It is 3 a.m. and a customer named John has not bought anything in eleven months. Somewhere in lower Manhattan - or rather, somewhere in the cloud that a lower-Manhattan company runs - software is quietly reading his order history, noticing he is a Knicks fan, and drafting him a note about a new bomber jacket. No human is awake for this. By the time one is, the message is written, on-brand, and waiting for a one-click approval. This is Remarkable AI on a slow night, which is to say a working night.
The company used to be called Chatdesk. The name changed; the obsession did not. Remarkable AI exists because of one stubborn fact about modern commerce: the place where brands talk to their customers - the support inbox, the comment section, the DM - is treated like a cost to be minimized, when it is actually the most honest sales channel a brand owns. Every angry email is a customer still paying attention. Every comment is a hand raised.
Support is the one channel where the customer started the conversation. Most companies treat that like a problem to be closed. Remarkable AI treats it like a door someone just opened.
A backlog is just unanswered revenue
Direct-to-consumer brands grew up fast and broke their support along the way. A product goes viral, ten thousand questions land overnight, and a four-person team drowns. The traditional fixes were both bad: hire a call center that has never heard of your brand, or make customers wait three days and watch them leave. Neither felt remarkable. Both felt like the cost of doing business.
Remarkable AI's founders looked at the same backlog and saw something less depressing. Most support tickets are not unique. They are the same forty questions, asked in ten thousand voices. That is a machine-learning problem wearing a customer-service costume. And the tickets that are genuinely tricky - the ones that need warmth, judgment, a human who actually owns the product - are exactly the ones worth handing to a person who loves the brand.
So the bet was a split. Let an AI engine handle pattern and volume. Let vetted human experts handle nuance and the moments that build loyalty. Connect the whole thing to the tools brands already use - Shopify, Zendesk, email, SMS, the social inbox - so nobody has to rip anything out. Then, quietly, do the thing nobody else was doing: measure how much money the support channel actually makes.
It turns out the answer is: a lot, if you bother to look. Which, irony of ironies, almost no one did.
The cheapest growth channel in e-commerce was hiding inside the department everyone was trying to spend less on.
A Googler and a McKinsey alum walk into an inbox
Spent 7+ years as a Product Manager at Google, working on Voice Search and the Google Assistant. He knows what natural language processing can and - more usefully - cannot do, which is why humans never left the loop.
Came from McKinsey & Co., where he spent his days on customer experience and digital operations for large companies. He had seen, up close, how much value leaks out of a badly run support function.
Okonkwo did not build a chatbot, which is notable given his resume. He had watched enough AI demos at Google to know that a customer who needs a refund does not want to argue with a script. The interesting product was not "AI instead of people." It was "AI plus the right people," with the software deciding, in real time, which is which. The founders raised about $2 million in 2017 to test the idea, then $7 million more in early 2022 once the idea had customers willing to vouch for it.
Anyone can automate the easy questions. The trick is knowing - automatically - which questions should never be automated.
Three jobs, one engine: Support, Retain, Acquire
Answer everything, everywhere, always
24/7 coverage across email, social, SMS, and chat. The AI clears the repetitive volume; experts who actually use the brand handle the rest. Backlogs shrink without a single new in-house hire.
Win back the ones who drifted
Personalized win-back messages aimed at lapsed customers. The company reports up to 50x revenue per recipient - and notes that more than half of re-engaged customers reorder without any extra nudging.
Turn comments into customers
Proactive 1:1 engagement on TikTok, YouTube, X, and Reddit. Every comment and DM becomes a chance to convert a follower into a buyer, in a voice that sounds like the brand and not a help desk.
One-click integrations
Plugs into Shopify, Zendesk, and the social inbox a brand already runs. No rip-and-replace, and deployments measured in days rather than quarters. Pricing starts around $1,500 a month.
A short history of an unglamorous idea
Numbers that survived a skeptic
Claims are easy. So here is the part where the support channel stops being a feel-good story and starts being a spreadsheet. During one peak season, Remarkable AI resolved more than 20,000 support tickets for the underwear brand SAXX - and converted 9.4% of those tickets into sales along the way. That second number is the whole pitch: roughly one in ten "I have a problem" messages turned into "I'll take it."
What brands report after switching it on
Figures are customer-reported and self-selected - the brands happy enough to put their name on a case study. Treat them as evidence of what's possible, not a guaranteed average. Source: beremarkable.ai case studies.
The roster reads like the inside of a millennial's bathroom cabinet and closet: SAXX, Hot Topic, New Era, Curology, Sheertex, Thinx, SheaMoisture, For Love & Lemons, Storyworth, Kindra. More than a thousand brands in total. The investors are not strangers to the space either - Menlo Ventures, Harlem Capital, Fika Ventures, and Serena Williams' Serena Ventures all wrote checks, the last two with a long record of backing founders the rest of the industry overlooks.
More than half of re-engaged customers have ordered again - without additional prompting.
"Every customer experience should be remarkable"
It is a slogan, yes, and slogans are cheap. But this one happens to be load-bearing. The entire company is an argument that "remarkable" and "scalable" are not opposites - that you can answer a million messages and still make each one feel like a person wrote it, because for the ones that matter, a person did.
The rebrand from Chatdesk to Remarkable AI in 2024 was not a fresh coat of paint. It was the company admitting what it had become: less a chat tool, more an engine that decides, message by message, where the machine ends and the human begins. The name now says the quiet part out loud.
There is a worldview underneath it. Most software wants to remove humans from the loop entirely; it is cheaper, it scales, it never asks for a raise. Remarkable AI made the unfashionable bet that the human is not the bug - the human is the product, and AI's job is to make sure that human only shows up where they are worth it.
That is a harder business to build than a pure chatbot. It is also, conveniently, a much harder one to copy.
The inbox isn't going anywhere
As every brand bolts a generic AI agent onto its website, the support inbox is about to fill with software that is fast, tireless, and slightly soulless. Customers will notice. The brands that win the next few years will be the ones that figured out how to be fast and human at the same time - automated where it's invisible, personal where it counts. That is precisely the line Remarkable AI has spent a decade learning to draw.
So return to John, asleep, eleven months lapsed, a Knicks fan who forgot he liked this brand. In the old world, he was a churned number on a dashboard, written off. In the world Remarkable AI is building, he is a conversation that hasn't ended yet - read, understood, and gently reopened before sunrise. The support inbox used to be where customers went to complain. Now it's where the brand goes to grow. Same inbox. Different company watching it.