The Customer Conversation Is Becoming a Control Surface
Plus: utilities are letting AI get closer to billing and outage support, Apple wants the assistant closer to customer intent, and one service tool still looks built for the real work.
Your daily signal on AI and CX — minus the hype.
DCX Stat of the day: Kraken says its platform now supports more than 90 million customer accounts across 15+ countries. That gives you a sense of the scale here. When AI moves into billing, outages, and account support, it is stepping into live infrastructure, not a contained pilot. Kraken
In this issue:
→ Utilities are letting AI get closer to live account context
→ Platform owners want to own the front door to customer intent
→ Scale is showing up faster than most control models
→ DigitalGenius still looks more useful than flashy
→ The opportunity is better service with clearer operating lines
🔎 Deep dive
Utility AI just moved closer to the account
Kraken Technologies is partnering with Sierra to bring generative AI customer service into utility environments. This is not the usual chatbot setup. Utility service sits inside billing questions, outage moments, tariff confusion, payment pressure, and account anxiety. When it works well, it can remove a lot of friction from moments customers already find stressful.
Axios reports that the system is meant to combine Sierra’s conversational AI with Kraken’s utility software, including account, meter, and rate data. So the model is getting closer to the operating system behind the journey, not just the words around it.
Once the conversation can touch live service context, the upside gets bigger and so does the need for clarity. You need to know what the AI can explain, what it can steer, when a human steps in, and how the whole experience stays easy to trust. In a utility setting, that can mean faster answers, cleaner guidance, and less customer effort, as long as the lines are clear.
Bottom Line: The customer conversation is becoming part of the operating layer. If AI can see the account, your team needs to know where explanation stops and action begins so the experience gets easier, not murkier.
📬 Copy-Paste Take
If your AI can touch live customer context, this stops being just a chatbot project. It becomes a service-design and operating-model question.
OPERATOR PLAYBOOK
Audit the moments where conversation can turn into action
Most teams are still reviewing prompts and containment rates. Here, that is only part of the job.
Audit every service flow where AI can shape a customer decision for four things:
The exact data the model can see, and how fresh it is
The moments where explanation turns into guidance or a recommended next step
The handoff points that keep the experience smooth when a human needs to step in
The owner responsible for keeping the whole experience coherent end to end
Then test whether your current QA process can tell you if the experience is actually getting easier for the customer.
Ask your team: Where does our AI currently shape a customer decision, and who owns that experience end to end?
Signal: If your team can name the assistant but not the owner of the experience around it, you still have work to do.
📈 Market Reality Check
Walmart is treating AI prompts like customer context
Walmart’s Sparky shows what shopping prompts can reveal before the customer ever clicks a filter or types a keyword. A customer who says they need help finding a snack for a child with allergies is giving you context that normal search often never captures.
That makes AI shopping bigger than a retail media story. It becomes a customer-understanding story. If conversational prompts carry richer intent, preference, and constraint data, retailers have a chance to make guidance more useful. They also have a new temptation to over-monetize the moment.
Why it matters: The real shift is not just that AI can recommend products. It is that customers may explain what they need in fuller, more human language, and that changes how brands should think about discovery, relevance, and interruption.
Better prompts = richer customer signals
🧰 Tool Worth Knowing
DigitalGenius
What it does: DigitalGenius is a customer service AI platform built around orchestration across email, chat, and voice, with agent copilots, autonomous flows, knowledge search, analytics, and a headless SDK for embedding support experiences.
CX use case: It is useful when the real problem is not writing prettier answers. It helps clean up the gaps between channels, automation, agent assist, and workflow execution in high-volume support environments.
Worth watching because: A lot of AI support tooling still feels like a thin wrapper around response generation. DigitalGenius is more useful when you need routing, reasoning, human handoff, and deployment flexibility to work together in a way that makes service feel more connected.
Bottom line: If your service operation is dealing with handoffs, channel fragmentation, and inconsistent assist behavior, this is the kind of stack I would pressure-test.
The DCX AI Today - AI Tool Directory - If you lead a CX team and want a curated shortlist of tools worth evaluating, this is your starting point.
⚡ 90-Second CX Radar
Apple wants Siri to feel less clingy and more useful
On the surface, this is about personality tuning. In practice, assistant design is now a CX decision. Tone, interruption style, and response posture all shape trust, friction, and repeat use.
Why it matters: Customers are judging whether AI works, and whether it feels bearable to live with.
🧭 Your Move
The question underneath this whole issue is simple: where can your AI influence a customer decision before a human or a brand-owned channel gets a second look?
That is usually where the work gets messy, and where the customer notices the experience first.
Next week, map one service or commerce journey where AI can see live customer context. Then mark the exact point where the answer becomes a recommendation, a nudge, or an action trigger.
If nobody clearly owns that line, sort that out before you chase another launch.
The real test for CX AI is simple: does it make the next step easier for the customer, while still being clear about who owns what?
Until Monday,
👥 Share This Issue
Think of one person who’s wrestling with AI in CX right now
and forward this to them.
I’m obsessed with Wispr Flow Pro! Get a Free Month on me.
If someone forwarded this to you, they thought you needed to see it before your next AI planning meeting. Get your own copy.