Author, a San Francisco-based synthetic intelligence startup, is launching a unified AI agent platform designed to let any worker automate advanced enterprise workflows with out writing code — a functionality the corporate says distinguishes it from consumer-oriented instruments like Microsoft Copilot and ChatGPT.
The platform, referred to as Author Agent, combines chat-based help with autonomous activity execution in a single interface. Beginning Tuesday, enterprise clients can use pure language to instruct the AI to create shows, analyze monetary information, generate advertising campaigns, or coordinate throughout a number of enterprise techniques like Salesforce, Slack, and Google Workspace—then save these workflows as reusable "Playbooks" that run mechanically on schedules.
The announcement comes as enterprises battle to maneuver AI initiatives past pilot packages into manufacturing at scale. Author CEO Could Habib has been outspoken about this problem, not too long ago revealing that 42% of Fortune 500 executives surveyed by her firm stated AI is "tearing their firm aside" because of coordination failures between departments.
"We're delivering an agent interface that’s each extremely highly effective and radically easy to remodel particular person productiveness into organizational affect," Habib stated in an announcement. "Author Agent is the distinction between a single gross sales rep asking a chatbot to write down an outreach e mail and an enterprise guaranteeing that 1,000 reps are all sending on-brand, compliant, and contextually-aware messages to focus on accounts."
How Author is placing workflow automation within the fingers of non-technical staff
The platform's core innovation facilities on making workflow automation accessible to non-technical workers—what Author executives name "democratizing who will get to be a builder."
In an unique interview with VentureBeat, Doris Jwo, Author's director of product administration, demonstrated how the system works: A person varieties a request in plain English — for instance, "Create a two-page partnership proposal between [Company A] and [Company B], make it a branded deck, embody affect metrics and partnership tiers."
The AI agent then breaks down that request into discrete steps, conducts internet analysis, generates graphics and charts on the fly, creates particular person slides with sourced data, and assembles a whole presentation. The whole course of, which could take an worker hours or days, may be accomplished in 10-12 minutes.
"The agent principally seems on the request, breaks it down, does analysis, understands what items it wants, creates an in depth plan at a step-by-step degree," Jwo defined throughout a product demonstration. "It would say, 'I must do internet analysis,' or 'This person wants data from Gong or Slack,' and it reaches out to these connectors, grabs the info, and executes the plan."
Crucially, customers can save these multi-step processes as Playbooks—reusable templates that colleagues can deploy with a single click on. Routines enable these Playbooks to run mechanically at scheduled intervals, primarily placing information work "on autopilot."
Safety and compliance controls: Author's reply to enterprise IT issues
Author positions these enterprise-focused controls as a key differentiator from rivals. Whereas Microsoft, OpenAI, and Anthropic provide highly effective AI capabilities, Author's executives argue these instruments weren't designed from the bottom up for the safety, compliance, and governance necessities of enormous regulated organizations.
"All the merchandise you talked about are nice merchandise, however even Copilot may be very a lot targeted on private productiveness—summarizing e mail, for instance, which is essential, however that's not the element we're specializing in," stated Matan-Paul Shetrit, Author's director of product administration, in an unique interview with VentureBeat.
Shetrit emphasised Author's "belief, safety, and interoperability" strategy. IT directors can granularly management what the AI can entry — for example, stopping market analysis brokers from mentioning rivals, or limiting which workers can use internet search capabilities. All exercise is logged with detailed audit trails exhibiting precisely what information the agent touched and what actions it took.
"These fine-grained controls are what make merchandise enterprise-ready," Shetrit stated. "We will deploy to tens of 1000’s or lots of of 1000’s of workers whereas sustaining the safety and guardrails you want for that scale."
This structure displays Author's origin story. In contrast to OpenAI or Anthropic, which began as analysis labs and later added enterprise choices, Author has focused Fortune 500 firms since its 2020 founding. "We're not a analysis lab that went to client and is dabbling in enterprise," Shetrit stated. "We’re before everything concentrating on the International 2000 and Fortune 500, and our analysis is in service of those clients' wants."
Inside Author's technique to attach AI brokers throughout enterprise software program techniques
A essential technical element is Author's strategy to system integrations. The platform consists of pre-built connectors to greater than a dozen enterprise purposes—Google Workspace, Microsoft 365, Snowflake, Asana, Slack, Gong, HubSpot, Atlassian, Databricks, PitchBook, and FactSet—permitting the AI to retrieve data and take actions throughout these techniques.
Author constructed these connectors utilizing the Mannequin Context Protocol (MCP), an rising normal for AI system integrations, however added what Shetrit described as an "enterprise-ready" layer on prime.
"We took a first-principle strategy of: You have got this MCP connector infrastructure—how do you construct it in a means that's enterprise-ready?" Shetrit defined. "What we have now right this moment within the trade is unquestionably not it."
The system can write and execute code on the fly to deal with sudden situations. If a person uploads an unfamiliar file format, for example, the agent will generate code to extract and course of the textual content with out requiring a human to intervene.
Jwo demonstrated this functionality with a day by day workflow she runs: Each morning at 10 a.m., a Routine mechanically summarizes her Google Calendar conferences, identifies exterior members, finds their LinkedIn profiles, and sends the abstract to her by way of Slack — all with out her involvement.
"This was fairly easy, however you’ll be able to think about for a salesman it’d say, 'On the finish of the day, wrap up a abstract of all of the calls I had, ship me motion objects, publish it to the account-specific Slack channel, and tag these people to allow them to accomplish these workflows,'" Jwo stated. "That may run repeatedly every day, every week, or on demand."
From mortgage lenders to CPG manufacturers: Actual-world AI agent use instances throughout industries
The platform is attracting clients throughout a number of industries. New American Funding, a mortgage lender, makes use of Author Agent to automate advertising workflows. Senior Content material Advertising and marketing Supervisor Karen Rodriguez uploads Asana venture tickets with inventive briefs, and the AI executes duties like updating e mail campaigns or remodeling articles into social media carousels, video scripts, and captions.
Different use instances span monetary companies groups creating funding dashboards with PitchBook and FactSet information, client packaged items firms brainstorming new product traces primarily based on social media traits, and advertising groups producing partnership shows with branded belongings.
Author has added clients together with TikTok, Comcast, Keurig Dr Pepper, CAA, and Aptitude Well being, becoming a member of an present base that features Accenture, Qualcomm, Uber, Vanguard, and Marriott. The corporate now serves greater than 300 enterprises and has secured over $50 million in signed contracts, with projections to double that to $100 million this 12 months.
The startup's internet retention charge — a measure of how a lot present clients increase their utilization — stands at 160%, that means clients on common enhance their spending by 60% after preliminary contracts. Twenty clients who began with $200,000-$300,000 contracts now spend about $1 million yearly, in keeping with firm information.
'Vibe working': Author's imaginative and prescient for AI-powered productiveness past coding
Author executives body the platform as enabling what they name "vibe working" — a playful reference to the favored time period "vibe coding," which describes AI instruments like Cursor that dramatically speed up software program improvement.
"We used to name it transformation once we took 12 steps and made them 9. That's optimizing the world as it’s," Habib stated at Author's AI Leaders Discussion board earlier this month, in keeping with Forbes. "We will now create a brand new world. That’s the greenfield mindset."
Shetrit echoed this framing: "Vibe coding is the theme of 2025. Our view is that ‘vibe working’ is the theme of 2026. How do you convey the identical productiveness beneficial properties you've seen with coding brokers into the workspace in a means that non-technical customers can maximize them?"
The platform is powered by Palmyra X5, Author's proprietary giant language mannequin that includes a one-million-token context window — among the many largest commercially obtainable. Author skilled the mannequin for about $700,000, a fraction of the estimated $100 million OpenAI spent on GPT-4, by utilizing artificial information and methods that halt coaching when returns diminish.
The mannequin can course of a million tokens in about 22 seconds and prices 60 cents per million enter tokens and $6 per million output tokens — considerably cheaper than comparable choices, in keeping with firm specs.
Making AI Selections Seen: Author's Strategy to Belief and Transparency
A particular facet of Author's strategy is transparency into the AI's decision-making course of. The interface shows the agent's step-by-step reasoning, exhibiting which information sources it accessed, what code it generated, and the way it arrived at outputs.
"There's a really clear exhibition of how the agent is considering, what it's doing, what it's touching," Shetrit stated. "That is essential for the top person to belief it, but in addition essential for the IT individual or safety skilled to see what's occurring."
This "supervision" mannequin goes past easy observability of API calls to embody what Shetrit described as "a superset of observability" — giving organizations the flexibility to not simply monitor however management AI habits via insurance policies and permissions.
Session logs seize all agent exercise when enabled by directors, and customers can submit suggestions on each output to assist enhance system efficiency. The platform additionally emphasizes offering sources and citations for generated content material, permitting customers to confirm data.
"With any type of chat assistant, agentic or not, belief however confirm is admittedly essential," Jwo stated. "That's a part of the pillars of us constructing this and making it enterprise-grade."
What Author Agent Prices—and Why It's Included within the Base Platform
Author is together with all the brand new capabilities—Playbooks, Routines, Connectors, and Persona customization—as a part of its core platform with out extra prices, in keeping with Jwo.
"That is absolutely included as a part of the Author platform," she stated. "We're not charging extra for utilizing Author Agent."
The "Persona" function permits particular person customers, groups, or whole organizations to customise the AI's communication fashion, guaranteeing generated content material matches model voice and tone tips. This works alongside company-level controls that implement terminology and elegance necessities.
For extremely structured, repetitive duties, Author additionally affords a library of greater than 100 pre-built brokers and an AI Studio for constructing customized multi-agent techniques aligned with particular enterprise use instances.
The Race to Outline Enterprise AI: Can Objective-Constructed Platforms Beat Tech Giants?
The launch crystallizes a elementary rigidity in how enterprises will undertake AI at scale. Whereas consumer-facing AI instruments emphasize particular person productiveness beneficial properties, firms want techniques that work reliably throughout 1000’s of workers, combine with present software program infrastructure, keep regulatory compliance, and ship measurable enterprise affect.
Author's wager is that these necessities demand purpose-built enterprise platforms reasonably than client instruments tailored for enterprise use. The corporate's $1.9 billion valuation — achieved in a November 2024 funding spherical that raised $200 million — suggests traders see benefit on this thesis. Backers embody Premji Make investments, Radical Ventures, ICONIQ Development, Salesforce Ventures, and Adobe Ventures.
But the aggressive panorama stays formidable. Microsoft and Google command monumental distribution benefits via their present enterprise software program relationships. OpenAI and Anthropic possess analysis capabilities which have produced breakthrough fashions. Whether or not Author can keep its differentiation as these giants increase their enterprise choices will take a look at the startup's core premise: that serving Fortune 500 firms from day one creates benefits that analysis labs turned enterprise distributors can not simply replicate.
"We're getting into an period the place in the event you can describe a greater solution to work, you’ll be able to construct it," Jwo stated. "The brand new Author Agent democratizes who will get to be a builder, empowering the operational consultants and inventive problem-solvers in each division to turn out to be the architects of their very own transformation. That's the way you unlock innovation that rivals can't replicate."
The promise is alluring — AI capabilities highly effective sufficient to remodel how work will get achieved, accessible sufficient for any worker to make use of, and managed sufficient for enterprises to deploy safely at scale. Whether or not Author can ship on that promise on the pace and scale required will decide if its imaginative and prescient of "vibe working" turns into the 2026 theme Shetrit predicts, or simply one other formidable try to resolve enterprise AI's execution downside.
However one factor is definite: In a market the place 85% of AI initiatives fail to flee pilot purgatory, Author is betting that the winners gained't be the businesses with probably the most highly effective fashions—they'll be those that make these fashions really work contained in the enterprise.
