Product demonstration
E9 Workflow Form-Filling Agent
Multi-turn completion in natural language · Missing fields recommended each turn · Fill multiple fields at once with one sentence · Real-time readback from 5/5
Chinese narration / original recording
About this demo
E9 Workflow Form-Filling Agent | Understands a sentence, intelligently recommends entries, and helps complete an entire workflow form The core capability in this scenario is not mechanically writing a sentence into a field. It is enabling AI to understand employees’ flexible, conversational language and, based on the fields, required-field rules, existing content, and actual options in the current E9 workflow, determine what is missing, recommend what to enter, and keep helping the user complete the form. Users don’t need to remember field names, figure out the form structure, or follow a fixed command template. The BoltFusion E9 Workflow Form-Filling Agent sits beside the workflow and lets users get started just as they would when talking with a colleague. This live demo took just three rounds of conversation: “Help me fill out a general approval request. Put ‘office network optimization’ as the application details.” “Use the default department.” “Write ‘optimize the office network’ in the details, and make me the approver.” BoltFusion AI completes the following steps automatically: Understand natural language → Read the E9 page context → Match business meaning to fields → Identify actual options → Determine missing fields → Recommend entries → Let the user confirm or adjust → Populate the native form → Read the page again → Verify all required fields are complete. After the first round of entries, AI doesn’t stop at “written to the form.” It continues analyzing the current form, proactively lists any required fields that are still blank, and recommends entries such as the applicant’s department, approval details, and approver based on business context and the actual options on the page. “Default” isn’t text to copy directly into the form; it means matching an actual department option on the current E9 page. “Make me the approver” isn’t ordinary text either; it requires selecting the approver based on the currently logged-in user. The last sentence contains two intents—approval details and approver—and AI understands and fills them both in one go. After each round, the agent rereads the current form. Completed entries stay in place, and it proactively explains what’s still missing and suggests entries the user can apply directly. Users can accept the suggestions or change any item in natural language. They don’t need to think of everything at once—the agent helps fill in the form step by step over multiple turns. These recommendations aren’t answers generated in isolation. They’re grounded in the current E9 workflow context, field rules, and real business data. AI handles understanding, matching, and recommendations; users confirm the final entries, which are then returned directly to the familiar native E9 business page. This goes beyond simply entering a few fields for the user. It turns people’s business requests into validated, actionable E9 workflow data, enabling AI to truly understand the form, the context, and what the user wants to accomplish—and proactively suggest the next actionable entries. BoltFusion Work, Agent Factory, and Spec together form an enterprise agent development platform, covering task execution and training, agent building and governance, multi-agent collaboration, and knowledge and context management. It focuses on connecting with Weaver E9, the E10 full API ecosystem, and enterprise-wide context, spanning organizations, people, permissions, forms, workflows, business objects, attachments, portals, and data knowledge. Agents can be sold independently for specific scenarios or built, reused, and operated on the platform. #BoltFusion #E9 #WorkflowAgents #AIFormFilling #EnterpriseAI