At Kovai.co, we are constantly exploring ways to improve how teams create, manage, and deliver knowledge. With our product, Document360, we aim to empower technical writers and documentation teams to keep up with the ever-increasing pace of SaaS development. One of the key challenges we identified is the growing gap between product releases and the availability of corresponding documentation. Our research revealed that nearly 70% of a technical writer’s time is spent in stakeholder meetings from initial ideation with product managers to walkthroughs with designers, developer showcases, and QA sessions, before even beginning a first draft. This often results in a backlog of documentation, delaying product readiness and customer enablement.
This growing need for speed and scale in content creation has led us to explore cutting-edge technologies. One such breakthrough is LLM-based Agents, intelligent systems powered by Large Language Models with reasoning and function-calling capabilities. These Agents can plan, coordinate with software tools, determine the best course of action, and execute tasks autonomously. As more agentic frameworks are released by the open-source community, LLM providers, and tech leaders like Microsoft, we are witnessing a surge in tools that drive significant efficiency across businesses. At Document360, we see immense potential in applying this technology to transform the way product documentation is created, and that’s exactly the problem we set out to solve.
Saravana Kumar, Founder and CEO of Kovai.co says,
Innovation drives us forward. Our evolution to agentic AI positions us at the cutting edge of Artificial Intelligence, where AI doesn’t just respond, it reasons and acts.
The Challenge: When Writing Takes a Backseat
Based on our user research, we also found out that it takes about 10 days to publish new feature documentation in an enterprise. This is astonishing as an enterprise wishes to move faster. Limitation of number of technical writers coupled with increased product feature velocity impedes publication of new documentation. In summary, the problems faced by enterprises are:
- Technical writers spend about 70% of their time in stakeholder meetings rather than writing.
- It takes 2 weeks to publish feature documentation.
- Despite quick approvals, editorial review takes time as reviewers need to check content against corporate style guide.
At Document360, we are passionate about solving business problems using new technology toolkits. We are thrilled to announce that we are introducing “Eddy AI writing Agent” that quickly produces a first draft of an article adhering to the style guide and content templates.
Key Features that Power AI Writing Agents
As part of “Eddy AI Writing Agent”, we are introducing three new functionalities that empower technical writers to produce content quickly:
- How-to article generator that generates step-by-step procedures along with screenshots if meeting videos are uploaded.
- Agent that writes an article based on audio, audio transcriptions and video file.
- Agent that browses the internet and your existing knowledge base to produce new content.
Technical writers can give a prompt to provide context to the “Eddy AI Writing Agent” and specify how they want their article to be written and formatted. Writing Agents do planning in the background, synthesize the information present in different sources and produce new knowledge that is compliant with style guide. They can choose a specific style guide for compliance and template for content structure. Eddy AI Writing Agent is versatile as it offers choice to technical writers on:
- What content needs to be written
- How content should be written
- How content should be formatted
- Which portion of content need more emphasis
More importantly, all administrative overhead on article management such as generating tag, creating article labels, adding SEO meta description, and so on are automatically done by the Writing Agent.
Seamless Design and Workflow
The design is intuitive and simple and available in the default state of category manager. The design provides options for choosing the writing operation, specifying the source, choosing the right style guide and template (if required).
The Business Value
Eddy AI Writing Agent would reduce the time to create a first draft of an article from weeks to a few minutes. Given that produced content complies with corporate style guide, it would eliminate the effort required by Editorial reviewer. The produced content has built-in trust as content is only produced from source artefacts within the boundary of customer knowledge base. Terminologies are used consistently, and brand voice is unanimous. This helps the content operations team to publish the content without any further delay. Eddy AI Writing Agent boosts the productivity of documentation teams and meets the expectations of product teams with increasing product feature velocity.
Our Journey Forward
Building Eddy AI assistant required meticulous effort from our team involving UX designers, product managers, developers, data scientists, solutions architects, and other stakeholders. We pushed ourselves hard in optimizing every single component of Writing Agent to be more robust. We are already working on future roadmap items to expand capabilities of Writing Agent. We’re excited to learn from our community and customers as we evolve our AI Assistant Eddy for the next era of technology.
Customer behavior is also shifting towards accomplishing goals rather than reading a lengthy knowledge base article or mobilizing knowledge gained through interacting with chatbot. Customers are more interested in accomplishing goals and “get things done” quickly rather than acquiring product knowledge. This problem statement sets precedence for what comes next!
So, are you ready to boost your documentation game with Eddy AI Writing Agent?
Check out Eddy in action and explore what’s next.
