The Scenario
You're a data engineer on a recruiting analytics team. Your company scrapes job postings from 30 different job board pages — indeed clones, niche boards, aggregators — and loads them into a pipeline. The problem: every board returns data in a slightly different format. One calls the salary field "compensation", another "pay range", another buries it in a description paragraph.
The pipeline breaks every time a new board is added. The fix is always the same: someone maps the new board's schema manually. That someone is always you.
The job board URLs are already in column A of an Excel workbook. The extraction schema you want is always the same five fields: job title, salary range, location, company name, remote flag.
The bad version:
- Open job board 1, inspect the page structure, figure out what selectors or field names it uses for each of your five target fields
- Repeat for job board 2, discover the salary field is missing entirely, decide whether to write "not listed" or leave it blank
- Realize by board 8 that you're manually writing extraction logic that ScrapeGraph AI could generate for you — and you're still doing it the hard way
The engineering sprint ends Friday and this is blocking two downstream tasks.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent inside your Excel workbook. It uses its built-in ScrapeGraph AI integration to generate a consistent extraction schema from a natural language description, then applies that schema uniformly across every URL in your column.
Paste this into the SheetXAI sidebar:
Generate a ScrapeGraph AI JSON schema for extracting job title, salary range, location, company name, and remote flag from job listing pages, then apply it to every URL in column A and write results into columns B through F
What You Get
- A consistent schema is generated once and applied to all 30 URLs
- Columns B through F fill with job titles, salary ranges, locations, company names, and remote flags
- Rows where a field wasn't found get a standard "not listed" placeholder
- The schema definition is written to a reference cell for reuse
What If the Data Is Not Quite Ready
Some job boards return salary as a range and others as a single number
Normalize all salary values in column C: convert ranges like "50,000-70,000" to a midpoint numeric value; convert single numbers to the same format; flag rows where the salary field was entirely absent in column G
The remote flag uses inconsistent labels across boards
Normalize the remote flag values in column F: convert all variants of "yes", "fully remote", "remote ok" to "remote"; convert "on-site", "in office", "no" to "on-site"; write "hybrid" for values containing "hybrid" or "flexible"
You need to add a new field (posted date) to the existing schema
Update the extraction schema to also capture the job posting date; re-run ScrapeGraph AI against the 30 URLs in column A and write the posting date into a new column G
Full pipeline: generate schema, apply uniformly, normalize, flag gaps
Generate a ScrapeGraph AI JSON extraction schema for job title, salary range, location, company, remote flag, and posting date; apply it to all 30 URLs in column A; normalize salary to a midpoint numeric; normalize remote flag to remote, on-site, or hybrid; flag any row with more than 2 missing fields in column H; write a schema reference to cell J1
One prompt that resolves the field-mapping problem that has been breaking your pipeline every time a new board is added.
Try It
If you have a batch of structured-data URLs and need a uniform extraction schema applied across all of them, Get the 7-day free trial of SheetXAI and let ScrapeGraph AI generate the schema and run it in one pass. For related tasks, see bulk scrape competitor pricing or crawl supplier category pages.