> ## Documentation Index
> Fetch the complete documentation index at: https://www.landbase.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Advanced Audience Search
> How Landbase lets you define custom GTM signals in natural language instead of being limited to fixed B2B data filters.
# Advanced Audience Search
Most B2B data platforms are catalogs with a fixed set of filters: industry, title, employee count, geography, tech stack, and a few predefined intent signals. If the right filter exists, you can build the list. If it does not, the work moves into exports, spreadsheets, stitching, and manual enrichment. In the CLI, the really powerful mode for this is `advanced_dataset_creator`: it lets Landbase propose custom SQL-backed audience builds, aggregations, exact filters, and output columns when the default search agent cannot express the query.
Landbase is different because the filter does not have to exist before you ask. You can describe the signal you need, and Landbase can compute it over companies, people, job postings, uploaded account lists, and historical career or hiring data.
## High-value search examples
These are the kinds of searches that advanced audience search is built for:
* Find people who were founding sales hires at a previous startup and are now sales managers or directors at B2B SaaS companies in the US or Canada.
* Find companies where the AE-to-SDR ratio is between 1.5 and 3.0 and the Manager-to-IC ratio in Sales is less than 0.15.
* Find companies where hiring for AI or ML titles in the last 30 days is more than 2x the prior 30 days.
* For an uploaded list of 1,500 accounts, count software engineers in North America and Western Europe, then flag accounts with more than 5.
* Find CISOs at companies where Security headcount is under 2 but Engineering headcount is at least 200.
## The seven capability families
| # | Capability | What it unlocks | Example |
| - | ------------------------------------ | ------------------------------------------------------------------------------- | ----------------------------------------------------------------------------- |
| 1 | Career history match | Find people by what they have done, not just their current title. | Founding sales hires who now manage sales teams. |
| 2 | Custom company signals | Count a custom population, posting type, location, or role group per company. | Companies with at least 10 engineers but fewer than 2 security engineers. |
| 3 | Team shape | Filter by medians, percentiles, concentration, variance, or other org patterns. | Companies where the 90th percentile of engineering tenure is under 18 months. |
| 4 | Change over time | Ask about past states, acceleration, churn windows, and event order. | AI or ML hiring in the last 30 days is more than 2x the prior 30 days. |
| 5 | Team mix and ratios | Compare custom groups inside the same company. | AE-to-SDR ratio between 1.5 and 3.0. |
| 6 | People-connected companies | Find companies connected by shared ex-employees or seed-company alumni. | Companies with at least 3 executives from Google, Meta, or Stripe. |
| 7 | Uploaded accounts plus Landbase data | Treat your uploaded list as a queryable table with custom output columns. | Add engineer counts, growth, and custom flags to an uploaded account list. |
## What this changes for GTM teams
Advanced audience search moves the hard part of list building into the search itself. Instead of exporting a broad list, computing the real signal elsewhere, and reshaping the output by hand, you can ask for the exact account or contact definition you need.
That is especially useful when the buyer's question is:
* "Which accounts are showing a signal that our sales team can act on this week?"
* "Which contacts have actually done this work before?"
* "Which companies look under-invested in a function relative to their size?"
* "Which accounts from our own list match this custom profile?"
* "Which outputs should go directly into our CRM, warehouse, or outbound workflow?"
## Why this works
These searches are possible because Landbase treats the underlying data as something you can compute over, not just something you can filter with a fixed menu.
* Uploaded customer data is treated as a queryable table, not only an enrichment target.
* Career history is structured across roles, dates, descriptions, skills, and headlines.
* Job postings and career timelines are preserved historically instead of flattened to today's snapshot.
* Output columns can be defined by the customer at search time, including counts, ratios, booleans, and labels.
## Use it with the CLI
The Landbase CLI gives agents, scripts, and operators a direct way to run these searches and save structured output. Use `--agent=advanced_dataset_creator` when you need precise logic, custom SQL, aggregations, or output columns beyond the default search agent.
Advanced Dataset Creator runs in two steps. First, ask for the audience and reuse a session label:
```bash theme={null}
landbase-cli search --agent=advanced_dataset_creator --session=sales-ratio \
"Find companies where the AE-to-SDR ratio is between 1.5 and 3.0. Return company name, website, AE count, SDR count, and ratio."
```
Then confirm the proposed query in the same session:
```bash theme={null}
landbase-cli search --agent=advanced_dataset_creator --session=sales-ratio "yes, run it" --download=sales-ratio.jsonl
```
You can also ask for custom output columns in the prompt:
```bash theme={null}
landbase-cli search --agent=advanced_dataset_creator --session=security-gap \
"Find CISOs at companies where Security headcount is under 2 but Engineering headcount is at least 200. Return company name, website, engineering headcount, security headcount, CISO name, and CISO LinkedIn URL."
landbase-cli search --agent=advanced_dataset_creator --session=security-gap "yes, run it" --download=security-gap.jsonl
```
## Keep reading
* [What can you do with the CLI?](/docs/what-can-you-do-with-the-cli) shows the broader CLI command surface for search, enrichment, uploads, datasets, and workflows.
* [Use Landbase CLI in Claude Code and Codex](/docs/landbase-cli) shows how to run searches through an agent.