.
The fix is not a reorganization. It is a governance decision: CX and AI share a single accountability structure, a single measurement framework and one definition of what success looks like for the customer.
B2B customers in 2026 are not evaluating your AI. They are evaluating their experience. They do not know — and do not care — whether the response came from a human, a model, or a hybrid. What they know is whether it was fast, accurate, contextual and worth staying for.
The companies that understand this are not running two strategies. They are running one.
The ones still printing two decks are paying for the gap — whether it shows on a spreadsheet yet or not.
Learn how you can join our contributor community.