Overview
StackAI built a no-code layer for connecting large language models to the systems where enterprises keep data and execute work. Founded in 2023 by MIT PhDs Toni Rosinol and Bernard Aceituno, the YC Winter 2023 company evolved from a developer API into a visual workflow builder with more than 100 integrations, private deployment options, and controls for regulated buyers.[1][2]
Its outcome was an acquisition, not a collapse. Asana bought StackAI in May 2026 for about $75 million in cash, subject to adjustments, while offering employees equity awards and an earnout opportunity.[3][4] The deal compressed two complementary assets: StackAI had built cross-system execution and enterprise controls; Asana already had collaborative context, approvals, distribution, and thousands of customers. The same complementarity explains both StackAI's value and the difficulty of scaling it independently as a horizontal platform.
Founding Story
Rosinol and Aceituno began working on the company in 2022, before incorporating the business that YC lists as founded in 2023. Both held MIT doctorates. Public sources do not establish how they met, so any more specific origin story would be conjecture. Their initial problem was concrete: enterprises wanted to use language models, but useful work required connecting those models to private data and operating systems.[1][2]
The first product was an API for developers. Customer work on requests for proposals and sales automation changed the founders' view of the bottleneck. Training another model was less important than reliably moving context into a model and approved actions back into business software. By May 2024, StackAI had become a low-code canvas where users arranged data sources, models, triggers, and actions into workflows.[5]
That pivot moved StackAI toward business teams while increasing the enterprise burden around security, permissions, deployment, evaluation, and implementation. It also created the asset Asana later wanted. A workflow agent that only drafts text is easy to copy. One that can read from SharePoint, reason over governed data, request an approval, and write into an ERP is much harder to ship safely.
Aceituno summarized the early insight to TechCrunch: “Our platform allows people to build workflows that require connecting different tools to work together.”[5] At acquisition, Rosinol drew the sharper product boundary: “General-purpose agents talk; specialized agents act.”[6] The first quote explains the original product; the second explains why Asana bought it.
Timeline
- 2022: Rosinol and Aceituno began building around enterprise data-to-model integration.[2]
- 2023: StackAI joined YC's Winter 2023 batch. YC records San Francisco as its base.[1]
- May 2024: TechCrunch described a six-person company with more than 200 customers and $3 million raised. The product had shifted from an API to a visual workflow builder.[5]
- May 2025: StackAI announced a $16 million Series A led by Lobby VC, with LifeX and several strategic operators joining prior backers.[2]
- May 27, 2026: Asana's filing records the merger with legal target Eigen Inc. as closed.[3]
- May 28, 2026: Asana announced the acquisition and said StackAI would continue as its own product and brand.[6]
What They Built
StackAI gave teams a drag-and-drop canvas for assembling an agentic workflow. A user could connect a source such as SharePoint, SAP, Workday, Salesforce, Snowflake, or Miro; add a model; supply governed context through a knowledge-base node; expose approved tools; and define what happened next. The workflow could read, write, or execute across enterprise systems rather than merely produce an answer in a chat window.[7]
The mature product separated the model from the operating layer. Customers could select providers including OpenAI, DeepSeek, and Perplexity while StackAI handled connections, retrieval, tools, deployment, and monitoring. Pre-built interfaces let teams deliver workflows without building a custom front end. This mattered because an enterprise agent is rarely one model call. It is a chain of access checks, retrieval steps, model decisions, human approvals, and writes into systems of record.
StackAI paired that flexibility with controls intended for sensitive environments: role-based access, approval flows, version control, production status monitoring, private cloud and on-premises deployment, and data-residency options. Its published compliance posture included SOC 2 Type II, HIPAA, GDPR, CCPA-CPRA, and ISO 27001.[7]
The product was differentiated less by a novel canvas than by the work surrounding it. Forward-deployed engineers helped customers build workflows. Weekly co-building sessions, internal Stackathons, and quarterly business reviews moved teams from a prototype to an operating process. That service layer reduced adoption risk, but it also made the business harder to scale like ordinary software.
Market Position
Target Customers
StackAI pursued enterprises that wanted agents but could not send sensitive data through an ungoverned consumer tool. Named customers and users included Nubank, LifeMD, MIT Sloan, SmartAsset, BAE Systems, Monzo, Morgan & Morgan, and Goodwill, plus a top-five US defense agency.[1][2] The jobs ranged from compliance audits and legal intake to underwriting and due diligence.
Market Size
No observed source provides a defensible market-size estimate specific to StackAI's category. The broader enterprise automation or generative-AI markets are too broad to substitute without inflating the opportunity. The better demand evidence is behavioral: more than 200 customers after roughly six months of a working product, a $16 million Series A, named deployments in regulated organizations, and an acquirer willing to pay about $75 million in cash.[5][3]
Competition
StackAI competed on two axes: breadth of execution across systems and readiness for enterprise governance. Zapier and n8n had automation ecosystems and broad connector libraries. Microsoft and Google controlled productivity suites and identity surfaces. OpenAI and Anthropic could move up the stack from models into agent tooling. Internal engineering teams could build narrower workflows with better knowledge of local systems.[8]
StackAI's advantage was its combination of model choice, deployment flexibility, controls, connectors, and implementation help. Its structural weakness was that every adjacent platform could absorb part of that bundle. A generic canvas faced pricing pressure from automation vendors, distribution pressure from cloud suites, and technical pressure from model providers. StackAI could win difficult deployments, but horizontal category ownership required it to out-distribute companies that already owned the workflow, data, or model layer.
Business Model
The current public pricing page shows a free plan with 500 runs per month, two projects, one seat, and Discord support. Enterprise pricing is custom and includes unlimited projects, custom runs and seats, dedicated infrastructure, solution engineers, on-premises or private-cloud deployment, single sign-on, and compliance features.[9]
This points to an enterprise software and services hybrid. The free tier made evaluation easy, but regulated deployments required solution engineering and repeated co-building. Public sources do not disclose revenue, annual recurring revenue, retention, gross margin, contract values, or customer-acquisition cost. Any claim that the company had already proven standalone software economics would exceed the evidence.
An inference is still possible. Asana said the acquisition added roughly 50 employees across engineering and AI-focused go-to-market roles, while YC later listed 62 people.[4][1] That organization could support high-value contracts, but it also suggests meaningful labor per deployment. The acquisition may have offered a better path to distribute that work than continuing to finance a standalone horizontal sales motion.
Traction
By May 2024, TechCrunch reported more than 200 customers after about six months of a working product.[5] StackAI's later customer page claims $86,000 in annual savings for a regional bank, 8,000 hours per month for a global bank, more than 475,000 hours for LifeMD, and 24 MIT assistants launched in two weeks that saved over 500 student hours.[10]
Those numbers are vendor-reported case-study claims, not independently audited results. They still reveal the company's sales logic: justify custom deployment through labor or cycle-time savings large enough to fund an enterprise contract. Funding followed that traction. StackAI raised $3 million by May 2024, then announced a $16 million Series A one year later; TechCrunch later estimated total funding before acquisition at just under $20 million.[5][8]
Post-Mortem
StackAI did not fail in the ordinary sense. This section examines why acquisition became the stronger strategic endpoint for a working product, a growing team, and a category converging around larger platforms.
A valuable execution layer
Asana's acquisition logic was unusually concrete. It said StackAI could orchestrate work across ERP, CRM, ITSM, data systems, and teams. In Asana's own proof of concept, StackAI pulled live data from five marketing systems, summarized it, then handed work to AI Teammates.[6] That is the missing bridge between a work graph and an agent that can act outside it.
Rosinol's acquisition-day formulation was blunt: “General-purpose agents talk; specialized agents act.”[6] The sentence captures the primary mechanism. Models were becoming commodities; governed execution inside enterprise systems was not.
The transaction reflected more than a distressed team transfer. Asana disclosed about $75 million in cash consideration, subject to customary adjustments, and planned equity awards for employees. Its earnings transcript described an equity-based earnout opportunity and roughly 50 incoming employees across engineering and AI-focused go-to-market.[3][4] StackAI's product and brand also continued after the deal.
Strategic compression
The non-obvious mechanism was strategic compression. StackAI had already paid the technical and organizational cost of making agents useful across fragmented enterprise systems. Asana already owned team context, approvals, work history, governance, distribution, and customer relationships. Combining them removed duplicated work on both sides: StackAI no longer needed to build a broad enterprise distribution layer, and Asana did not need to recreate cross-system orchestration from scratch.
The counterargument is that StackAI could have remained independent and grown into the neutral agent layer across all work software. Its model choice, deployment options, and connector breadth support that case. The problem was not a lack of value. It was the cost of staying neutral while competing with platforms that owned distribution and could bundle adjacent capabilities. Each complex customer also pulled StackAI toward forward deployment and workflow-specific service work. The acquisition gave that labor a larger installed base over which to spread.
What remains unknown
Public evidence does not establish StackAI's revenue, retention, margins, acquisition earnout formula, or investor returns. It also does not show whether the product will remain a separate SKU or become infrastructure inside Asana AI Studio. The website, docs, customer pages, pricing page, and status page remained live as of the research cutoff, which supports continuity but not commercial independence.[6]
Key Lessons
- Integration depth created acquisition value. StackAI's canvas was legible, but connectors, private deployment, approvals, and implementation support made it useful to regulated enterprises. Builders should distinguish the visible interface from the operational layer a buyer cannot cheaply reproduce.
- Forward deployment can prove demand and constrain scale. Co-building helped StackAI turn agent demos into production workflows. The same labor made a broad independent platform harder to scale, especially as automation, cloud, and model vendors converged on the category.
- A complement can be worth more inside the right system. StackAI supplied execution across enterprise software; Asana supplied team memory, work context, approvals, and distribution. The $75 million cash component reflected the value of closing that gap faster than internal development.
- Acquisition is not evidence of completed economics. Named customers, funding, and vendor-reported savings show demand, but public sources disclose no revenue, retention, or margin data. The correct conclusion is that StackAI built a strategically valuable wedge, not that it proved every part of a standalone business.