--- title: "Projects and Workflows: Two Ways to Give an Agent Structure" slug: "projects-and-workflows" date: "2026-08-24" tag: "Application" summary: "Projects give a body of work its own files, standing instructions, and assigned agents, so every conversation started from it carries that context. Workflows lay a multi-step agent run out as a graph — branches, loops, guardrail checks, and human approval steps." author: "ibl.ai Engineering" linkedin: | There are two ways to make an AI agent reliable on real work, and they solve different halves of the problem. The first is context. An agent that has to be re-briefed at the start of every conversation isn't working on your project — it's guessing at it, again. Projects on ibl.ai give a body of work its own files, its own standing instructions, and its own assigned agents. Open the project, and every conversation you start already carries all three. The instructions are scoped to that project, so setting house rules for one piece of work doesn't change how the same agent behaves anywhere else. The second is control flow. A single prompt asking one agent to do everything is a bet that it gets the whole sequence right in one turn. Workflows let you draw the run instead: an agent step here, a guardrails check there, a branch, a loop, and a point where a person has to approve before anything continues. The node palette is the interesting part. Guardrails is a node — PII detection, moderation, jailbreak detection, hallucination checking — which means you decide exactly where in the run the screening happens rather than flipping a global switch. User approval is a node too. And every agent step carries a "continue on error" switch, so you choose per step whether a failure stops the run or the work carries on past it. Publishing validates the graph before promoting it. Both are surfaces you own outright. On ibl.ai you own all the code and the data — your project files, your standing instructions, and the workflow graphs themselves are yours, running on your infrastructure, not process locked inside a product you rent. #iblai #AgenticAI #EnterpriseAI #AIAgents #Workflows --- Two surfaces shipped for giving agents structure, at different levels. **Projects** organize the context a body of work needs. **Workflows** organize the steps a task runs through. ## Projects ![The project landing page, showing the project title, the Ask anything chat input with Canvas and Prompts controls, the Project files and Add project instructions cards, and the Project Agents grid](/images/updates/projects-landing.webp) A project is a named workspace holding three things: the **files** the agents in it can draw on, the **instructions** that tailor how they respond inside it, and the **agents** assigned to the work. Opening a project replaces the usual chat welcome screen with a landing page for that work, and the chat input at the top of it behaves exactly as it does elsewhere — the same composer, the same Canvas and Prompts controls — except that what you send is scoped to the project. Project instructions are the part that changes the day-to-day: they layer on top of an agent's own prompts and apply only inside the project, so a project is where you set the conventions, the audience, and the output format for one piece of work without touching how that agent answers anyone else. ![The Project Files dialog, showing a dataset search, an Add Files button, and a table of files with Name, Type, Tokens, Interval, Visibility, and Status columns](/images/updates/projects-files.webp) **Project files** manages the project's own document set, each with its refresh interval, visibility, and an active switch. Several agents can share one project when the work spans more than one specialism. ## Workflows ![The workflow editor, showing the node-type sidebar grouped into Core, Tools, Logic, and Data, a canvas with Start, Agent, Guardrails, and End nodes connected by curved edges, a Guardrails configuration panel with safety-check checkboxes, and Connectors, Save, and Publish actions](/images/updates/workflows-editor.webp) A workflow is an agent run you draw rather than describe. The sidebar is the palette of node types, in four groups: - **Core** — an agent step, an **End** node, and a **Note** for annotating the graph. - **Tools** — **File search**, **Guardrails**, and **MCP**. - **Logic** — **If / else**, **While**, and **User approval**. - **Data** — **Transform** and **Set state**. The canvas gives you drag-and-drop, pan and zoom, undo and redo, and auto-save, with the header showing the workflow's status and whether there are unsaved changes. Each node type opens its own configuration. An agent step takes instructions, a model, and a **Continue on error** switch that decides — per step — whether a failure stops the run or the work carries on. A **Guardrails** node offers a checklist of **PII Detection**, **Moderation**, **Jailbreak Detection**, and **Hallucination Check**, which means the screening happens at exactly the point in the run you place it. The **Start** node declares the state variables the run carries, and **End** defines what it returns. **Publish** validates the graph before promoting it to the running version, so a structurally broken workflow is refused rather than shipped. ## Also in this release **Explore Agents** is documented as its own surface — the searchable, filterable agent browser with starred, featured, custom, and all-agents sections, and category, subject, type, and LLM-provider filters. ![The Voice instructions card, showing the style prompt with Edit and Copy actions, a character counter, and one-click example presets](/images/updates/voice-instructions.webp) **Voice instructions** let an agent be told *how* to speak, as distinct from which voice says it — a free-form style prompt with a rich-text editor, a 1,000-character cap, and one-click presets. OpenAI reads it as speech instructions and Google as a synthesis prompt; both treat it as guidance on delivery, not content. ![The Sandbox Prompts section, listing the agent workspace files — Identity, Soul, User Context, Tools, Agents, Bootstrap, Heartbeat, and Memory — each with an information tooltip and an Edit button](/images/updates/sandbox-workspace-prompts.webp) **Sandbox workspace prompts** — the eight files that define a sandboxed agent on its runtime (Identity, Soul, User Context, Tools, Agents, Bootstrap, Heartbeat, and Memory) — are each editable from the Sandbox panel, with **Auto Push on Save** keeping the runtime in step. **LTI launch links** now build asynchronously. Creating one provisions a course behind the scenes, so the link appears immediately with a status of its own — pending, building, ready, or failed — with a refresh while it builds and a retry on the rare failure, rather than holding a browser request open for the duration. ## Documentation - [Projects](/docs/os/chat-canvas/projects) and [Workflows](/docs/os/chat-canvas/workflows) - [Explore Agents](/docs/os/chat-canvas/explore) - [Agent Settings: Voice](/docs/os/agent-settings/voice), [Sandbox](/docs/os/agent-settings/sandbox), and [LTI Links](/docs/os/agent-settings/lti-links)