The Client Who Went Quiet - What AI Relationship Risk Actually Catches

Every consultant has a version of this story. A promising deal was moving well - a few good calls, a proposal sent, maybe even a verbal "let's do this" - and then nothing. No reply to the follow-up email. No response to the "just checking in" message three days later. You tell yourself they're busy. Weeks later you realize the deal quietly died somewhere in that silence, and you only noticed because you happened to scroll back through your contact list.

Picture a solo consultant juggling eight or nine active client relationships at once, each at a different stage - some just qualified, some deep in a proposal, one or two that should have closed already. That's a lot of context to hold in your head, and the deals that need attention most are often the ones you're least likely to think about, precisely because they've stopped generating anything to think about. A loud client emailing you daily is easy to track. A quiet one isn't a task on your list - it's an absence, and absences don't show up in a normal to-do view.

This is the actual problem SparkyContacts' AI Relationship Risk feature is built to solve, and it's worth being precise about what it does and doesn't do, because "AI flags risk" can mean almost anything depending on who's selling it.

What it actually looks at

Relationship Risk lives inside Reports, under a button labeled "Analyze risks." When you run it, it scans your open deals - specifically the high-value ones tied to contacts who've gone quiet - and returns a set of results, each with a severity rating (like "high") and a plain-language description of what looks off. Not a score out of 100, not a color-coded dashboard widget with no explanation attached - an actual sentence describing the pattern it noticed, such as a lead not having been heard from in several days despite an open, valuable deal sitting on the board.

That distinction matters more than it sounds like it should. A lot of "AI insights" features in CRM tools amount to a number with no reasoning behind it, which means you either trust it blindly or ignore it entirely. When the tool tells you why something looks risky, you can actually evaluate whether it's right. Maybe that quiet lead is quiet because they're on vacation, not because the deal is dying - you know that context, the tool doesn't. But you can only make that judgment call if the flag comes with a reason attached.

Once you've run an analysis, there's a "Regenerate" option to refresh it, which is worth using after you've followed up on the deals it flagged - rather than staring at a stale list, you get a current read on where things stand.

Why it works off stages, not guesswork

Here's the part that's easy to miss but actually matters for trusting the feature: Relationship Risk doesn't hardcode what "Proposal" or "Qualified" means. Your win rate, your open-deal totals, and the risk analysis all classify deals by how your stages are configured, not by fixed stage names. If you've renamed your pipeline stages to match how you actually think about a deal's progress, the risk analysis still respects that structure instead of silently assuming everyone uses the same five default stage names.

For our hypothetical consultant, that's not a small thing. Say their pipeline runs Lead, Discovery Call Booked, Proposal Sent, Verbal Yes, Contract Out - a sequence that maps to how their actual sales conversations go, not a generic template. Relationship Risk still knows which of those stages count as "open" and which contacts in those open stages are worth flagging, because it's reading the underlying stage type, not string-matching a label.

The plan reality worth knowing up front

Relationship Risk is available on the Max plan. If your plan doesn't include Reports at all, the page loads but shows no data rather than an error - which is a reasonable design choice, but worth knowing so you're not confused wondering why the page looks empty instead of assuming something's broken.

Using it without over-trusting it

The honest way to use a feature like this is as a second pair of eyes, not an oracle. It's genuinely good at surfacing the deal you forgot to think about - the one that isn't loud enough to interrupt your day but is exactly the one worth a check-in email. It is not going to know that your contact told you last week they'd be traveling, or that a "quiet" period is actually just the normal gap between a proposal going out and a decision-maker committee getting around to it.

Treat a flagged result the way you'd treat a colleague tapping you on the shoulder and saying "hey, have you talked to this person lately?" Sometimes the answer is "yes, everything's fine, I just haven't logged it." Sometimes it's "you're right, I completely lost track of that one." Either way, it's a useful nudge, and for a solo operator without a sales team checking in on each other, that nudge is doing work a whole team would otherwise need to do manually.

How to do this in SparkyContacts

To try it yourself: open Reports from the sidebar, and click Analyze risks near the top of the page. You'll get a list of flagged contacts, each with a severity level and a plain description of what looks off - not just a score, an actual reason. Once you've followed up on what it surfaces, hit Regenerate to get a fresh read. It's part of the Max plan, alongside the rest of Reports' pipeline breakdown, win rate, and task completion view. You can see the full feature set at SparkyContacts' features page.

If you're managing more than a handful of active relationships by memory alone, this is exactly the kind of gap it's meant to close - not by replacing your judgment, but by making sure the quiet ones don't slip past it.