BREAKING* ASSEMBLED'S AI AGENT NOW RESOLVES 70%+ OF SUPPORT CHATS* $70.7M RAISED ACROSS 3 ROUNDS* STRIPE, ROBINHOOD, ETSY, INTERCOM RUN ON IT* SF-BUILT, EX-STRIPE FOUNDED, 150-PERSON STRONG* AGENTIC SCHEDULE GENERATION GA - JAN 2026* BREAKING* ASSEMBLED'S AI AGENT NOW RESOLVES 70%+ OF SUPPORT CHATS* $70.7M RAISED ACROSS 3 ROUNDS* STRIPE, ROBINHOOD, ETSY, INTERCOM RUN ON IT* SF-BUILT, EX-STRIPE FOUNDED, 150-PERSON STRONG* AGENTIC SCHEDULE GENERATION GA - JAN 2026*
Assembled logo
YesPress Dossier · Company · SaaS / AI

Assembled.

The quiet operating system behind the world's busiest support desks - now teaching AI agents to answer the phone.

FOUNDED 2018 SAN FRANCISCO SERIES B / $70.7M ~150 EMPLOYEES
ABOVE: The Assembled mark, photographed in its natural habitat - the second tab of a support manager's browser, three minutes before the morning standup.
SHARE THIS DOSSIER: Twitter / X LinkedIn Facebook Instagram Copy URL

The Quietest Software in the Call Center

Right now, at 10:47 a.m. Pacific, a support manager at a fintech you've heard of is opening a browser tab. The forecast for the afternoon shift is already there. Two agents called in sick. The schedule has rebalanced itself. An AI agent has handled 612 chats since breakfast and escalated 19. None of this is dramatic. All of it is Assembled.

Assembled does not make a product you would put on a billboard. It makes the thing that makes Stripe's refund queue move, the thing that quietly absorbs Robinhood's option-expiry Fridays, the thing that lets Etsy's seller support team take Tuesday off without anyone noticing. It is, in the founders' own early phrasing, "the operating system for support teams" - and over the past seven years it has gone from a polite Google Sheets replacement to the place where humans and AI agents share a single calendar.

Customer support is the largest unsolved labor problem in tech. Assembled is the company that decided to solve it without writing a single inspirational blog post about empathy.- YesPress observation, May 2026

The Problem They Saw

Modern support is hard for reasons most software companies find boring. There are thousands of agents in a dozen time zones. There are tickets in chat, email, voice, and SMS. There are SLAs measured in minutes. There are vendors in Manila and Mexico City and Manchester. And there is, traditionally, one tired ops manager updating a spreadsheet at 5:30 a.m. so the queue doesn't fall over by 9.

Before Assembled, the tools for this job came in two flavors. The first was the legacy workforce-management software your bank uses, the kind that requires a six-month implementation and a consultant who remembers Lotus Notes. The second was Excel. Most teams, faced with this choice, picked Excel. They picked it loudly. They picked it angrily. They picked it for ten years.

What the incumbents missed

The Wang brothers and Brian Sze saw something the incumbents had stopped noticing: support was no longer a cost center hidden in the basement. At companies like Stripe - where they all met - it had become a product surface, a brand surface, and a place where engineers wanted to ship. The software had to feel that way too: real-time, opinionated, fast, and built by people who had actually carried a pager.

The legacy WFM industry was a museum. Assembled walked in and quietly turned on the lights.- A former contact center director, paraphrased

The Founders' Bet

In 2018, Ryan Wang, his brother John Wang, and Brian Sze left Stripe to build a company in a category nobody at a dinner party wanted to discuss. Ryan had been an early Stripe engineer working on fraud and support automation. John had been one of Stripe's first infrastructure hires. Brian had run support operations during the years it scaled from hundreds of merchants to millions.

Their bet was uncomplicated and unfashionable: that scheduling and forecasting, done properly, was infrastructure - the same way payments are infrastructure - and that whoever became the system of record for support operations would also, eventually, become the place where AI agents got slotted into the rota next to humans.

Origin Story Receipt

March 2020: Stripe led a $3.1M seed round into Assembled. Two months later, the world's support queues lit on fire. A pandemic made the demo unnecessary.

The first year was, in Ryan Wang's own retelling on a podcast, an eight-month stretch of zero revenue caused by a pricing mistake. They survived it. They priced it correctly. Then they raised $16.6M from Emergence Capital, then $51M from NEA, and then they did the thing that startup narratives rarely describe: they kept building the same product.

We saw firsthand at Stripe how support transformed into a complex, geographically distributed function of thousands of people - and how the tools simply hadn't kept up.- Ryan Wang, co-founder & CEO

A Company in Six Dates

2018

Three ex-Stripe colleagues found Assembled in San Francisco.

2020

$3.1M seed led by Stripe. The pandemic makes scheduling existential.

2021

$16.6M Series A from Emergence Capital - WFM goes mainstream.

2022

$51M Series B led by NEA. Stripe, Robinhood, Etsy on board.

2025

Launch of Assist - omnichannel AI agent for chat, email, and voice.

2026

Agentic AI schedule generation goes GA. The spreadsheet finally retires.

The Product, Such As It Is

Assembled today is three things glued together by a single source of truth about who is working when and what they're working on. There is the workforce management platform itself: forecasting, scheduling, real-time monitoring, reporting, and the small mercy of a calendar agents can actually trust. There is the vendor management layer, for the awkward reality that most large support orgs run on a mix of in-house staff and BPO partners scattered across continents. And there is Assist - the AI agent platform that handles chat, email, and voice from one brain.

What you can actually do with it

Forecast call volume for a Black Friday you haven't seen yet. Build a schedule in minutes instead of days. Watch a queue in real time and reroute before the SLA blows. Hand a customer to an AI agent for a refund, then back to a human when they want to vent. Get a clean dashboard out the other end - one that, blessedly, your CFO can read.

Assembled treats support like an engineering problem. Which is, it turns out, exactly what most large support teams have been begging someone to do.- The category they helped define

The Proof, In Numbers

Assembled is the kind of company that lets the receipts do the talking. The funding history is one set of receipts. The customer logos are another. The deflection numbers, the ones Assist quietly publishes, are the third - and probably the most important.

Assembled in Five Useful Numbers

Sources: company materials, TechCrunch, Crunchbase, G2
Total raised
$70.7M
Series B size
$51M
AI chat deflection
70%+
Team size
~150
Years operating
8

The customer list, when you look at it, reads less like a sales pitch and more like a who's who of companies whose support volume can take down a startup overnight: Stripe. Robinhood. Etsy. Intercom. HubSpot. GoFundMe. Rover. Poshmark. Lyra Health. DailyPay. Lulu and Georgia. These are the orgs where bad scheduling has measurable balance-sheet consequences.

Partner Roster

Five9 marketplace partner. Salesforce AppExchange listing. Zendesk-deep integrations. Foundation-model relationships with Anthropic and OpenAI behind the AI agent. The plumbing is intentional.

The Mission, Sincerely

Assembled's stated mission is to power exceptional customer experiences at scale. That sentence does most of the heavy lifting that mission statements usually try to avoid. Power - because they want to be infrastructure. Exceptional - because most support is, statistically, not. At scale - because the interesting problems start at a thousand agents, not ten.

The unstated mission is more interesting. Assembled wants to be the place where AI agents and human agents share one schedule, one performance review, and one set of metrics. In a world where every CX vendor is racing to ship "AI for support," Assembled is making the boring, important argument that what matters is not the model - it is the workflow that decides whether a question goes to the model or the human, who picks up after, and who gets paged when both of them get it wrong.

The future of support is not AI replacing humans. It's AI and humans on the same Gantt chart. Assembled drew the Gantt chart.- The argument, condensed

Why It Matters Tomorrow

Customer support is the part of every software company that touches the most humans and gets discussed at the fewest off-sites. It is also, almost certainly, the function that AI changes most over the next five years. If that change goes well, customers get faster, kinder answers and agents get to do the work that actually requires a human. If it goes badly, we get a decade of robotic refusals and ticket loops.

Assembled has spent eight years quietly positioning itself to be the company that makes it go well. Not by writing manifestos. By building the unsexy connective tissue - the schedules, the forecasts, the vendor reports, the AI handoffs - that decides which path the industry actually walks down.

It's now 10:51 a.m. Pacific. That same support manager has closed the browser tab. The afternoon shift is staffed. The AI agent has handled another 89 chats. Nothing is on fire. Nobody is paged. The product worked, exactly the way good infrastructure is supposed to work: invisibly. Which is, in the end, the most flattering thing you can say about software this important.