Skip to content

Platform / AI Fields

Useful CRM fields. Less manual sorting.

Find the CRM detail you need. Review each suggested value and its source before applying a change.

Find the useful details
Review extracted details before accepting them. Example data.

HOW IT WORKS

Start small. See what will change.

  1. 01

    Choose the field

    Use a supported text, long-text, select or multi-select field on Leads, Companies or People. Define the information you want.

  2. 02

    Review a quoted sample

    Select up to ten records. Review the credit quote before running the AI sample.

  3. 03

    Inspect values and evidence

    Check each proposal against its available source. Existing manual values remain protected.

  4. 04

    Apply the approved suggestions

    Confirm the reviewed changes. If the underlying record changed, refresh the review before applying it.

Useful answers. In your CRM.

A useful AI field answers a question that someone would otherwise investigate record by record.

Read the details

For a company, that might be an adoption concern mentioned in a conversation. For a person, it might be a stated contact preference. Keep the question specific enough that a teammate can judge whether the proposed answer is supported.

Start in CRM → Fields with a saved custom field, then open AI suggestions. Choose the instructions, permitted sources and time window. Text and long-text fields work for concise explanations. Select and multi-select fields work when your team needs one of the options already defined for that field.

The record type matters. A company-level field should describe the company; a person-level field should describe that person. A detail about one deal should not quietly become a general fact about the whole business. Clear instructions and a small reviewed sample help you catch that mismatch before applying values.

Ask an answerable question
Prefer “What concern did this company state?” to an instruction that asks the model to infer everything about the account.
Choose the right shape
Use an explanation when context matters and an allowed category when consistent grouping is the goal.
Test different records
Include records with useful context, little context and an existing value so you can inspect the different outcomes.
Review each suggestion
Review extracted details before accepting them. Example data.

Review the details before handoff.

Imagine a team wants a short Company field called Adoption concern.

Read the details

It defines the question, chooses the permitted CRM and meeting sources, and selects a few companies. Spacebrain shows the cost quote before the sample runs through Orbit. Saving the field configuration alone does not run the sample or change the CRM.

For each result, compare the current value, suggested value and source excerpt. One company might have a clear concern about migrating its records. Another might have too little evidence. Leave unsupported results unapplied; a missing answer is not a reason to clear an existing field.

Values entered by people or other tools receive protection. Replacing an existing value requires explicit overwrite review, and a changed record can require a fresh sample. Where private context is involved, review the visibility choice before publishing a derived value into a shared CRM field. These controls make the proposed change inspectable at the point of use.

Run a small sample
Up to ten records lets you assess the instruction and quoted usage before expanding your use.
Keep the approved result
Apply selected suggestions after checking the evidence. The result is a value on the native CRM record.
Refresh deliberately
Optional refresh uses a selected scope and daily credit cap. It creates new suggestions; applying them remains a separate review.

Keep the field useful as the work changes.

Use a clear question and a field your team can act on. The controls keep review close to the data.

Describe a practical detail

Summarize a stated need, extract a discussed next step or suggest an allowed category. Results depend on the context available for the selected records.

Explore the workflow

Protect your team’s entries

AI proposals do not silently replace values your team entered. Review both the current field and the proposed value before applying.

Refresh suggestions within a limit

Optionally enable bounded refresh for a small selected set and a daily credit cap. Refresh creates suggestions for review; applying them remains a separate action.

Explore the workflow

Before you start

Which records and field types are supported?

AI field suggestions support Leads, Companies and People, using text, long text, select and multi-select custom fields. This does not make every CRM field type AI-editable.

Will AI overwrite my manual values?

Manual values are protected from silent replacement. Replacing a protected value requires explicit overwrite review. The workflow checks the current record again before applying the approved change.

How much does a sample cost?

Spacebrain shows a credit quote before you confirm a sample of up to ten records. Optional refresh also uses a configured daily credit cap. Check the current quote and your workspace usage rather than assuming unlimited AI is included.

Does refresh keep applying changes automatically?

No. Optional refresh produces new suggestions within the selected scope and cap. You still review and apply values separately.

Can I use an AI field to enrich missing external data?

The field workflow proposes values from available supported context. It does not promise a complete external enrichment database or verified facts when the source is missing.