Customer engagement teams have never had more data to work with.
They can see the customer who has stopped buying, the campaign that is losing momentum, the survey response hinting at a new need, and the support interaction that should change what happens next.
The problem is often what comes after the signal.
“The signal that you might see in a dashboard or you might see in the data can still trigger several handoffs between several teams,” Kabir Shahani, Co-Founder and Co-CEO of Amperity, told CX Today.
That delay is where relevance can be seriously affected. By the time analytics has checked the data, marketing has built the audience, and another team has approved a journey, the customer may have purchased elsewhere, resolved the problem, or simply lost interest.
This week, Amperity, Treasure AI, and Adobe all pushed toward a similar answer: give AI more responsibility for the slow middle between insight and action.
Yet each announcement also carried an important caveat: someone still needs to own the decision.
Editor’s Pick: Amperity Wants to Close the Gap Between Insight and Action
Amperity’s new Pér agent is designed to investigate customer-data questions, recommend a response, and prepare the next action across connected tools for human approval.
That sounds like another agent announcement, but it is more interesting than that.
Most organizations do not need help finding a segment; they need help deciding whether a signal matters, identifying the customers affected, checking the evidence, and acting before the moment passes.
Shahani gave CX Today an early example from MacKenzie-Childs, a homewares brand examining survey feedback from shoppers buying wedding, bridal, registry, or engagement-related gifts.
Pér identified 3,000 distinct customers connected to those occasions, with 90% already opted into email.
It then proposed a welcome and registry-engagement journey, including potential email treatments for the team to review.
“It’s the person; it’s the user that’s deciding. But it’s Pér that’s doing the work,” Shahani said.
That model may become more common. Brands are happy to shorten the work between customer intelligence and execution, but they remain reluctant to hand over the judgment entirely... for good reason.
Treasure AI Sends Its Digital Worker into the Journey
Treasure AI has launched Digital Worker, a governed AI teammate designed to turn customer signals into insights, recommendations, and actions.
The vendor also announced Personalization Studio, a real-time website personalization workspace for marketers, alongside an engagement-based email pricing model that charges according to actions such as clicks rather than send volume.
Digital Worker can run on demand, to a schedule, or proactively, according to Treasure AI. Its role is to take on defined marketing work within permissions set by the team.
In the official release, Kaz Ohta, CEO and Co-Founder of Treasure AI, said:
“Digital Workers run on demand, on a schedule or proactively, always within team-defined permissions and guardrails.”
The emphasis on guardrails is doing a lot of work here.
Customer engagement platforms are beginning to describe AI as a colleague, not merely a feature. That creates an attractive proposition for overstretched marketing teams: fewer manual checks, faster audience analysis, and more timely personalization.
However, it also creates a more difficult question: when an AI worker has acted on a signal, who decides whether the message was appropriate, the offer made sense, or the customer should have been contacted at all?
Treasure AI’s answer is to keep teams in control of strategy and boundaries. The practical test is whether that oversight is usable when campaigns, channels, and customer signals begin moving faster than a human review process.
Adobe and NVIDIA Want Faster, More Capable Brand Conversations
Adobe and NVIDIA have expanded their collaboration around AI-powered creative work, production workflows, and personalized customer experiences.
Adobe Brand Concierge is a particularly interesting element. The enterprise conversational experience tool is being optimized with NVIDIA Nemotron models, with availability planned later this year.
NVIDIA has selected Brand Concierge for its own use, according to Adobe, integrating company data to provide more customized and responsive answers to queries.
Jason Paul, Vice President of Local AI at NVIDIA, said:
“AI agents are helping creators unlock deeper and richer workflows using the Adobe applications they know and love.”
The bigger customer engagement opportunity is obvious. Faster creative production and stronger conversational agents could help brands respond to customers with more useful, on-brand content.
Still, speed will not repair a broken experience by itself. A polished answer that cannot access accurate availability, order status, policy information, or customer context remains a polished dead end.
Talking Point: Faster Is Not Always More Personal
This week’s announcements are built around acceleration. Amperity wants to speed up the decision between customer signal and journey. Treasure AI wants to automate defined marketing work. Adobe and NVIDIA want to improve the performance behind brand conversations.
All worthwhile goals.
But customer engagement does not become personal simply because a brand can react faster. The critical question is whether the action has a clear reason to exist from the customer’s perspective.
AI can investigate, recommend, and prepare. It can even execute within a tightly defined workflow. But brands still own the relationship when that action misses the mark.
Look Ahead
The next test for customer engagement AI will be outcome, not output.
Vendors can demonstrate an agent creating a segment, proposing a campaign, or generating a response. The stronger proof will come when they show that the agent reduced customer effort, improved relevance, and knew when not to act.
