How Predictive Insights Improve Lead Follow-Ups

What are predictive insights in a real estate CRM?

Predictive insights are calculated signals derived from a lead's behavior history — how often they engage, how fast they respond, which channels they prefer, and how recently they were active. Unlike a simple timestamp ("last contacted 3 days ago"), a predictive signal synthesizes multiple data points into an actionable recommendation.

In a real estate context, the most useful predictive signals fall into four categories:

• Engagement scoring — a composite score that rises when a lead opens emails, returns calls, or replies to texts, and decays when they go silent.
• Velocity — how quickly engagement is accelerating or decelerating relative to recent history.
• Recency of inbound activity — days since the lead last initiated contact, which often outpredicts agent-initiated contact in determining conversion intent.
• Preferred channel — the medium (call, text, email) through which the lead has historically been most responsive.

Why traditional follow-up sequences miss high-intent leads

Most CRM action plans are time-based: send a text on day 1, call on day 3, email on day 7. This is better than nothing, but it treats a lead who just visited three listing pages and texted back within minutes the same as a lead who submitted a form six months ago and has been unresponsive since.

The result is that agents spend significant time on low-intent contacts while genuinely hot leads — who behaved differently but fell into the same sequence — receive the same cadence as everyone else. Predictive scoring breaks that tie by surfacing the leads whose signals indicate purchase intent right now.

How Follow Up Ace tracks predictive signals inside Follow Up Boss

Follow Up Ace maintains two free predictive custom fields on every contact in Follow Up Boss — updated in real time as webhook events come in and re-synced nightly from its calibrated model suite — plus four more with the Ace Trove add-on. These fields give every agent on the team a live predictive view without manual work:

Ace Win Score and Ace Churn Risk are available on every plan including free, and all six write directly into Follow Up Boss custom fields so agents can build Smart Lists and action plans using them. In June 2026 this set replaced the earlier heuristic fields (Ace Score, Tier, Status, Velocity, Preferred Channel, and friends) — those signals are still computed on every webhook, but they now feed the models and the ranked queue instead of occupying field slots. See the Ace Trove page for details on the full field set, or the field-by-field guide.

How does Ace Win Score actually work?

Ace Win Score is a 0–100 number that represents a calibrated probability: a contact showing 30 has roughly a 30% chance of converting to a closed deal, as measured against actual outcomes. It comes from a machine-learning model trained on engagement history — recency, inbound-vs-outbound direction, response latency, channel behavior, deal context — and its raw output is passed through a calibration layer that is refit against real labeled outcomes, so the number keeps meaning what it says as your market shifts.

That calibration loop is the difference between a vanity score and a probability you can plan on: the system records what it predicted for each contact, checks what actually happened at 24 hours and 7 days, and adjusts. (Read the full loop in our predictive scoring deep-dive.)

What is Ace Churn Risk and why does it matter?

Ace Churn Risk answers a different question than Win Score: not "will they convert?" but "is this engaged contact about to go cold?" A separate churn model watches for decaying engagement — slowing reply cadence, lengthening gaps since the last inbound touch — and classifies each engaged contact Low, Medium, or High.

Churn Risk is the signal that separates agents who catch the cooling moment from those who call a day too late. A Smart List filtered to Ace Churn Risk = High is your save-list: the pipeline you already earned, about to slip away.

How does channel and velocity tracking change follow-up outcomes?

Behind the fields, Ace still tracks which communication medium historically gets the fastest response from each lead, how fast their engagement is accelerating or decelerating, and how many days since they last reached out. Those signals feed the models, the ranked daily queue, and the suggested opener each agent sees — so a lead who answers texts in minutes but ignores calls gets a text-first play without anyone checking a field manually.

Matching your outreach channel to a lead's demonstrated preference reduces friction and increases response rate. A call that goes to voicemail is a missed moment; a text to a lead who already engages via text is a much higher-probability touch.

How to build a predictive follow-up workflow in Follow Up Boss

1. Create a Smart List sorted by Ace Win Score descending. This becomes your daily priority dial list. Ace fields update as engagement changes, so the list refreshes automatically.
2. Add a second Smart List filtered by Ace Churn Risk = High. These are engaged contacts the model expects to go cold — your save-list, worth a touch before the window closes.
3. With Ace Trove, sort by Ace Expected GCI descending. Same leads, ranked by estimated commission at stake instead of raw probability — the team lead's view of where the money is.
4. Use Ace Opportunity to route the play. "At Risk" gets a save sequence, "Likely Seller" gets a home-value conversation, "Active Buyer" gets inventory. One field, one recommended play.
5. Let the ranked queue handle the rest. Speed to lead matters enormously in real estate; Ace's daily queue already folds response urgency, channel preference, and days-since-inbound into its ordering, so the highest-priority contacts surface without extra filters.

Can predictive insights replace human judgment in follow-up?

No — and they're not designed to. Predictive signals are information inputs, not decisions. A lead's Ace Win Score tells you who deserves your first call of the day; it doesn't tell you what to say when you make it. The human judgment involved in reading a conversation, understanding a life situation, and building a relationship is not something a score replaces.

The value is time allocation. Most agents have more leads than hours. Predictive scoring makes the prioritization problem tractable without requiring agents to manually review every contact's timeline every morning.

What about leads that score low but are actually serious buyers?

Low-engagement leads who are serious buyers do exist — the person who reads every email but never replies, or the client who is simply not a communicator. Ace Win Score is trained largely on behavioral engagement with your outreach, which is a strong but imperfect proxy for underlying purchase intent.

The best practice is to use predictive signals as a prioritization layer, not an exclusion filter. Low-Win-Score leads should still receive nurture touches — the difference is that high-Win-Score leads receive your personal attention first. You can also prompt Follow Up Ace's chat to surface recently created or recently active contacts that may have been overlooked in a score-based sort.

For a broader look at how AI continuously improves your contact database, see the agentic tools that work alongside these predictive fields, or explore how AI instantly scores new real estate leads as they enter Follow Up Boss.