What Salesforce's CX Research Actually Tells Us in 2026

What Salesforce's Research Actually Tells Us About Customer Experience in 2026

Every year, Salesforce publishes one of the most widely cited pieces of CX research in the industry. The State of the Connected Customer report — now in its seventh edition — surveys tens of thousands of consumers and business buyers across dozens of countries. It is not perfect research, and Salesforce is not a disinterested party. But it is large, longitudinal, and honest enough to be useful. The question is whether the people reading it are drawing the right conclusions.

Most aren't. The report gets quoted selectively — a headline statistic here, a percentage there — stripped of the structural argument underneath. This article reconstructs that argument, tests it against what practitioners actually see, and draws out what it means for anyone building or leading a customer experience strategy in 2026.

What the "State of the AI Connected Customer" Report Is — and What It Isn't

The seventh edition of Salesforce's flagship CX report carries a new subtitle: State of the AI Connected Customer. The shift in naming is deliberate. Artificial intelligence is no longer a chapter in the report; it is the organising lens through which every other finding is interpreted. Trust, personalisation, data, service quality — all of it is now framed against the backdrop of AI adoption.

That framing matters for how you read the data. Salesforce sells AI-powered CRM and service software. The report's conclusions — that customers want AI-driven personalisation, that trust in AI is rising, that companies must accelerate AI integration — are not wrong, but they are shaped by the worldview of the organisation commissioning the research. Treat the findings as a serious signal, not gospel.

With that caveat stated plainly, the structural patterns in the data are worth taking seriously. Salesforce surveys at scale across geographies and sectors. The directional trends — even if the precise percentages should be held lightly — reflect something real about how customer expectations are shifting.

The Central Tension: Trust Is the Constraint on Everything Else

The most important finding in the seventh edition is not about AI capability. It is about trust. Salesforce's research consistently finds that a significant majority of customers say they trust companies less than they did in previous years, and that this erosion of trust is the primary barrier to customers engaging with AI-powered experiences.

This is a behavioural economics problem before it is a technology problem. Loss aversion — the well-documented tendency, identified by Daniel Kahneman and Amos Tversky, for people to weight potential losses roughly twice as heavily as equivalent gains — means that customers are far more sensitive to the risk of a bad AI interaction than they are motivated by the promise of a good one. A single experience where an AI assistant gives wrong information, fails to escalate appropriately, or feels manipulative does disproportionate damage to the relationship.

The trust deficit is not a communications problem. It is a design problem. Companies that treat trust as a marketing message rather than an operational output will keep losing it — regardless of how sophisticated their AI becomes.

The practical implication is that AI deployment in customer-facing contexts must be conservative in scope and transparent in operation before it can be ambitious in scale. Companies that lead with capability before establishing credibility are, behaviourally speaking, asking customers to accept a large potential loss in exchange for an uncertain gain. That is not a trade most people make willingly.

Personalisation: The Gap Between What Customers Expect and What Companies Deliver

Salesforce's research has tracked the personalisation expectation gap for several editions now. The pattern is consistent: customers increasingly expect companies to understand their context — their history, their preferences, their current situation — and to act on it without being asked. What they receive, in most cases, is segmentation dressed up as personalisation.

There is a meaningful difference between the two. Segmentation says: "You are in the 35–44 age bracket, so we will show you this offer." Personalisation says: "You called us three times last month about the same issue, so we will resolve it before you have to ask again." The first is a marketing convenience. The second is a genuine service act.

The Salesforce data suggests that customers are increasingly able to tell the difference — and that their patience for the former is diminishing. This connects directly to what the report frames as "connected experiences": the expectation that data shared in one channel or interaction is available and used in the next, without the customer having to repeat themselves.

For organisations in sectors with complex, multi-touchpoint journeys — banking and financial services, healthcare, telecommunications — this is where the gap is most painful. A customer who has completed onboarding, submitted documents, and spoken to three different agents should not have to re-explain their situation on the fourth call. That they routinely do is not a technology failure; it is a journey design failure.

What Salesforce's Research Reveals About AI in Service — and What It Misses

The seventh edition gives significant attention to AI in customer service: autonomous agents, AI-assisted human agents, predictive service, and proactive outreach. The data suggests growing customer openness to AI-assisted service — with important conditions attached.

Customers are more willing to interact with AI when the stakes are low, the task is transactional, and a human escalation path is clearly available. They are significantly less willing when the issue is emotionally charged, financially consequential, or involves a complaint. This is not surprising. It maps precisely onto dual-process theory: System 1 (fast, automatic, low-effort) is comfortable with AI for routine queries; System 2 (slow, deliberate, high-stakes) demands human judgment and accountability.

What the Salesforce report handles less well is the employee side of this equation. AI in service is not just a customer-facing technology; it changes the nature of the work that human agents do. When AI handles the routine, agents are left with the complex, the emotional, and the escalated. That is harder work, not easier. Organisations that deploy AI without redesigning the agent role — their training, their authority, their wellbeing — will find that service quality deteriorates precisely where it matters most.

This is why employee experience is not a separate workstream from CX transformation. It is the upstream condition on which downstream customer outcomes depend.

The Data Sharing Paradox — and Why It Matters for CX Strategy

One of the more nuanced findings in Salesforce's research concerns the relationship between data sharing and personalisation. Customers say they want personalised experiences. They also say they are uncomfortable sharing the data that makes personalisation possible. Both things are true simultaneously.

This is not hypocrisy. It is a rational response to an asymmetric exchange. Customers have learned, through repeated experience, that sharing data often produces targeted advertising rather than better service. The value exchange feels extractive rather than reciprocal. Until companies can demonstrate that data is used to serve the customer — not to sell to them — the paradox will persist.

The resolution is not better privacy communications or more granular consent flows, though those matter. It is demonstrating, through actual service acts, that data sharing produces tangible benefit for the customer. A bank that uses transaction data to flag a potential fraudulent charge before the customer notices has earned the right to ask for more data. A retailer that uses browsing history to show the same product in five different formats has not.

The data-sharing paradox resolves when the value exchange becomes visible. Customers do not object to being known; they object to being used.

Customer Experience Roles and Career Paths: What the Research Implies

Salesforce's research does not directly address the labour market for CX professionals, but its findings have clear implications for customer experience roles and customer experience career paths in 2026. As AI absorbs more of the transactional and analytical work in CX functions, the roles that remain — and the roles that grow — are those requiring judgment, empathy, and strategic thinking.

The profile of a CX leader is shifting. Technical literacy matters more than it did five years ago: understanding how AI systems work, what their failure modes are, and how to design human-AI handoffs is now a core competency, not a specialist skill. So is the ability to translate customer insight into business cases that finance teams will fund.

For those building or developing CX teams, a few structural shifts are worth noting:

• Journey ownership is becoming a dedicated function. Organisations that treat journey management as a project rather than a permanent role consistently underperform on consistency and follow-through.
• Voice of Customer is moving upstream. The most effective CX teams are embedding customer feedback into product and service design decisions, not just measuring satisfaction after the fact. This requires analysts who can connect VoC data to operational levers — a genuinely scarce skill.
• CX governance is gaining board-level attention. As customer experience becomes a measurable financial driver rather than a soft aspiration, accountability structures are formalising. Chief Experience Officers, CX councils, and cross-functional governance models are becoming standard in mature organisations.
• Behavioural design is entering the mainstream. The ability to apply behavioural economics principles to journey and service design — not just to marketing — is increasingly valued in senior CX roles.

For professionals navigating customer experience salary negotiations in 2026, the premium is on this combination: strategic thinking, data fluency, and the ability to drive cross-functional change. Generalist CX coordinators face more pressure from automation; specialists who can design, measure, and govern experience at a system level do not.

Customer Experience Certifications and Learning: What Actually Builds Capability

The Salesforce report does not address professional development directly, but the capability gaps it implies are instructive for anyone thinking about customer experience certifications or structured learning.

The most common mistake organisations make in CX training is treating it as content delivery — a course to complete, a certificate to earn — rather than capability building. A practitioner who has memorised the NPS formula but cannot explain why a particular touchpoint drives churn has not developed useful capability. The test is always application, not recall.

The most valuable learning programmes for CX professionals in 2026 share three characteristics. First, they are grounded in real journey work — participants map, score, and redesign actual customer journeys rather than hypothetical ones. Second, they integrate behavioural economics as a practical lens, not an academic module. Third, they connect CX metrics to business outcomes in language that non-CX stakeholders understand.

For organisations that want to build this capability internally rather than relying on external certification, bespoke training programmes designed around the organisation's own journeys and data tend to produce faster and more durable results than off-the-shelf courses.

What the Research Means for Specific Sectors

The Salesforce data is global and cross-sector, which means its implications vary considerably depending on where you operate. A few sector-specific readings are worth making explicit.

Banking and financial services face the sharpest version of the trust challenge. Financial decisions are high-stakes, emotionally loaded, and — in the customer's mind — irreversible. AI adoption in this sector must be paced against trust-building, not ahead of it. The institutions that will win are those that use AI to make human advisers more effective, not to replace them in moments that require genuine accountability. The CX dynamics in banking are distinct enough to warrant sector-specific strategy.

Retail and e-commerce are further along the AI adoption curve, and the Salesforce data reflects this. Customer tolerance for AI in retail is higher because the stakes are lower and the failure modes are more recoverable. The frontier here is not whether to use AI, but how to use it to create genuinely differentiated experiences rather than operational parity.

Public services and government face a different version of the personalisation paradox. Citizens are often the most data-rich constituents — governments hold more information about individuals than most private companies — but the least likely to experience personalised service as a result. The gap between data held and value delivered is a significant opportunity, and a significant accountability risk.

The Structural Argument Salesforce Is Making — and Whether It Holds

Beneath the individual findings, Salesforce's research advances a consistent structural argument: that the future of customer experience is AI-powered, data-connected, and trust-dependent, and that companies which cannot integrate these three elements will lose customers to those that can.

This argument is largely correct, with one important qualification. Technology is a necessary condition for competitive CX in 2026, but it is not a sufficient one. The organisations that consistently outperform on customer experience — across sectors, geographies, and economic cycles — are those that have made customer centricity a cultural and operational reality, not just a technological investment.

Salesforce's research is well-placed to identify what customers say they want. It is less well-placed to explain why so many organisations fail to deliver it despite having access to the same technology. The answer to that question lives in how leaders signal and sustain customer centricity through daily decisions — not in the sophistication of their CRM stack.

Understanding customer experience at a structural level means understanding that technology enables but culture determines. A company with mediocre tools and a genuine commitment to its customers will consistently outperform a company with excellent tools and a customer-centricity that exists only in its brand guidelines.

Using the Research Well: A Practitioner's Reading Guide

If you are using the Salesforce State of the AI Connected Customer report to inform your CX strategy, here is how to read it with appropriate rigour:

1. Treat directional trends as more reliable than precise percentages. Large-scale surveys are good at identifying the direction and rough magnitude of shifts in customer expectation. They are less reliable on exact figures, which vary by methodology, sample, and question framing.
2. Localise before you act. Global averages mask significant regional and cultural variation. Customer expectations around AI, data sharing, and service norms differ materially between, say, the Gulf and Western Europe. Validate global findings against local data before building strategy on them.
3. Separate the signal from the sales narrative. Salesforce has a commercial interest in AI adoption. That does not make its research wrong, but it does mean you should read the AI-related findings with particular care, and triangulate against independent sources.
4. Use it to start conversations, not end them. The most valuable use of this research is as a prompt for internal inquiry: "The data suggests customers want proactive service — where in our journeys are we reactive when we could be proactive?" That question is worth more than any statistic.
5. Connect findings to your own VoC data. A robust Voice of Customer strategy gives you the local, specific evidence to confirm, challenge, or contextualise what the Salesforce report suggests at a global level.

If you want to understand where your organisation currently stands relative to the capability expectations the research implies, a structured CX maturity assessment is a useful starting point — it maps your current state across the dimensions that the Salesforce data suggests matter most: personalisation, trust, data integration, and service design.

The Real Lesson from Seven Years of Connected Customer Research

Salesforce has now been publishing this research for seven editions. The most striking thing about reading them in sequence is not how much has changed, but how much has not. The expectation gap — the distance between what customers want and what companies deliver — has been documented consistently across every edition. The technologies cited as solutions have changed; the gap itself has narrowed only modestly.

That persistence is the real finding. It tells us that the problem is not primarily informational — organisations know what customers want. It is not primarily technological — the tools to deliver it have existed for years. It is organisational: the structures, incentives, and cultures that determine what actually happens at the moment of truth.

Seven years of data and the gap is still there. The question is no longer what customers want. It is why organisations keep failing to deliver it — and whether you are willing to look honestly at the answer.

The Salesforce report is a useful mirror. What it reflects back is a consistent picture of rising expectation and uneven delivery. The organisations that close that gap are not the ones with the best technology or the most sophisticated research subscriptions. They are the ones that have made the hard organisational choices — about accountability, about investment, about what customer centricity actually requires of leadership — that the research never quite captures but always implies.

That is the work. The report tells you where to look. What you find when you look honestly is up to you.