Today ETAPX is rolling out native chat across Influxx — a rebuilt, first-class conversational layer inside the ADE, the ADE mobile app, and the Podz workspace experience, replacing what used to feel like a secondary panel bolted onto the side of the real work. This is not Influxx becoming a chat product. It is the part of Influxx that was always a real-time exchange between you and your agent — check-ins, clarifying questions, quick follow-ups, reviewing what got done — finally getting an interface built specifically for it, instead of borrowing space from the agentic workspace that was never designed to host it.
That distinction matters enough that we are going to spend this entire piece on it, because it is easy to hear "native chat" and assume the opposite of what actually happened. Influxx's whole reason for existing is that it is not a chat window — it is an outcomes-driven agentic environment where you describe a goal and the work gets carried through, end to end, with check-ins only when a real decision needs you. Nothing about that has changed. What changed is that the check-ins themselves — the moments that were always part of using Influxx — now happen in an interface that was actually designed for them, rather than one that was borrowed from somewhere else and never quite fit.
The Conversation Was Always Part of the Job
Every agentic workflow, no matter how autonomous, has moments that require a real exchange. Influxx asks a clarifying question before committing to an interpretation of an ambiguous request. It surfaces a decision when a task branches into genuinely different directions and your judgment is the only thing that can pick one. It shows you what an agent did and waits for a thumbs-up, a correction, or a follow-up before moving to the next step. None of that is new to Influxx — it has been part of the product since the day it launched, because an environment that carries work through multiple steps still needs a human in the loop at the moments that matter.
What has changed is how those moments felt to actually use. The check-in layer was functional, but it read as secondary — a panel that opened next to the real workspace rather than something that felt built into it, with its own inconsistencies depending on whether you were in the ADE on a desktop, pulling up the mobile app between meetings, or working inside a Pod. It worked, in the sense that messages sent and replies arrived, but it never felt like part of the same product as the agentic workspace surrounding it. Engineers described switching into it as a small context break in the middle of a task that was otherwise moving smoothly.
That gap became more noticeable, not less, as people leaned on Influxx for more ambitious work. The more a task branches into real decisions, the more the check-in layer gets used — and the more a slightly clunky interface for that layer costs you. A tool that is excellent at carrying work through ninety steps but awkward at the handful of moments where you actually need to weigh in has a design problem sitting exactly where it matters most. We heard that from engineers, from Pod owners, and from the mobile app's early users with enough consistency that it became the clear next thing to fix.
What "Native" Actually Means Here
To be specific about what shipped, because "native" gets used loosely: this is a product and interface change, not an infrastructure change we are asking anyone to understand or care about. The conversational layer is now built directly into the ADE, the ADE mobile app, and Podz as a first-class part of each surface, rather than a separate panel with its own visual language, its own loading behavior, and its own set of quirks depending on where you encountered it.
Concretely, that shows up as one consistent design for check-ins, questions, and follow-ups everywhere you use Influxx — the same interaction patterns whether you are reviewing a decision on your laptop mid-afternoon or catching a clarifying question on your phone between meetings. It shows up as replies and reactions that feel immediate rather than deliberate, in a layout that was actually designed around fast back-and-forth instead of adapted from something else. And it shows up as continuity: a conversation started in the ADE is the same conversation you pick up in the mobile app, not a separate thread you have to go looking for.
- One consistent interface: the same chat experience across the ADE, the ADE mobile app, and Podz, instead of a slightly different version of it depending on which surface you happen to be using.
- Built into the workspace, not beside it: check-ins and follow-ups now live inside the same flow as the agentic work itself, so responding to a question doesn't feel like leaving the task to go somewhere else.
- Faster, lighter exchanges: quick clarifying questions and short follow-ups are designed to be answered in a glance and a tap, not treated as a full context switch.
- Continuity across devices: a conversation that starts on desktop picks up cleanly on mobile, and vice versa, without hunting for where it went.
- Clear review moments: when an agent finishes a step and wants your read on it, that review now has a dedicated, legible place to happen instead of getting buried in a side panel.
Every one of those is a product and design decision. We are intentionally not describing how any of it works under the hood, because that is not the point of this piece and it is not what changed for you as a user. What changed is what the check-in layer feels like to actually use — and it now feels like it belongs to the same product as the rest of Influxx, because it does.
No, Influxx Is Not Becoming a Chatbot
We want to address this head-on rather than let it sit as an implication, because we know exactly how "native chat" reads out of context: it sounds like Influxx is pivoting toward being a chat product, and that would contradict everything Influxx has stood for since it launched. It is not, and it does not.
The test is simple. Influxx is still outcome-first: you describe what needs to get finished, and Influxx works backward from that outcome through however many steps it takes, rather than waiting on you to supply the next instruction one message at a time. Influxx still follows through across steps on its own, working in the background while you do something else, and it still checks in only when a decision genuinely needs your judgment — not after every step, not to narrate its own progress, not to keep a conversation going for its own sake. Native chat did not touch any of that. It touched the interface for the moments that were already check-ins, questions, and reviews, and made that interface feel like it belongs to the product instead of feeling borrowed.
Put another way: a chatbot's job is the conversation. Influxx's job is the outcome, and the conversation is a small, necessary part of getting there — the part where your judgment plugs in. Making that part well-designed does not make it the main event. It makes it a better-built supporting piece of a product whose main event was, is, and remains getting things done without you having to babysit every step of how.
"I want to be direct about this because I know how it can sound from the outside: we did not build a better chat window and decide to make that the product. We looked at the moments where Influxx genuinely needs to talk to you — a clarifying question, a decision, a quick review — and we built those moments a proper home instead of leaving them stapled to the side of the workspace. The agentic core hasn't moved an inch. We just stopped making the one part of Influxx that's actually a conversation feel like an afterthought next to the part that isn't."
— John Ridge, CTO, ETAPX
Where You'll Feel the Difference
Native chat touches three surfaces, and it touches each of them a little differently because each one is used a little differently.
In the ADE, the check-in layer now sits alongside the agentic workspace as a peer, not a drawer you pull open. When an agent working through a multi-step engineering task hits a genuine fork — a naming decision, a trade-off between two reasonable approaches, a question about scope — that question now shows up in a place built for a quick, legible exchange, and answering it drops you right back into watching the work continue, rather than requiring you to close one interface to get back to another.
In the ADE mobile app, native chat matters even more, because mobile is where check-ins tend to actually happen — the quick decision between meetings, the review of what an agent finished overnight, the clarifying answer typed from a phone while you are technically doing something else. A conversational layer that was designed for a desktop panel and squeezed onto a phone screen shows its seams fastest in exactly these moments. The native version was built for mobile as its own surface, not adapted to fit it, which is the difference between a check-in that takes ten seconds and one that takes two minutes of scrolling and zooming to get through.
In Podz, the dedicated workspaces John Ridge and his team built for organizing separate projects, native chat means each Pod's check-ins, questions, and reviews live inside that Pod's own context rather than spilling into a single undifferentiated inbox shared across everything you're working on. If you run three Pods for three different projects, the conversational layer for each one now feels like it belongs specifically to that Pod, the same way the agentic work inside it already does.
"I run two Pods for two different clients and I used to genuinely lose track of which check-in belonged to which project — the chat side of things felt like an afterthought compared to how well-organized everything else was. Now it doesn't. A question from one Pod looks and feels like it belongs to that Pod, and I can answer it from my phone in the time it used to take me to just find the right thread."
— Priya Anand, early native chat user
A Day With Native Chat
The best way to understand what actually changed is to walk through an ordinary day, because the difference is not in any one feature — it is in how many small frictions quietly disappeared.
You hand a task to Influxx first thing in the morning, the same way you always have: describe the outcome, not a script of instructions, and let it get to work. An hour later, a notification arrives — not a generic "check your messages" alert, but a specific, legible question about a decision the task ran into, shown in the same visual language as everything else in the ADE. You answer it in the time it takes to read one sentence and tap a response, and the task keeps moving without you having to go find where the question lives or figure out which panel it opened in.
Midday, you are away from your desk and a Pod you are running finishes a step that is worth your review before the next one starts. You pull up the mobile app, and the review is right there, formatted for a phone screen instead of shrunk down from a desktop layout, with the context you need to actually make a call rather than a wall of text you have to piece together. You respond, and by the time you are back at your laptop the task has already moved forward on the strength of that answer.
Late afternoon, a second Pod surfaces a genuinely open-ended question — one of the moments where your taste matters more than the agent's competence, exactly the kind of decision Influxx has always been built to route to you rather than guess at. The conversation that follows is quick, because the interface is built for quick exchanges rather than treating every reply as a fresh page load. You settle it in two messages and get back to the part of your day that was never going to be handed off to anyone, human or otherwise.
None of that day involved more conversation than a day using Influxx always has. What changed is that none of it felt like a detour. The check-ins were still exactly as infrequent as Influxx's whole design promises they'll be — the environment still did the vast majority of the work without asking anything of you — but the moments where it did ask stopped being the clunky part of an otherwise smooth experience.
Why We Built It This Way
There was a simpler path available here, and we want to be honest that we considered it: leave the check-in layer as a lightweight, secondary panel and put our effort entirely into the agentic core, on the theory that the conversational part is minor by design and doesn't deserve much investment. We rejected that, for a reason that should sound familiar if you followed how Influxx itself came to exist as its own product in the first place: something a lot of people rely on, even briefly, deserves to be built well rather than tolerated.
Influxx was built as a dedicated environment instead of a feature bolted onto something else because getting work done deserves its own room, built around its own rhythm. The check-in layer inside Influxx deserves the same standard applied at a smaller scale. It is a small part of the overall experience by design — that was never in question, and it is not changing — but small does not mean it should feel like an afterthought. A well-built five-second exchange is still worth building well, especially when it is the exact moment where your judgment enters the loop and the task's outcome actually depends on getting your input clearly and quickly.
There is also a trust dimension to this that is easy to underrate. The check-in is the moment Influxx is, in effect, asking for your attention and your judgment. If that moment feels clumsy, it quietly teaches you to dread check-ins, to answer them faster and less carefully than they deserve, or to put them off — which undermines the entire premise that Influxx only interrupts you when it genuinely matters. A well-designed, native check-in earns the trust that the interruption was worth having, which is precisely what lets Influxx keep the interruptions rare.
Rolling Out Across the Board
Native chat is live now across the ADE, the ADE mobile app, and Podz, and it rolls out as a direct replacement of the previous check-in experience rather than an opt-in layered on top of it — there is no separate mode to switch into. Existing conversations and check-in history carry forward, so nothing about this changes what you'd find if you looked back at a Pod's history from last month; it only changes how future check-ins look and feel.
This is also, deliberately, not the finish line. The near-term work is about deepening consistency across the three surfaces even further — making sure a check-in feels exactly as immediate on a Pod accessed from a phone as it does on a desktop ADE session, and continuing to tighten the handful of interactions, like reviewing a multi-part decision or scanning back through a longer exchange, that benefit most from careful design attention. None of that roadmap touches the agentic core. It is entirely about making the conversational layer earn its keep as well as the rest of Influxx already does.
We are treating this the same way we treat every part of Influxx: as something that should keep getting better without asking you to relearn how to use it. The goal was never to make chat the centerpiece. The goal was to stop making it the weak point.
Frequently Asked Questions
Does this mean Influxx is becoming a chatbot?
No. Influxx remains an outcomes-driven agentic environment — you describe a goal, and the work is carried through multiple steps on its own, with check-ins only when a real decision needs your judgment. Native chat is a redesign of that check-in layer, not a shift toward conversation as the main way you use Influxx. The interface for the small, necessary moments where Influxx talks to you got better. The fact that Influxx mostly works without talking to you at all has not changed.
What actually changed with native chat?
The conversational layer inside the ADE, the ADE mobile app, and Podz was rebuilt as a first-class, consistent part of each surface, rather than a secondary panel that felt bolted on and varied depending on where you encountered it. Check-ins, clarifying questions, quick follow-ups, and reviews now look and feel the same everywhere, and they load and respond faster. This is a product and interface change — we are not describing any technical or infrastructure detail behind it, because that isn't the point of the update.
Will Influxx check in with me more often now?
No. The frequency of check-ins is governed by the same principle it always has been — Influxx surfaces a moment only when a genuine decision needs a human. Native chat did not add check-ins or make Influxx more chatty; it made the check-ins that were always going to happen feel native to the product instead of like a detour away from it.
Does native chat work the same way across the ADE, mobile, and Podz?
It's built as one consistent experience across all three, with each surface tuned for how it's actually used — the ADE mobile app is designed for quick exchanges on the go, while Podz keeps each Pod's conversation scoped to that Pod's own context. A conversation you start on one surface carries over cleanly if you pick it up on another, rather than existing as a separate thread you have to track down.
Do I need to do anything to get native chat, or change how I use Influxx?
No setup is required. Native chat is rolling out as the default check-in experience across the ADE, the ADE mobile app, and Podz, and existing conversation history carries forward. How you use Influxx day to day — handing off outcomes, letting the work run, responding when a real decision comes up — stays exactly the same.
Native chat is a small change described plainly: the part of Influxx that was always a real conversation now feels like it belongs to the product, instead of feeling like something borrowed and squeezed in beside it. Influxx's job hasn't moved — it is still where you hand off an outcome and get back something finished, with your judgment pulled in only when it's genuinely needed. What changed is that the moments where your judgment gets pulled in finally have an interface worthy of them.
