In the rush to reap AI benefits in CX, many organizations have adopted AI, but they haven’t assigned ownership for AI’s business outcome to a single leader. That lack of leadership can lead to failed pilots, lackluster results or outcomes that show potential but can’t be definitively linked to business value.
One issue is that AI is more than another layer in the enterprise tech stack. It’s not simply a tool to improve outcomes in a CRM, contact center as a service, unified communications as a service, or workforce management platform. AI should operate within and across each of those organizational silos.
“AI is a cultural and behavioral transformation disguised as a tech stack,” Behnam Behzadifar, a CX expert and senior strategist with Mofid Securities, told CX Dive in an email.
Achieving that potential requires organizations to facilitate integrations and teamwork across internal groups — and that requires leadership. To deliver successful AI solutions, CX leaders need to do what far too few have done — seize accountability for AI business outcomes.
A mere 17% of organizations have assigned a named owner for AI business outcomes, according to a recent Metric Sherpa research report. Produced in partnership with CX vendor Sprinklr, Metric Sherpa surveyed more than 800 U.S. marketing, service or customer experience practitioners.
In customer service, the right person to own an AI deployment is the head of service or contact center operations because they’re already responsible for cost to serve and customer satisfaction, Justin Robbins, founder and principal analyst of Metric Sherpa, told CX Dive in an email.
“A simple test identifies the right person: If this AI made the customer experience worse next quarter, whose scorecard would show it first?” Robbins said.
The benefits of naming an AI owner
Having one person accountable for AI’s business outcomes means they’re invested in the project’s success. Compared with respondents that haven’t designated an owner, those with a named owner can precisely quantify results almost twice as often, according to the report. They also use controlled tests more often, can more often cite financial reporting as evidence and report on goals and baselines more often.
Companies who appoint an owner also uncover more problems in the deployment, Robbins said.
These enterprises report governance, compliance, privacy or security friction at more than twice the rate and data quality problems more often than organizations that don't, according to Robbins.
“That counts in the owner’s favor, because teams can fix the problems they see,” Robbins said. “When budgets tighten, organizations with an owner can defend the program with evidence.”
Assigning an owner doesn't automatically mean an AI rollout will deliver the desired results, but it means someone can answer questions about where a rollout goes wrong and what's working.
AI in CX requires coordination across multiple internal departments, and potentially downstream and upstream partners, as well. Just because that CX leader is accountable for the results doesn’t necessarily mean they have the authority over executing the rollout or its daily operations. They need to be given a clear mandate to require those individual departments to make changes that support the success of the AI rollout.
The CX leader should be able to audit and influence processes across the silos that the AI rollout touches, from logistics to ordering to billing and perhaps beyond. They also need the mandate to pause feature rollouts if quality benchmarks are missed, Behzadifar said.
“If an organization hires or appoints a CX leader without giving them real organizational teeth, they are simply hiring a glorified scapegoat,” Behzadifar said.
That CX leader therefore needs a clear chain of command that leads to the CEO, COO or a joint steering committee.
“Without top-down air cover, legacy department heads will prioritize their local KPIs over enterprise-wide customer outcomes,” Behzadifar said. “If accountability for an AI initiative is split across multiple business teams, no one owns the end-to-end customer outcome and the initiative dissolves into finger-pointing.”
This siloed approach is what can lead to failed AI pilots, lackluster results or deployments that produce metrics that may look like outcomes but don’t necessarily demonstrate business value.
“Many organizations bought AI as a tool rather than managing it as a business change. What started as an experiment moved into production across multiple teams, without anyone becoming accountable for the result,” Robbins said. “When everyone touches it, everyone assumes someone else owns it.”
