Next-Gen Custom Fields: Let Ace AI Write, Update & Report on Data You Never Thought Possible

What are AI-written custom fields in Follow Up Boss?

Follow Up Boss has always let teams add custom fields to contact records — a great feature for capturing notes, loan status, or referral source. The limitation has always been the same: someone on your team has to fill those fields in. If nobody does, they sit empty.

AI-written custom fields work differently. Instead of relying on manual entry, a scoring layer evaluates supported contact evidence and writes structured values to a deliberately small field catalog. Each value carries a scoring cadence and freshness boundary; a blank or stale state is surfaced instead of being described as current.

Follow Up Ace installs a unified set of these fields into your Follow Up Boss account. The system is split into two tiers based on how they are produced and what plan you are on.

Which custom fields does Ace write, and what do they contain?

Ace maintains seven model-grounded custom fields in two groups: two fields for every connected account and five additional fields for Ace Trove accounts. Values appear only when the scoring pipeline has enough supported evidence; blank is better than an invented answer.

Free fields — available on every connected account

These two predictive fields are available at no additional cost and are refreshed on the account's scoring cadence:

Ace Trove fields — account-wide predictive intelligence

When you enable Ace Trove, the scoring pipeline adds five sortable fields backed by account-wide and contact-level evidence:

• customAceExpectedGCI — calibrated win probability multiplied by the account's own median commission, clearly presented as an estimate
• customAceOpportunity — At Risk / Active Buyer / Active Seller / Likely Seller / Nurture
• customAceSellerTier — Very High / High / Moderate / Low / None relative seller-propensity band
• customAceBuyerTier — Very High / High / Moderate / Low / None relative buyer-propensity band
• customAceLastAnalyzed — date of the most recent completed analysis

Modeled values are estimates and relative rankings, not guaranteed outcomes. Prose summaries, property context, and recommended next steps stay live in Companion and the Ace dashboards instead of becoming stale FUB fields.

How does Ace AI actually write data into Follow Up Boss custom fields?

After a scoring run completes, supported values are queued through Ace's guarded Follow Up Boss write path. The resulting fields are visible on contact records and can be used for native sorting and filtering; unsupported or stale model outputs are left blank rather than guessed.

Model-score delivery (free tier)

Ace uses Follow Up Boss activity as model evidence. Scores are delivered on the account's configured cadence, and meaningful activity can schedule a debounced refresh. The displayed freshness state tells you when the last completed analysis ran.

Ace Trove analysis (paid tier)

Ace Trove performs account-wide analysis and writes only the five canonical Trove fields when their source evidence is sufficient. The customAceLastAnalyzed date tells you when each contact's completed analysis was produced.

The Seller Tier field is additionally refreshed nightly for Ace Trove accounts via a Seller Radar scan that incorporates property tenure, estimated equity, local appreciation data from the FHFA, area market pressure, and CRM seller signals.

What can you actually build with AI-written custom fields in FUB?

The value of any custom field depends on what you do with it. Because Ace fields are native FUB custom fields, they plug directly into everything Follow Up Boss already supports: filtering, sorting, Smart Lists, and automations. Below are the built-in recipe templates Ace ships with the Ace Trove.

Recipe: Highest Win Score review

Sort by customAceWinScore descending and save the view as a Smart List such as "Highest Win Score." This treats the score correctly as a within-account rank and keeps the highest-ranked supported contacts at the top.

Recipe: Stale high-value re-engagement

Combine a high customAceWinScore rank with customAceChurnRisk = High. This creates a review queue of relatively promising contacts whose relationship needs attention. Keep any outreach agent-reviewed and use the documented POST /actionPlansPeople endpoint only when the workflow has been intentionally configured.

Recipe: Ghost / at-risk owner alert

Filter on customAceChurnRisk = High and sort by Win Score. Team leads can use the resulting evidence-backed review queue for coaching or a save attempt; the model flag is a prompt to investigate, not an instruction to reassign or archive someone automatically.

Recipe: Opportunity review (Ace Trove)

Sort numeric Ace Win Score descending or filter Ace Churn Risk = High, then review current FUB stage, activity, response-time reporting, and communication history before acting. Those native signals are not additional Ace custom fields.

Recipe: Ready buyers fast-track (Ace Trove)

With Ace Trove enabled, filter on customAceBuyerTier = Very High or High. These are relative model-ranked propensity bands, not guarantees that a contact is ready to transact. Review the underlying contact evidence before routing or outreach.

How does Ace use custom fields to surface seller opportunities?

The customAceSellerTier field — and the Seller Radar score behind it — represents a different category of AI-written data: propensity modeling applied to your existing database rather than lead scoring based on inbound behavior.

The Seller Score (0–100) is computed from a weighted set of signals: property tenure, estimated equity modeled from a conservative mortgage calculation (not observed lien data), FHFA appreciation data for the area, local market pressure indicators, and CRM-level seller tags you or your team have applied. Nightly, the Seller Radar scan refreshes these scores across your account's contacts.

What this means practically: you can build a FUB Smart List filtered on customAceSellerTier = Very High or High and have a standing view of the homeowners in your database most likely to consider selling — without any manual outreach tracking or spreadsheet maintenance.

A few important caveats. Equity figures are modeled estimates, not appraisal data. The Seller Radar surfaces suggested outreach in the dashboard for agent review — suggestions are not auto-sent. The banded Seller Tier writes to a FUB custom field nightly; the underlying 0–100 Seller Score and its evidence stay in the Seller Radar view rather than syncing to a field.

Learn more about how Ace approaches property intelligence in the Agentic overview.

Can Claude or ChatGPT read and use these custom fields?

Yes. Follow Up Ace exposes your Follow Up Boss data — including all Ace custom fields — to Claude and ChatGPT via an MCP (Model Context Protocol) connector. Claude connects via the Streamable HTTP endpoint at https://followupace.com/mcp; ChatGPT connects via the streamable-HTTP endpoint at https://followupace.com/mcp?tools=full. Both surface the same 189 tools covering contacts, activities, tasks, and pipeline data.

With that connector active, an agent can ask Claude directly: "Summarize the last three interactions with Sarah Chen and suggest a follow-up message given her Ace Buyer Tier and preferred channel." The AI reads the custom field values you already have in FUB and incorporates them into the answer without you needing to copy-paste anything. See the Agentic page for a full walkthrough of the connector setup.

How to set up AI-written custom fields in Follow Up Boss — step by step

1. Connect Follow Up Boss. Sign into your Follow Up Ace account and complete the Follow Up Boss OAuth connection. Ace needs read and write access to populate fields on contact records.
2. Let Ace install the free fields. On first connection, Ace creates the free predictive fields — Ace Win Score and Ace Churn Risk — in your FUB account automatically. No manual field setup is required.
3. Verify the fields appear in FUB. Go to Admin → Custom Fields in Follow Up Boss and confirm you see the customAce* fields. If you have existing contacts, Ace begins scoring them as events arrive.
4. Build your first Smart List. In FUB, go to People → Filter. Sort by Ace Win Score descending (or filter Ace Churn Risk = High for the save-list). Save the view. This is your live priority queue.
5. Enable the Ace Trove for deeper fields. From the Ace admin dashboard, activate the Ace Trove. Pricing scales from $49/month for up to 5,000 contacts to $899/month for up to 500,000 contacts. The paid fields become available immediately after activation.
6. Let the account-wide Trove analysis run. Ace writes supported values to Expected GCI (est.), Opportunity, Seller Tier, Buyer Tier, and Last Analyzed.
7. Build review views on verified fields. Sort Win Score, filter Churn Risk, or use Trove's Opportunity and buyer/seller tiers; confirm the contact evidence before routing or action-plan changes.

What is the difference between free and paid Ace custom fields?

Do AI-written custom fields work with Fair Housing compliance?

This is a legitimate concern. Any AI system that classifies contacts and influences which ones receive outreach needs to be evaluated against Fair Housing standards. Ace's approach addresses this in two ways.

First, the published opportunity and propensity fields are derived from product-supported CRM, activity, and property evidence. They are not designed to use protected-class characteristics.

Second, Ace includes a Fair Housing compliance layer that scans outbound messages for language that could violate Fair Housing rules. The scanForComplianceViolations() function in the platform checks generated text before it is sent, flagging steering language, prohibited terms, and other compliance risks. You can read more about how this works on the Compliance page.

That said, how you act on AI-populated fields is ultimately your responsibility. A Buyer Tier is a relative model rank and must not be used as a proxy for a protected characteristic. Treat AI-written fields with the same Fair Housing review you apply to manual segmentation.

How do AI custom fields compare to building your own FUB custom fields manually?

You can absolutely build custom fields in Follow Up Boss manually — and many teams do. The difference comes down to who fills them in and how consistently that happens in practice.

Manual custom fields depend on agents entering data. That works well for fields that agents genuinely know and care about — closing date, referral source, buyer pre-approval status. It works less well for fields that require synthesizing activity data, like "how engaged is this lead" or "what properties are they looking at," because agents rarely have time to update those consistently across hundreds of contacts.

AI-written fields fill a narrower gap: the seven canonical model outputs are delivered on the applicable scoring cadence when evidence supports a value. Behavioral prose, channel history, and next-step suggestions remain live in Companion or FUB rather than becoming stale custom fields. Manual and modeled fields can coexist when each has a clear owner.

For a broader view of what Ace adds to a Follow Up Boss workflow, the Guides section covers specific use cases including Zillow lead handling and pipeline velocity analysis.

Key takeaways

• Free catalog: Ace Win Score and Ace Churn Risk are available on every connected account.
• Ace Trove adds Ace Expected GCI (est.), Ace Opportunity, Ace Seller Tier, Ace Buyer Tier, and Ace Last Analyzed.
• Published values are native FUB custom fields. Live prose, evidence, and suggested next steps stay in Companion and the dashboards so they do not become stale CRM fields.
• The Ace Trove is priced at $49–$899/month based on the number of contacts in your Follow Up Boss account.
• Claude and ChatGPT can read these fields via the MCP connector (Streamable HTTP at https://followupace.com/mcp; SSE at https://followupace.com/mcp?tools=full).