Automat
Agentic automation that replaces manual work and brittle legacy robotic process automation.
Website: https://runautomat.com/
Cover Block
Open sources
| Field | Value |
|---|---|
| Name | Automat |
| Tagline | Agentic automation that replaces manual work and brittle legacy robotic process automation. [Y Combinator, 2026] |
| Headquarters | San Francisco, CA [Y Combinator, 2026] |
| Founded | 2022 [Y Combinator, 2026] |
| Stage | Seed [Y Combinator, 2026] |
| Industry | Other |
| Technology | AI / Machine Learning |
| Founding Team | Lucas Ochoa, Gautam Bose [Y Combinator, 2026] |
| Funding Label | Seed |
| Total Disclosed | $15.5M Series A announced in November 2025, following earlier seed financing reported by Y Combinator and TechCrunch-linked company coverage [Axios, November 2025] [TechCrunch, September 2023] |
Links
Open sources
- Website: https://runautomat.com/
- LinkedIn: https://www.linkedin.com/in/lucasrunautomat/
- Other: https://www.ycombinator.com/companies/automat
What an Investor Needs First
PUBLIC Automat is building enterprise workflow automation from video demonstrations and plain-language instructions, a product angle that merits attention because it aims to move beyond brittle scripted RPA at a moment when buyers are again testing AI-led automation in production settings [TechCrunch, September 2023] [Axios, November 2025]. The company was founded in 2022, and public reporting ties the origin story to Lucas Ochoa and Gautam Bose, who met at Carnegie Mellon University and later worked together at Google Creative Lab before starting Automat [TechCrunch, April 2023] [Y Combinator, 2026].
The product has been described publicly as software that can turn screen recordings into workflow automations, with users able to provide either a video demonstration or natural-language instructions; more recent company positioning frames the broader ambition as agentic automation for enterprise workflows [TechCrunch, September 2023] [TechCrunch, April 2023] [Axios, November 2025]. That is a credible narrative wedge, but the public record still says more about promise than about deployed outcomes, and key claims around contextual reasoning, error recovery, and self-improvement rely largely on company-controlled sources [Y Combinator, 2026] [Automat, 2026].
The team reads as product-forward rather than sales-forward. TechCrunch reported that Ochoa and Bose spent roughly three and a half years at Google Creative Lab working on products including AI Test Kitchen, Teachable Machine, and Pixel Buds Pro, which supports the view that Automat was founded by operators with hands-on experience shipping user-facing AI products [TechCrunch, September 2023].
On financing, the cleanest confirmed public milestone is a $15.5 million Series A led by Felicis, reported by Axios in November 2025 and also covered by FinSMEs the same month; Y Combinator is also publicly associated with the company through its accelerator profile [Axios, November 2025] [FinSMEs, November 2025] [Y Combinator, 2026]. The business model is less clear from public materials, though the current website suggests a mix of software and service delivery through "AI Teammates and Managed Automations," which could help early adoption but may complicate margin profile and implementation scalability if services remain material [Automat, 2026].
Over the next 12 to 18 months, the main public markers to watch are whether Automat can translate its product narrative into visible enterprise traction, clarify where software ends and managed service begins, and show that the post-RPA pitch produces repeatable adoption rather than bespoke deployments [Axios, November 2025] [Automat, 2026]. The setup is promising, particularly given the founder-market fit around workflow capture and interface design, but the public evidence base is still thin on customers, metrics, and operating proof points.
Partially corroborated -- Core company history and product framing are corroborated by TechCrunch, Axios, Y Combinator, and company materials, but several material operating claims remain company-sourced and traction data is limited.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Industry / Vertical | Other |
| Technology Type | AI / Machine Learning |
| Founding Team | Solo Founder |
| Funding | Seed, total disclosed approximately $500,000; later Series A of $15.5 million publicly reported |
Inside the Company
PUBLIC
Automat appears to have been formed in 2022 and is now based in San Francisco, with the public record pointing to a company built around workflow automation learned from a product-minded founder bench rather than from a legacy RPA lineage [Y Combinator, 2026]. Y Combinator's company profile identifies Automat as founded by Lucas Ochoa and Gautam Bose, while also describing the product thesis as "agentic automation" that can understand context, recover from errors, and improve over time [Y Combinator, 2026].
The founding story is more concrete than the operating history. TechCrunch reported in April 2023 that Ochoa and Bose met as students at Carnegie Mellon University, and in September 2023 reported that the pair had been hired together out of college at Google Creative Lab, where they spent roughly three and a half years working on products including AI Test Kitchen, Teachable Machine, and Pixel Buds Pro [TechCrunch, April 2023] [TechCrunch, September 2023]. By September 2023, TechCrunch described Automat, then also referred to as Lasso, as a company that could turn videos into workflow automations, which marks the clearest public milestone in its early product evolution [TechCrunch, September 2023].
The next dated milestone in the public record is financing. Y Combinator lists Automat as one of its companies, establishing the accelerator affiliation, and Axios reported in November 2025 that Automat had raised a $15.5 million Series A led by Felicis to automate enterprise workflows [Y Combinator, 2026] [Axios, November 2025]. No state filing or company-published legal entity name was provided in the source set used for this section, so the legal entity remains unstated here.
Partially corroborated -- Confirmed in part by Y Combinator and TechCrunch; headquarters, founding year, and founding team are public, but legal entity detail is not established in the cited source set.
Under the Hood
MIXED
Automat is presenting itself as an AI automation platform aimed at replacing manual workflow execution and older rule-based robotic process automation, although the strongest independently reported product description is narrower: the company can turn video demonstrations into workflow automations for enterprise use [TechCrunch, September 2023] [Axios, November 2025]. TechCrunch reported in 2023 that users could provide a workflow either through a screen recording or through natural-language instructions, which gives the product a practical wedge into process capture rather than requiring customers to author automations from scratch [TechCrunch, April 2023] [TechCrunch, September 2023]. The broader claims that Automat's agents understand context, recover from errors, and improve over time come from company-controlled or low-confidence sources, so they are better read as positioning than as independently validated capability [Y Combinator, 2026] [Automat, 2026].
Public materials also suggest the product surface has expanded since the earlier Lasso branding. The company website in 2026 describes two offers, Automat Workforce, framed as an AI teammate called Ace, and Managed Automation, where Automat's team builds and maintains automations for customers [Automat, 2026]. Separate company and LinkedIn references also point to AI document extraction and to an "external engine" for business logic, custom interfaces, dashboards, and document generation, but those descriptions are company-originated and do not yet appear in independent reporting or a verified public demo in the source set [LinkedIn, 2026] [Automat, 2026]. That leaves the core product thesis reasonably clear, enterprise workflow automation learned from demonstrations and text, while the breadth of the current platform remains only partially corroborated.
Claim stands unchecked -- This section mixes independent reporting from TechCrunch and Axios with several material product claims that are only described in company-controlled sources such as Automat's website and Y Combinator profile.
Market Research
PUBLIC
Enterprise workflow automation is drawing fresh budget because large language models have widened the range of tasks software can plausibly execute, but the public record here is thinner than investors would usually want for a clean market sizing exercise. No named third-party TAM, SAM, or SOM specific to Automat appears in the supplied materials, so the market has to be framed through adjacent categories that the company is clearly targeting: enterprise workflow automation, robotic process automation replacement, and AI-assisted task execution inside back-office and knowledge-work flows [TechCrunch, September 2023] [TechCrunch, April 2023] [Axios, November 2025]. That framing matters because Automat is presented in public sources as a system that turns screen recordings or natural-language instructions into workflow automations, which places it at the overlap of RPA, process intelligence, and newer agent-based enterprise software rather than in a narrowly defined single-purpose SaaS category [TechCrunch, September 2023] [TechCrunch, April 2023].
The demand signal in public coverage is qualitative, but it is consistent. TechCrunch's 2023 reporting described Automat, then referred to in coverage around Y Combinator, as software that could convert videos into workflow automations, while a separate TechCrunch roundup said the product could be used by sending a workflow in a screen recording or with natural language, both of which point to a buyer desire to reduce the setup cost that has historically limited automation deployments [TechCrunch, September 2023] [TechCrunch, April 2023]. Axios's 2025 funding report places the company more directly in enterprise workflow automation, with the pitch centered on moving beyond brittle scripts and manual processes, which aligns with a broader market shift from rules-based automation toward systems that can absorb messier inputs and exception handling [Axios, November 2025].
| Market lens | Public evidence | Relevance to Automat |
|---|---|---|
| Legacy RPA substitution | Automat is described as replacing brittle legacy robotic process automation with agentic automation, according to company and directory descriptions [Y Combinator, 2026] [Automat, 2026] | Suggests spending may come from existing automation budgets rather than entirely new line items |
| Workflow capture and orchestration | Public reporting says the product turns videos or plain-text instructions into workflow automations [TechCrunch, September 2023] [TechCrunch, April 2023] | Points to a wedge in process capture, training, and automation authoring |
| Enterprise AI assistants | Automat's website describes "AI Teammates" and managed automations, according to the company site [Automat, 2026] | Expands the addressable use case from script creation to ongoing task execution and service delivery |
The table points to an important nuance: this is likely a budget reallocation story as much as a net-new software category. If the product works as described, Automat is competing for dollars currently spent on RPA tooling, internal operations labor, outsourced process work, and emerging AI copilots, not just for a standalone "agent platform" budget [Axios, November 2025] [TechCrunch, September 2023].
Adjacent markets matter because the product description spans more than one buyer problem. Public materials suggest at least three neighboring spend pools: document extraction, which appears in a LinkedIn reference to an Automat self-service tool; managed automation services, which the company site describes as built, run, and maintained by its team; and enterprise application-layer tooling for dashboards, business logic, and document generation, described on a company web page [LinkedIn, 2026] [Automat, 2026]. The overlap is strategically useful because automation buyers often begin with a narrow pain point and then expand into broader workflow coverage, but it also complicates market definition because these are distinct categories with different incumbents, sales motions, and margins [Automat, 2026].
Macro and regulatory forces are supportive in broad terms, though the available sources do not tie Automat to any single compliance-driven wedge. Enterprise buyers continue to face pressure to improve productivity without proportionate headcount growth, and automation projects that can be demonstrated through existing user behavior, such as screen recordings, may be easier to pilot than traditional process-rewrite efforts, according to how the product is described in public reporting [TechCrunch, April 2023] [TechCrunch, September 2023]. At the same time, the more a vendor handles enterprise workflows and documents, the more procurement scrutiny tends to shift toward data handling, auditability, and error recovery. Public sources say Automat's agents are intended to understand context and recover from errors, but those claims are still mostly company-asserted rather than independently validated, which limits how far the market argument can be pushed from public evidence alone [Y Combinator, 2026] [Automat, 2026].
Partially corroborated -- Market framing is supported by TechCrunch and Axios reporting, but no named third-party market sizing data was provided in the source set, and parts of the category definition rely on company and Y Combinator descriptions.
Competition and Substitutes
Competitive Positioning
MIXED Automat is positioning itself against two different alternatives at once: legacy robotic process automation on one side, and newer AI-native workflow tools that learn from demonstrations or plain-language instructions on the other [TechCrunch, September 2023] [Axios, November 2025] [Y Combinator, 2026].
The public record is thin on named direct rivals in Automat's own materials, but the functional comparison is still clear enough. The company began by pitching a product, then called Lasso, that could turn screen recordings into workflow automations, with users either sending a workflow in a video or describing it in natural language [TechCrunch, April 2023] [TechCrunch, September 2023]. That puts it in the same broad buying conversation as incumbent RPA platforms such as UiPath and Automation Anywhere, both of which were built around scripted flows and structured process mapping, even though those companies are not named in Automat's source set and therefore cannot be benchmarked here on funding or product specifics from this dataset. It also places Automat near a newer cluster of AI automation products that promise more flexible execution than rule-based bots, though the available sources do not identify specific startups in that cluster by name [Axios, November 2025].
That distinction matters because Automat is not merely selling task automation in the abstract. According to TechCrunch's 2023 reporting, the product's wedge was usability: capture a process in video, add text instructions, and let the system infer the workflow [TechCrunch, April 2023] [TechCrunch, September 2023]. If that workflow generation works reliably, Automat competes less on classic RPA implementation depth and more on time-to-automation for non-technical teams. If it does not, buyers can fall back to known substitutes, namely human operations teams, outsourced process execution, or incumbent automation software that is slower to configure but easier to govern once deployed. The competitive map, then, is best read as incumbents owning control and procurement familiarity, adjacent service providers owning execution certainty, and AI-native challengers trying to win on setup speed and adaptability.
The strongest defensible asset visible in public sources today is team credibility, not yet distribution. Lucas Ochoa and Gautam Bose previously worked together at Google Creative Lab, where TechCrunch reported they spent three and a half years helping launch products including AI Test Kitchen, Teachable Machine, and Pixel Buds Pro [TechCrunch, September 2023]. That background is relevant because demonstration-driven automation depends on product design, model behavior, and human-computer interaction, not only on back-end workflow plumbing. Y Combinator backing also provides a modest distribution and recruiting signal, and Axios reported a $15.5 million Series A led by Felicis in November 2025, which suggests enough capital to keep iterating if early enterprise deployments are promising [Y Combinator, 2026] [Axios, November 2025]. Still, this edge looks perishable rather than durable at this stage. Ex-Google talent and early investors can open doors, but they do not create the kind of data moat, ecosystem lock-in, or compliance estate that historically protects automation vendors over time.
Automat appears most exposed wherever reliability, governance, and enterprise integration matter more than workflow capture elegance. The company says its agents understand context, recover from errors, and improve over time, but those claims are presently company-led rather than independently validated in the source set [Y Combinator, 2026]. By contrast, established enterprise automation vendors typically benefit from longer implementation histories, larger partner ecosystems, and procurement familiarity, even if their tools are less intuitive to configure. Automat is also exposed to adjacent AI copilots and general-purpose model vendors that can add automation layers into broader productivity suites, because those companies may control the customer relationship before a standalone automation startup gets considered. Public sources do not yet show Automat owning a channel, a regulated niche, or a proprietary dataset that would shut out those entrants [Automat, 2026] [Axios, November 2025].
The most plausible 18-month scenario is a split market rather than a single category winner. UiPath is the likely winner if enterprise buyers continue to prioritize governance, auditability, and broad systems integration over ease of workflow capture, because those are the conditions that usually favor established automation stacks. Automat is the likely winner if video- and text-based automation authoring reduces deployment friction enough for business teams to automate work without heavy process engineering, especially inside mid-market or departmental use cases where speed matters more than central IT standardization [TechCrunch, April 2023] [TechCrunch, September 2023]. The likeliest loser if large model platforms commoditize workflow generation is any startup whose differentiation rests mostly on natural-language orchestration without unique execution infrastructure or proprietary operational data. Based on the public record, that is the strategic pressure Automat still needs to answer with proof points rather than narrative.
Partially corroborated -- Section relies on TechCrunch and Axios for independent reporting, with additional company and Y Combinator materials for current positioning.
Opportunity
PUBLIC
The prize here is straightforward: if Automat can turn workflow capture from a bespoke services exercise into reliable software, it could become a meaningful control layer for enterprise back-office automation, a category with budgets that have already supported multibillion-dollar outcomes for incumbents such as UiPath [TechCrunch, September 2023] [Axios, November 2025].
The headline opportunity is not that Automat merely adds another interface to automation. It is that the company is pursuing a simpler input method, video demonstrations and plain-language instructions, for a problem that enterprises already spend heavily to solve [TechCrunch, April 2023] [Axios, November 2025]. TechCrunch reported in 2023 that Automat, then referred to as Lasso, could turn videos into workflow automations, and that users could provide a screen recording or natural-language prompt rather than manually script flows [TechCrunch, September 2023] [TechCrunch, April 2023]. Axios then reported in 2025 that Automat raised a $15.5 million Series A led by Felicis to help automate workflows, which matters because it is a signal that at least one established venture firm saw enough product and market progress to underwrite a larger institutional round [Axios, November 2025]. That does not prove repeatable enterprise adoption, but it does make the upside case reachable rather than purely conceptual.
A useful way to frame the upside is through a few distinct paths, all of which depend on the same core bet: automation built from observed human work is easier to deploy and maintain than brittle rule-based RPA [TechCrunch, September 2023] [TechCrunch, April 2023].
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Workflow capture becomes the wedge | Automat becomes the preferred tool for teams that want to create automations from screen recordings and text prompts, then expands from single workflows into broader enterprise process coverage | A product improvement that consistently converts recorded workflows into dependable automations across common business systems | The product concept was described by TechCrunch in 2023, and the company is still presenting the same core thesis around context-aware automation in 2026, suggesting continuity rather than a sharp pivot [TechCrunch, September 2023] [Y Combinator, 2026] |
| Managed automation becomes a services-to-software bridge | Automat uses a managed offering to get into larger accounts, learns failure modes, and gradually converts custom work into reusable product modules | The current positioning of both AI teammates and managed automation on the company site, if translated into standardized playbooks | The company website describes two lines, Automat Workforce and Managed Automation, which is consistent with an entry strategy where service delivery informs product scope [Automat, 2026] |
| Enterprise orchestration layer | Automat moves from workflow generation into a broader layer for business logic, dashboards, and document generation, giving it a larger share of customer processes | Successful expansion from narrow task automation into adjacent workflow surfaces inside the same accounts | The company has publicly described itself as an external engine for business logic, custom interfaces, dashboards, and document generation, which implies ambition beyond point automation [Automat, 2026] |
The compounding mechanism, if this works, would likely come from three reinforcing loops. First, every successful workflow capture should reduce onboarding friction for the next one, because customers would start from examples instead of building from scratch [TechCrunch, April 2023]. Second, a managed automation layer could expose where automations fail in production, creating a path to improve recovery and reliability over time, a capability the company and Y Combinator both describe as central to the product [Y Combinator, 2026] [Automat, 2026]. Third, if customers begin to trust Automat for one repetitive process, the same buyer can extend usage into neighboring processes with lower sales cost than a net-new account motion. Public evidence does not yet confirm this flywheel is active at scale, but the product design and current packaging are directionally aligned with it [Automat, 2026] [Axios, November 2025].
The size of the win is large enough to matter even under a conservative scenario frame. UiPath went public in a market that had already shown enterprises will pay real budgets for workflow automation infrastructure, even if next-generation entrants attack the market with a different technical approach [TechCrunch, September 2023]. More specifically for Automat, a plausible upside case is that it grows into a differentiated enterprise automation platform with meaningful software revenue and strategic value to larger workflow, CRM, or service-management vendors. If the enterprise orchestration scenario plays out, Automat could support a multibillion-dollar outcome over time (scenario, not a forecast), because the company is targeting the same broad budget line, workflow automation, while trying to remove the implementation friction that limited earlier RPA systems [Axios, November 2025] [TechCrunch, April 2023]. The gap between that outcome and current public evidence remains wide, but the path itself is legible.
Partially corroborated -- This section relies on TechCrunch and Axios reporting for product concept and financing, with supplemental company and Y Combinator materials for current positioning.
Sources
Open sources
[Y Combinator, 2026] Automat: Agentic automation that understands context, recovers, and improves. | https://www.ycombinator.com/companies/automat
[Axios, November 2025] Automat raises $15.5M led by Felicis to help automate workflows | https://www.axios.com/pro/enterprise-software-deals/2025/11/19/automat-15-million-enterprise-workflows
[TechCrunch, September 2023] Y Combinator-backed Automat turns videos into workflow automations | https://techcrunch.com/2023/09/06/y-combinator-backed-automat-turns-videos-into-workflow-automations/
[TechCrunch, April 2023] These Y Combinator-backed startups are trying to build 'ChatGPT for X' | https://techcrunch.com/2023/04/04/these-y-combinator-startups-are-trying-to-build-chatgpt-for-x/
[Automat, 2026] Automat | AI Teammates and Managed Automations | https://runautomat.com/
[LinkedIn, 2026] Bhaskar Bose, PhD - AiXMotion, Inc. | LinkedIn | https://www.linkedin.com/in/bboseai/
[FinSMEs, November 2025] Automat Raises $15.5M in Series A Funding | https://finsmes.com/2025/11/automat-raises-15-5m-in-series-a-funding.html
Articles about Automat
- Automat's $15.5 Million Series A Funds a Bet on Video as the New Script — The Y Combinator alum, founded by ex-Google engineers, is using screen recordings to build agentic automation that aims to replace brittle RPA.