CX budgets are growing and practitioners are laser-focused on building their agentic AI, automation and data insights capabilities. At least that's how the last 12 months have played out. Next year, the experts say practitioners should focus on AI governance and their people.
Therefore, this article offers four spending tips and 11 questions CX practitioners can put to their prospective vendors. It covers:
- How have CX practitioners invested in CX in recent years?
- Spending tip 1: Invest in new roles and new people
- Spending tip 2: Less AI more governance
- Spending tip 3: Start working toward an integrated CX tech ecosystem
- Spending tip 4: Know what to ask of prospective vendors
This is the third article of our 2027 outlook series. Via these links you can catch up with part one, 9 ways CX will change in 2027, and part two, 6 AI challenges for CX to tackle in 2027.
For more insights into how practitioners can invest to drive impact, CX Network also has these resources:
- Where is CX driving value? Communicate and deliver impact through culture, technology, metrics and savvy investment
- How practitioners are spending their CX budgets in 2026
- Balancing the books on AI: How to deliver a return on your investment
How have CX practitioners invested in CX in recent years?
When CX Network conducted its latest research into how practitioners are investing in CX, the results showed that since 2022, there has been a clear migration from low to mid-size budget brackets (from US$250,000 to $1 million), while the number of practitioners reporting budgets of $1 million and more has also edged up year-on-year, rising from nine percent in 2022 to 11 percent in 2026.
What's more, practitioners have been spending heavily on AI. We asked the research respondents to tell us their top three investment priorities for the year ahead and the graphic below shows the 10 most selected responses.
The research also confirmed that AI spending was on track to increase significantly this year across all categories. On this point the survey found:
- 46 percent of respondents planned to increase spending on generative AI while eight percent expected it would decrease.
- 40 percent planned to increase spending on agentic AI and AI agents, with 11 percent anticipating a lower spend.
- 32 percent planned to increase spending on AI regulatory compliance, with nine percent expecting a decrease.
- 39 percent planned to increase spending on AI projects for CX and service, while 11 percent expected a decrease.
However, as we prepare for 2027, the experts say CX should stop chasing new tools and focus on the foundations: the governance that upholds responsible and ethical AI and the people who drive CX standards.
Spending tip 1: Invest in new roles and new people
We all know people drive experience and as the next generation of CX tools is embedded in organizations many of those people have seen – or will see – their roles change, particularly those on the front line of customer service.
Jeannie Walters, CEO of Experience Investigators and author of Experience is Everything, says that to support better customer outcomes, organizations should invest in their greatest assets… their frontline people.
"If we don't invest in the people who serve customers – and support their well-being – technology investments won't be optimized," Walters says.
With AI now capable of automating the routine, Walters says the human workers who pick up the rest will be handling "challenging, difficult and taxing situations."
"We need to support them with education, empowerment and emotional awareness. People are not robots, and we need to protect them from burnout so they can deliver the best for our customers and our organizations," she says.
Bill Staikos, founder and managing partner of Be Customer Led, takes a broader view. He says CX has changed "less than the technology conversation around it would suggest" and that roughly half of CX teams are still "heavily dependent" on surveys. In fact, only the progressive organizations are beginning to combine behavioral, operational, conversational, transactional and employee signals to understand the true customer experience.
"That is progress, but simply putting more data into the same old CX operating model will not produce a fundamentally different outcome," he cautions.
In 2027, Staikos hopes to see the emergence of the CX engineer: someone who can work across customer data, AI, workflows, journeys, automation, experimentation and measurement to deliver lasting impact.
"CX teams should begin building around people who can diagnose a problem, redesign the underlying system, deploy an intervention and prove whether it improved the customer and business outcome," he says. "That would give CX a very different kind of influence inside the company, because the function would increasingly create change rather than primarily identify where change is needed."
Spending tip 2: Less AI more governance
As the research shows, CX has invested heavily in AI in recent years – and with so many new capabilities on the market, it's easy to see why.
However, Hemang Upadhyay, a product and AI leader and CX Network Advisory Board member, says that where AI agents are concerned, it is time to focus on the infrastructure that makes them trustworthy at scale
He tells CX Network: "As AI agents move from answering questions to taking actions – changing orders, applying policies, issuing credits, making recommendations or coordinating across systems – the critical CX question changes. It is no longer simply, 'Did the AI give the right answer?' It becomes, 'Was this the right decision, based on the right context, with the right authority – and can we recover quickly if it was wrong?'"
To support this, Upadhyay says organizations need a "customer decision layer" that connects four capabilities:
1. Shared customer context.
2. Explicit decision rights.
3. Real-time observability.
4. Recovery mechanisms.
He explains: "Every consequential agent action should have a clear owner, an explainable source of truth, an escalation path and, wherever possible, a way to reverse or correct the outcome without forcing the customer to start over."
Upadhyay says this also changes how CX success should be measured. Containment and automation rates are useful, he says, but incomplete. Instead, leaders should measure decision quality, recovery quality, repeat-failure rates and whether the customer's intended outcome was actually achieved.
The companies that differentiate in 2027 will not necessarily be those with the most agents.
"They will be those that can safely give AI greater responsibility while preserving customer context, human accountability and trust. That foundation is what turns agentic AI from an impressive interface into a durable CX capability," Upadhyay says.
Watch Upadhyay talking about what to do when an AI agent is wrong in this video from Mastering CX: AI Agents.
LINK TO HIS MASTERING CX SESSION
For more on AI governance in CX, read this article from consultant and lecturer Sue Duris.
Spending tip 3: Start working toward an integrated CX tech ecosystem
Ekaterina Mironova, CX program manager for Uber, says that in 2027 – and beyond – the most important investments won't be individual tools but an integrated CX tech ecosystem.
She says CX leaders should prioritize:
- AI-powered Voice of Customer platforms that analyze conversations, surveys and reviews and support interactions at scale.
- CDPs that create a unified profile by combining behavioral, transactional and sentiment data.
- Journey orchestration platforms that trigger personalized experiences based on real-time customer behavior.
AI agents, she says, are "equally important" as they can resolve increasingly complex customer requests, enterprise knowledge management systems that ensure both employees and AI always provide accurate and consistent information, predictive analytics that identify churn risk and growth opportunities early, and strong AI governance to ensure insights remain transparent, reliable and actionable.
"The companies that lead in CX won't necessarily have the most AI. They'll be the ones that combine trusted data, intelligent automation and human judgment to turn customer insights into business growth," she adds.
Mironova's insights were echoed by practitioners in the CX Network report The state of CX tech 2026, which explored how organizations such as Bosch and Toyota migrated to connected and unified CX operations.
Spending tip 4: Know what to ask of prospective vendors
When selecting partners to deliver these solutions, Paul Weaver, VP of global digital product design for Hyatt Hotels, says practitioners must be prepared to quiz their prospective vendors on everything from real-world use cases to the meaning of capability.
He explains: "I would start by focusing less on deliverables and more on outcomes. It is easy to ask a partner what reports, workshops, designs, or frameworks they will provide. Those things may be useful, but they are not the same as value."
Weaver says the "more important" questions are:
- What customer or business outcome are we trying to improve?
- How will we know whether we have improved it?
- What evidence do we have that this is the right problem to solve?
- How will assumptions be validated before significant investment is made?
The vetting doesn't stop there. To ensure the investments truly tie back to business strategy, Weaver says prospective partners must also demonstrate where their work has improved measurable outcomes, "not just customer sentiment or design quality". On this point, there are further questions to ask:
- Did it improve conversion?
- Reduce effort?
- Increase adoption?
- Lower support demand?
- Improve task completion?
- Reduce risk?
The 11th and final question takes a much broader view.
What capability remains inside the organization after the engagement is complete?
It also relates to the growing trend for solutions providers to embed forward deployed engineers to their customer organizations.
"The best partners do not just help organizations build things; they help organizations make better decisions about what should be built in the first place, as well as how those decisions can be validated, measured and improved after the initial work is complete," Weaver explains.
"That is where the real value often sits. A good partner should reduce uncertainty, help teams focus on the right problems and leave the organization better equipped to keep improving after the project is done," he adds.
For more guidance on how to quiz your prospective vendors, download the whitepaper 6 questions to ask your AI vendor from Quiq.
Quick links
9 ways CX will change in 2027
https://www.cxnetwork.com/cx-experience/articles/9-ways-cx-will-change-in-2027
6 AI challenges for CX to tackle in 2027
https://www.cxnetwork.com/cx-experience/articles/6-ai-challenges-for-cx-to-tackle-in-2027
What does the future of customer insights look like in an AI-driven world?
https://www.cxnetwork.com/customer-insights/articles/future-of-customer-insights-ai-driven