When you create a custom property in a report, Aline asks you to choose an output type. The output type tells the AI what format to return its answer in — free text, a number, a date, a yes/no, or a choice from a set of options. Picking the right one is what makes your results sortable, filterable, and reliable across hundreds of documents, so it's worth understanding what each does.
For how to create a property in the first place, see Building AI Reports in Aline.
What an Output Type Does
Every custom property is a question you ask the AI about each document. The output type shapes the answer: a well-chosen type gives you clean, structured data you can filter, group, sort, and even feed into reminders. A poorly chosen one (usually defaulting everything to Text) gives you answers you can read but can't easily work with at scale.
Aline offers six output types: Text, Number, Date, Boolean, Single-select, and Multi-select.
💡 Best Practice: Choose the most structured type that fits your question. Prefer Number, Date, Boolean, or a select type over Text whenever the answer isn't genuinely free-form — structured outputs are what let you sort, filter, group, and drive reminders.
Text
Returns a free-form written answer. Use Text when the answer doesn't fit any fixed format — extracting the language of a specific clause, a short summary, or a governing-law reference.
Best for: clause language, summaries, anything open-ended.
Example: a Governing Law property that returns "State of Delaware."
Note: Text is the hardest type to filter or group precisely, because every answer can be worded differently. If you need the exact contract wording, include the keyword "verbatim" in your prompt (see Refining AI Prompts for Reports, Playbooks & Chat).
Number
Returns a numeric value. Use Number for any amount or count you'll want to sort or filter by size.
Best for: contract value, liability cap amount, quantity, number of days.
Why it's powerful: because the answer is a true number (not text), you can sort high-to-low and filter by range — for example, surfacing every agreement with a contract value over $25,000, or every liability cap below a threshold.
Example: a Contract Amount property you use to prioritize your highest-value agreements.
Date
Returns a date value. Use Date for any milestone you'll want to sort or track by time.
Best for: expiration date, execution/effective date, notice deadline.
Why it's powerful: because the answer is a true date, you can sort chronologically, filter by a date range, and — most importantly — point AI reminders at the field to get notified before a contract expires or renews (see Setting Up AI Reminders for Contract Expirations & Renewals).
Boolean (Yes / No)
Returns a simple Yes or No. Use Boolean for presence/absence questions.
Best for: "Is this a BAA?", "Does this contain an auto-renewal clause?", "Is there a limitation of liability?"
Why it's powerful: Boolean is the backbone of a staged report. Screen your whole set with a Yes/No property first, filter to the Yes results, then run detailed prompts only on the relevant subset — faster, more accurate, and far lighter on tokens (see AI Reports Best Practices).
Tip: When a result is Yes, click the document icon to open the source and jump straight to the relevant clause.
Single-select
The AI assigns each document exactly one option from a set of choices you define. Use Single-select for categories that are mutually exclusive — where a document can only be one thing.
Best for: contract type, renewal type, risk level (Low / Medium / High), governing-law state.
Why it's powerful: because every answer comes from the same list and there's only one per document, your data stays consistent and clean — perfect for grouping your portfolio into non-overlapping segments and filtering reliably.
Example: a Contract Type property with options MSA / NDA / Order Form / DPA, so every document is sorted into exactly one bucket.
Multi-select
Works like Single-select, but the AI can assign one or more options from your defined list. Use Multi-select when a document can legitimately fall into several categories at once.
Best for: payment terms (Net 30 / Net 60 / Net 90), data types processed, clause types present, applicable jurisdictions.
Why it's powerful: it captures the full picture for documents that aren't one-dimensional, while still keeping answers consistent (no "Net 30" vs "net-30" vs "30 days" variants), so filtering and grouping stay clean.
Single-select vs. Multi-select: choose Single-select when only one value can apply (a contract has one type); choose Multi-select when several can (a contract can process several data types).
Building on Existing Fields
Whatever the output type, you can create more advanced properties that use already-extracted fields as inputs. For example, create an Unfavorable Payment Terms property that takes your Payment Term and Payment Frequency fields as inputs and returns a combined assessment. This lets you layer contextual analysis on top of the raw data you've already pulled.
The output type is a small choice with a big payoff: match it to how you'll actually use the answer — Number to sort by size, Date to sort and remind, Boolean to screen and filter, Single- and Multi-select to group cleanly, and Text only when the answer is genuinely open-ended — and your reports come back as structured data you can act on, not just read.
Related articles: Building AI Reports in Aline · Report Filtering, Grouping & Custom Properties · AI Reports Best Practices · Setting Up AI Reminders for Contract Expirations & Renewals · Refining AI Prompts for Reports, Playbooks & Chat