Structured Output Template
- What it does: Takes a company name as input, uses web search to fetch current company information, and returns it in a fixed structured format (overview, headquarters, founding year, key people, revenue, employee count).
- Best for: sales research, data ops, analysts, anyone who needs consistent company data for CRMs, enrichment, or downstream apps
- Apps used: Web Search
- Setup time: 5 minutes
Structured Output is a no-code, single-agent workflow that turns a company name into a standardized company profile. The agent uses the web search tool to gather up-to-date information and returns it in a defined schema (Overview, Headq_Location, Founding_Year, Important_People, Revenue, Num_Employees), with "Not available" for missing fields. Output is suitable for API responses and integrations. This template acts as a starting point: any app, instructions, and agents can be updated as required for your own use case—simply ask Max (Agent Builder) to update the workflow as needed. Last verified from workflow config on March 2026. Includes common failure modes and fixes.
Problem this solves
- Company data is scattered and inconsistent; you need a single, structured profile for many companies.
- Manual lookup and copy-paste into CRMs or sheets is slow and error-prone.
- You want the same fields every time (overview, HQ, revenue, etc.) for downstream apps or AI features.
- Web search returns unstructured results; you need a fixed schema for pipelines and APIs.
What this agent does
This template can:
- Accept a company name (e.g. "Apple Inc.") as input via the text trigger.
- Use the web search tool to find and retrieve current information about the company.
- Extract and normalize: overview (business, industry, core activities), headquarters location, founding year, key people (e.g. CEO, founder), revenue (latest, USD), number of employees (latest).
- Return results strictly in the defined output schema; mark unavailable fields as "Not available."
- Exclude unrelated data or extra analysis; output is only the required structured fields.
- Support downstream use: structured output is suitable for the SketricGen Public API and for feeding into other apps or AI features.
How it works
- You provide a company name (e.g. "Acme Corp") via the text trigger (Playground, API, or your deployment channel).
- AI Agent (company information agent) receives the input and calls the Web Search tool to gather current company information.
- The agent maps search results to the required fields: Overview, Headq_Location, Founding_Year, Important_People, Revenue, Num_Employees.
- It formats the response in the defined schema; any field it cannot fill is set to "Not available."
- The workflow returns the structured output; you can consume it via the API, use it in another workflow, or display it in your app.
Requirements
- Web Search enabled (the template uses a web_search tool).
- Input: a single company name (string). No API keys required for the default setup.
- Optional: if you integrate with the Public API, you will need API access and to handle the structured response format (see docs: Public API usage).
Setup guide
- Clone the template and open it in AgentSpace.
- Ensure the Web Search tool is enabled (the template ships with it on). No connection is needed for the default web search tool.
- Run Test Workflow with a company name (e.g. "Stripe" or "Apple Inc."). Confirm the response has all schema fields and that "Not available" appears for missing data.
- (Optional) Expose the workflow via the SketricGen Public API and consume the structured output in your app or pipeline.
- Deploy or run on demand when you need company data.
Common issues and fixes
- Missing or wrong fields: web search results vary; the agent fills what it can and uses "Not available" for the rest. For rare or private companies, expect more "Not available" fields.
- Output not valid for my app: the schema is fixed in the agent (AgentOutput with the listed fields). You can edit the agent’s output schema in AgentSpace to add/rename fields or change types, then re-test.
- Slow or rate limits: web search can be slower for many calls; for bulk company lookup, consider batching or caching outside the workflow.
Customization knobs
- Change the output schema in the AI Agent node (add or remove fields, rename for your CRM or API). Keep the agent instructions in sync so it still fills the new schema.
- Add more tools (e.g. a dedicated company data API) and update the agent to use them; keep the structured output format for consistency.
- Use the same pattern for other entity types (e.g. person, product) by changing the agent instructions and schema.
Apps used
| App | What it is used for | Typical permission scope |
|---|---|---|
| Web Search | find and retrieve current company information from the web | internet search |
Use cases
- Lead or account enrichment: Get a standard company profile (overview, HQ, revenue, people) for a list of company names. Best for sales and SDRs.
- CRM or sheet population: One company name in, one structured row out for import into CRM or spreadsheets. Best for data ops and rev ops.
- API-backed company lookup: Expose the workflow via the Public API and return structured company data to your app or another service. Best for developers and product teams.
- Research and due diligence: Quick, consistent company snapshots for research or screening. Best for analysts and investors.
- Pipeline input: Feed structured company data into another workflow (e.g. Personalised Lead Outreach) or into an AI feature that needs fixed fields. Best for multi-step automations.
Example prompts and outputs
Example 1: Public company
Prompt (input): Apple Inc.
Output: Structured object with Overview (business, industry), Headq_Location (e.g. Cupertino, CA), Founding_Year, Important_People (e.g. CEO), Revenue (latest USD), Num_Employees. Unavailable fields: "Not available."
Example 2: Private or small company
Prompt (input): Acme Design Studio
Output: Same schema. Some fields may be "Not available" if web search has limited public data.
Example 3: API usage
Trigger the workflow via the Public API with a company name in the request body; response body is the structured output (e.g. JSON with the schema fields). See Public API usage for integration details.
Why you need this template
- Same schema every time: predictable fields for pipelines, APIs, and CRMs.
- No manual formatting: the agent normalizes web search results into your structure.
- Easy to extend: change the schema or add tools (e.g. company APIs) in AgentSpace.
- Integrates with the SketricGen API and other workflows for multi-step use.
FAQs
1) What input does the workflow accept?
A single company name (string), e.g. "Stripe" or "Apple Inc." You can pass it via the Playground, your deployment channel, or the Public API.
2) Can I change the output fields?
Yes. Edit the AI Agent node in AgentSpace: update the output schema (add, remove, or rename fields) and adjust the agent instructions so it still fills those fields from web search (or other tools you add).
3) Why do some fields say "Not available"?
The agent only fills fields it can source from web search. For private companies, very new companies, or limited public data, some fields will be "Not available."
4) Does it use an external company database or API?
The default template uses only Web Search. You can add another tool (e.g. a company data API) in AgentSpace and update the agent to use it while keeping the structured output format.
5) How do I use this with the Public API?
Trigger the workflow via the SketricGen Public API with the company name as input; the response will be the structured output. See Public API usage.
6) Can I use it for people or products instead of companies?
Yes. Duplicate the template and change the agent instructions and output schema to fetch and structure data for a person or product; add or swap tools as needed.
7) What are the main failure modes?
Web search returns no or poor results (try a more precise company name), schema mismatch (ensure your app expects the same field names and types), or rate/usage limits on web search for high volume. Fix by refining input, aligning schema, and batching or caching if needed.
8) Is the output always valid JSON / typed?
The workflow returns the agent’s structured output as defined in the schema. When consumed via the API, it is typically JSON. Ensure your consumer handles optional fields and "Not available" strings.
Related templates
- Personalised Lead Outreach – enrich leads (e.g. with Apollo, PredictLeads) and send personalized emails; can consume company or lead data from other workflows.
- Github Code Review – structured review output (reasoning, positives, changes, etc.) for code.
- Internal Team Support – internal Q&A from knowledge bases.
- Marketing Strategy – research and strategy with structured outputs possible.
- Personal Ops AI Assistant – calendar, comms, and retrieval in one workflow.
Structured Output
Fetch structured, up-to-date company info using AI and web search.
