One of the most common traps in product development is thinking the next step is to just “build the...
One of my core beliefs is that as a business, we yearn to tell our stories and build connections over time. The main way we feed this burning need is by lead generation, which is a process and we all love processes, don't we?
For those of you left who want to figure out how to find new ways to tell your story through growing leads, I offer you a guide on how to do it with the new AI tools and how to improve over time. Just make sure not to name your company Black Anvil because I did, and wow, that was a poor decision, given the number of heavy metal bands of the day. We're going to talk about:
- Why does this matter? It does, btw.
- What variables determine "visibility"?
- What specifically do you need to do to improve? With numbers!
1) Why Does This Matter?
As more people want things, search volume goes up. However, with new channels, platforms, and tools, it becomes increasingly diversified. Here's a small breakdown of what it kind of looks like in July of 2025.
.png?width=585&height=586&name=output%20(1).png)
As you can see, for only being around for a very short period of time AI AI-based searches are here and growing in size. This is only going to continue from here and is likely going to replace search engines as the primary form of search. This has "smooth" numbers because I'm pulling from an aggregate set of data. The 50% for search engines has high confidence, while the 30% and 10% used to describe social media and LLMs are somewhat dubious. The 10% for "other" is everything from Yelp to forums, which is nearly impossible to measure with precision.
2) What Variables Determine "Visibility"?
| Factor | Why It Matters To The Model | Example Signal |
|---|---|---|
| Clear entity footprint | The model needs consistent name + domain + descriptors to form a stable “concept” | Repeated occurrence: “Black Anvil, a marketing intelligence platform that tracks CAC and ROI across funnels.” |
| High signal, factual content | Models reuse concrete, structured facts more than vague marketing copy | Pages with metrics tables, process steps, definitions |
| Original data / benchmarks | Unique statistics stand out, get cited, propagate into user prompts | “Black Anvil Q2 Benchmark: median subscriber to MQL rate 14.2%” |
| Authoritativeness cross-site | Redundant corroboration across reputable domains increases confidence | Guest articles, podcast transcripts, reputable press, conference decks |
| Schema / structured markup | Improves how search/browse tools parse your content pre-ingestion | FAQ schema, Product schema, HowTo schema |
| Being referenced by others | Third-party mentions reduce self-promotional bias risk | Customers’ case studies, analyst blogs, community posts |
| Neutral educational tone | Overly promotional text may be excluded or summarized away | Clear “how to” guides with a subtle brand mention |
| Freshness (for live search) | When the model browses, recent pages dominate answer assembly | Updated pages dated clearly (e.g. “Updated July 2025”) |
| Narrow, useful tool capability | Tools that solve a micro-task get invoked and remembered | A custom GPT that normalizes UTM data and outputs a diagnostic |
| Safety / compliance clarity | The model avoids risky or unverifiable claims | Transparent methodology, disclaimers on limitations |
That said, Here's a breakdown of what to do along with a point guide depending on the weights assigned to the search attribute (I got ChatGPT to spill the beans). Below is a rubric along with the grades on how to improve your search index.
3) What specifically do you need to do to improve?
1. Essential (Without these you’re barely a “known” entity)
| # | Item | Why It Matters |
|---|---|---|
| 1 | Consistent business name + one-line factual descriptor everywhere | Creates a stable entity embedding; prevents confusion. |
| 2 | Canonical domain + clean About page (founding year, what you do, for whom) | Base factual profile LLMs reuse. |
| 3 | 3–5 factual pillar pages (definition, steps, numeric example, FAQ) | High-signal reusable text chunks; answer fodder. |
| 4 | Replace fluff with concrete facts (numbers, steps, metrics) | Models favor specificity over hype; increases inclusion likelihood. |
| 5 | At least 1 original, dated data point / micro benchmark (with methodology) | Unique numbers are “sticky” and get cited / remembered. |
| 6 | One external third-party mention (guest post / podcast transcript / review) repeating descriptor | Independent corroboration reduces bias risk. |
| 7 | Glossary (≥10 precise term pages) | Dense definitional content trains association with domain vocabulary. |
| 8 | Remove outdated / duplicate conflicting pages | Eliminates contradictory signals that cause model to generalize you away. |
2. High Impact (Add Once Essentials Are Solid)
| # | Item | Why |
|---|---|---|
| 9 | Case studies with numeric before/after | Supplies evidence snippets for “examples of X” queries. |
| 10 | Structured markup (FAQ / HowTo schema) | Improves extraction when assistants browse live pages. |
| 11 | Quarterly refresh cadence (visible “Updated Month Year”) | Keeps you eligible for “latest” answers. |
| 12 | Press/fact sheet page (concise stats list) | Central reference; reduces hallucination risk. |
| 13 | 2–4 additional external mentions (diverse domains) | Strengthens authority weighting. |
| 14 | Named proprietary framework (clearly defined steps) | Gives users & models a hook to reference you explicitly. |
| 15 | Simple on-site calculator or estimator (plain text logic) | High utility asset; can be described by browsing models. |
3. Nice To Have (Do If Resources Allow)
| # | Item | Why |
|---|---|---|
| 16 | Glossary expansion beyond 25 terms | Incremental coverage; diminishing returns after core terms. |
| 17 | Prompt examples (“Ask an AI: …”) | Encourages user-driven model association. |
| 18 | RAG site chat trained on curated docs | Improves conversion; secondary for being listed. |
| 19 | Prompt card / downloadable cheat sheet | Scales user seeding of your terms. |
| 20 | Versioned changelog for data pages | Extra freshness signal; trust reinforcement. |
| 21 | Micro tool public (e.g. custom GPT) | Discovery surface; not yet core unless your audience uses those directories. |
Weighted Scoring (Quick Audit)
Assign weights: Essential=3, High Impact=2, Nice=1, Marginal=0.
| Layer | Items Complete | Points |
|---|---|---|
| Essential (8) | x * 3 | 24 max |
| High Impact (7) | y * 2 | 14 max |
| Nice (6) | z * 1 | 6 max |
| Total | 44 max: |
-
30+ points: Strong presence foundation
-
24–29: Adequate; finish remaining Essentials + a few High Impact
-
<24: Fix gaps in Essentials before anything else.
In conclusion, if you want to continue to ongoing arms race of lead generation process improvement, follow my guide above to learn what you need and feel free to contact me if you're looking for pointers. I'm always available for questions and feedback.
Drop your comments below!