JustAI

A company built on one word

Here is a fact that ought to reorganize how you think about marketing, and mostly does not. At Twitter, someone changed the phrase "Elon tweeted" to "Elon just tweeted," and by the company's telling that single added word helped pull in something like 800,000 users. Not a redesign. Not a new feature. A four-letter adverb. If you are an engineer, this is horrifying, because it means enormous outcomes are hiding in changes too small to schedule a meeting about. If you are Neha Mittal, who was running growth at Twitter at the time, it is a business plan.

That is the founding observation behind JustAI, which for its first couple of years was called Just Words, a name that told you exactly what it believed: the words are the product. Most companies treat marketing copy as the thing you write once and forget. JustAI treats it as an asset that should be continuously tested and improved, the way you would never ship code once and refuse to touch it again. The problem is that testing copy is slow, manual and boring, so almost nobody does enough of it. The graveyard of unrun A/B tests is where most marketing growth quietly goes to die.

So the pitch is straightforward, even if the machinery underneath is not: take the part of marketing that is tedious and repetitive - generate a variant, ship it to a segment, measure it, keep the winner, do it again - and hand it to software that does not get bored. JustAI calls the software "agents," which is the word every AI company is contractually required to use in 2026, but here the term is doing real work, because the system genuinely passes tasks between specialized parts rather than dumping everything into one chatbot.

Four agents, one loop

The platform is organized around four agents, which is a tidy way of describing what used to be four different jobs on a growth team. The elegance is not that any one of them is magic; it is that they hand work to each other in a loop that keeps running after everyone has gone home.

Underneath, JustAI leans on reinforcement learning and causal inference - the difference between "this email got opened" and "this email caused a person who would not otherwise have converted to convert." That distinction sounds academic until you realize most marketing dashboards cheerfully take credit for customers who were going to buy anyway. The causal part is the unglamorous, correct part.

The payoff the company likes to cite is speed. A lifecycle experiment used to take three to six months, which is roughly the lifespan of a marketing manager's patience. JustAI compresses that to real time. And when your experiment cycle collapses from a quarter to an afternoon, your whole strategy changes - you stop placing a few big, nervous bets and start learning constantly, in small, cheap increments. Compounding is the entire point.

What the numbers say

JustAI reports that its agents delegated more than 600 marketing decisions a month and touched over $100 million in customer revenue last year, and that it compounded one customer's - Grammarly's - growth by 16%. Numbers a company reports about itself deserve the usual pinch of salt, but the shape of them is instructive: they are revenue and decision-volume numbers, not "emails sent." That is the metric that answers the only question an enterprise buyer actually asks, which is whether the thing moved money.

The customer list is the more persuasive artifact. Grammarly, ClickUp, Notion, Lemonade, Coursera, Thumbtack, Outschool - these are companies with real growth teams who could build a version of this themselves and have chosen not to. When several sophisticated buyers all quietly run the same engine underneath their marketing, the interesting signal is not the AI. It is that the loop is hard enough to build that even good teams would rather rent it.

The growth addiction

Neha Mittal's resume reads like a slow, deliberate walk toward the place where small changes reach the most people: Goldman Sachs, the Chan Zuckerberg Initiative, then product growth leadership at Twitter and Pinterest, where she ran teams for notifications, internationalization and paid marketing and founded Twitter's Women in Product Committee. Growth at that scale is a specific kind of addiction - once you have seen a one-word change move a million people, ordinary product work feels low-resolution. It usually ends with someone starting a company about it. She co-founded JustAI with Jeff Hara, who brings the machine-learning and recommendation-systems half of the story.

That line is easy to put on an About page and hard to live, but it is at least a useful tell about how the company wants to be read: not as a generative-AI novelty that writes copy - anything writes copy now - but as a system that decides which copy, for whom, when, and whether it worked. Generation is cheap. Judgment is the thing they are actually selling.

From Just Words to JustAI

The company graduated from Y Combinator's Winter 2024 batch and raised a $1.7 million seed round from Peak XV Partners and YC. Then, in mid-2026, it raised a $17 million Series A led by Base10 Partners, with YC and Peak XV returning and a notably operator-heavy cap table joining in: growth leaders from Anthropic and Chime, the CTO of HubSpot, and the founders of Eppo and Vapi. That is the kind of investor list that signals the round was less about the check and more about who could tell you whether the product actually works.

Around the same time, Just Words became JustAI. Rebrands usually smell like a company running from an old story; this one reads more like a company whose ambition outgrew its name. "Just Words" was a clever nod to how much a single word matters. But if the plan is to run an entire marketing function autonomously - not just wordsmith it - then "words" undersells the thing. The legal entity, amusingly, remains Choice AI Inc., which is the name nobody uses and the one on the paperwork.

• 2024Founded in San Francisco; joins Y Combinator W24 as Just Words.
• Aug 2024Closes $1.7M seed from Peak XV Partners and Y Combinator.
• Oct 2025Rebrands from Just Words to JustAI, an AI-first agentic marketing platform.
• Jun 2026Raises $17M Series A led by Base10; reports 5x ARR growth and plans to expand into e-commerce and B2B.

What comes next, per the company, is expansion: more engineers, more go-to-market, deeper agentic infrastructure, and a push beyond consumer apps into e-commerce and B2B marketing. That last move is the real test. Consumer lifecycle marketing is a friendly home for this approach because the volume is high and the feedback is fast. B2B, with its long sales cycles and committee decisions, is a harder place for a loop that thrives on quick signal. Whether JustAI's agents are learning something general about persuasion, or just something specific about consumer push notifications, is the question the next two years will answer - and it is a genuinely interesting one to watch.