As companies continue to seek ways to harness artificial intelligence for concrete productivity gains, a company called Writer offers AI tools specifically geared toward getting things done at the enterprise level.
Writer’s AI systems can connect to a wide variety of business software, including standard productivity tools from Google, Salesforce, and Microsoft, as well as a range of database systems. And customers can customize on a granular level what data the AI—and the humans using it—has access to read and write.
But Writer’s platform is also specifically designed to enable white-collar workers without an engineering background to reliably get things done without having to rely on their IT or software development colleagues to write code, set up new systems, or troubleshoot newly automated operations.
“So much of what enterprises are doing with AI is engineering-led,” says cofounder and CEO May Habib. “And with Writer, we are introducing a business-led paradigm, and that has been incredibly important.”
Writer enables companies to create what it calls skills, representing specific tasks that an AI agent can be trained to do, like bringing content into compliance with branding and legal requirements, classifying data, or performing financial analyses. The company provides more than 200 prebuilt skills, including some targeting specific industries like retail, healthcare, and financial services, and businesses can also build their own. Skills can then be harnessed in AI playbooks, which are built around detailed, step-by-step instructions for AI to follow in more complex tasks. Those can include drafting content, conducting research and building internal dashboards, onboarding customers, or setting up marketing content or slide presentations.
At the core of both skills and playbooks are written prompts to AI agents, along with other materials like example output for reference or assets to be used in dashboards or presentations. And while users can craft their own from scratch, Writer is now enabling users to also use AI to actually create and modify skills and playbooks by describing what they want these tools to do.
“It’s designed to be as simple as possible,” says Doris Jwo, VP of product at Writer. “So you can start with something as basic as a one-sentence prompt.”
Users can then test out how new skills and those complex playbooks perform, giving feedback to AI or making manual edits to optimize their operation. When playbooks get executed, users can see the AI describe what it’s doing in plain language, as well as more detailed information about its precise interactions with other systems, like SQL code sent to database engines. And the system also features facilities to learn and mimic a company’s distinct corporate voice, designed for content generation, and constructing a “knowledge graph” of trustworthy information.
Writer, which announced in November 2024 it had raised $200 million in a Series C funding round at a $1.9 billion valuation, points to prominent enterprise customers including Vanguard, Marriott, Dropbox, and Clorox. At Clorox, Writer’s AI is used for tasks including maintaining thousands of listings for the company’s sprawling array of products, from cleaning materials to Burt’s Bees lip balms and Hidden Valley salad dressings, across major e-commerce retailers.
Each retail site has its own requirements, from character limits for particular fields to terms that can’t be used in listings, and keeping product listings updated across products old and new can be a never-ending task, says Matt Harker, VP for consumer experience transformation at Clorox. Those various requirements can now be defined in Writer skills, with playbooks then built to craft those listings for various online stores. Humans are still involved in reviewing the AI output, but the AI can help by flagging potential areas of concern, and time savings so far are substantial, Harker says.
“It’s north of 85% savings in terms of time and tasks that go away,” he says.
That time savings means more time can be devoted to continuing to refine copy and images, including customizing online content for seasonal events like back-to-school shopping or the holiday season, he says. And playbooks and skills can be easily refined when stores inevitably change their listing requirements and formats, making it easy to update product listing fields to stay in compliance.
“Because the skill has learned the new requirement, you run the playbook, all of the product detail pages are now updated and in compliance, and you publish,” Harker says.
While Habib acknowledges that the AI market is a crowded one, with many companies testing and deploying different AI systems for different purposes—“every buyer in the enterprise systems is keeping all bets open,” she says—she believes reliability, enterprise-grade security features, traceability around what the AI is doing, and ease of use are continuing to win customers over. For many tasks, Writer also uses its own line of large language models, dubbed Palmyra, which Habib says are cheaper per API token than those from big competitors, enabling big savings for customers with many complex tasks.
And though a wide range of software tools, from Google and Microsoft Office Suite to productivity packages like Slack and Asana, are now effectively competing with AI-first vendors to be the go-to enterprise gateway to artificial intelligence, Habib says another advantage of Writer is that it’s effectively a neutral operator, able to connect without bias to the wide range of cloud-based applications that businesses use.
“I do think the legacy SaaS companies are doing a great job adding agentic layers to their products, but what the enterprise really needs is systems that can think a lot more and act a lot more like their employees can,” she says. “All of the chair-swiveling between the multiple systems that all of us are doing every day, we need agents who are able to do that, and that’s what we’re building.”
