How We Use AI to Build Water Quality Proposals That Pull Live EWG Data

There is one objection that kills more water treatment demos than price, than timing, than anything else a rep faces at the door.

"I didn't know my water was bad."

It sounds like curiosity. It is not. It is the exit ramp. The homeowner who says this is telling you they were never pre-sold on the problem. Your rep showed up to educate a skeptic instead of close a believer. The demo becomes an uphill argument and your close rate reflects it.

The fix is not a better objection handler. The fix is ensuring the homeowner already knows their water is bad before your rep rings the doorbell. Here is how we built that system using publicly available government data and a few hundred lines of AI logic.

The Database Most Water Treatment Companies Have Never Heard Of

The Environmental Working Group maintains a public database at ewg.org/tapwater that tracks over 500 contaminants in public water systems across every municipality in the United States. It is sourced directly from utility-reported data that water systems are legally required to submit. It is freely accessible. It is organized by zip code and utility.

It is also one of the most powerful sales tools in residential water treatment, and most sales orgs are not using it at all.

Here is why it matters: EWG does not just report whether a contaminant exceeds the EPA's legal limit. It reports whether a contaminant exceeds EWG's own health guidelines — which are based on the latest independent research and are almost always more conservative than EPA maximums. The gap between legal and safe is where your sales conversation lives.

A homeowner can look at a water quality report that says "passes all federal standards" and walk away feeling fine. That same homeowner, looking at an EWG report for their zip code, learns that their water passes EPA legal limits but contains chromium-6 at 14 times the level EWG considers safe, nitrates at 5 times the health guideline, and disinfection byproducts linked to bladder cancer.

That is not a sales tactic. That is the truth about their water. Your system solves a real problem. EWG gives you the documentation to prove it.

What Dallas Tap Water Actually Contains

Let us make this concrete. Here is what the EWG database shows for Dallas, Texas:

• Chromium-6 (hexavalent chromium): 14x above EWG health guidelines. This is the carcinogen made famous by the Erin Brockovich case — a known human carcinogen at elevated exposure levels.
• Nitrates: 5x above EWG health guidelines. Associated with increased risk for colorectal cancer and thyroid disruption, and particularly dangerous for infants and pregnant women.
• Total trihalomethanes (TTHMs): Disinfection byproducts formed when chlorine reacts with organic matter. Linked to bladder cancer and adverse reproductive outcomes.
• Haloacetic acids: A second class of chlorination byproducts, present above EWG health thresholds.
• Radium-226 and 228: Radioactive elements occurring naturally in Texas groundwater that end up in treated municipal supply.

That is 17 contaminants in Dallas tap water that exceed EWG's health guidelines, even though the utility is in full legal compliance. Every one of those contaminants has a citation, a health context, and a comparison number a homeowner can understand.

When a homeowner in Dallas receives an email or text before the demo with a report showing their specific water data — not generic language about "water quality concerns" but the actual compound names, the actual multiples above safe levels, the actual health associations — they do not show up to the demo wondering if they need a water system. They show up wanting to know which one.

How the AI Layer Works

The raw EWG data is structured but not readable. It is API output: contaminant codes, ppb measurements, utility identifiers, reference values. What we built is an AI layer that sits between the raw data and the homeowner.

Here is the pipeline:

Step 1: Zip code lookup. When a lead submits their form (name, phone, address), the zip code is passed to the EWG API or a pre-pulled data file for that territory. We get back the list of contaminants detected in that utility's water supply with measured concentrations and EWG health guideline values.

Step 2: AI formatting and contextualization. We pass the raw contaminant data to Claude (Anthropic's LLM) with a prompt that instructs it to: identify the contaminants that exceed EWG health guidelines, rank them by severity (multiple above guideline), write a plain-English description of each — what it is, why it matters, what the number means in context — and format the output as a clean, readable report.

The output is not templated language. It reads like a practitioner explaining the data. "Chromium-6 is a form of chromium that forms naturally in the environment but is also a byproduct of industrial processes. It has been classified as a known human carcinogen. Dallas tap water contains chromium-6 at 14 times the level EWG considers safe for long-term consumption."

Step 3: Delivery. The formatted report is sent to the homeowner as a pre-demo email or SMS — timed to go out when the appointment is confirmed. Reps also have it available on a tablet to pull up during the in-home visit as a third-party credibility reference. In some workflows, it generates a one-page PDF the rep can leave behind.

The whole process — form submit to formatted report delivered to the lead — runs automatically, with no rep involvement.

How Reps Actually Use This in the Field

The EWG data does three different jobs depending on where it appears in the sales process.

Before the demo (pre-sell): This is where it does the heaviest lifting. A rep who shows up to a home where the owner received a contamination report 24 hours earlier is not starting the conversation at zero. The homeowner has already Googled chromium-6. They have already looked at the EWG website. They are already asking "what do I do about this?" The rep's job shifts from problem identification to solution presentation. Show rates go up because homeowners do not cancel appointments they are already emotionally invested in.

During the demo (credibility): When a homeowner pushes back — "how do I know this is real?" — the rep opens the EWG website on a tablet and shows them their zip code directly. EWG is a nonprofit. The data comes from government reporting. It is not something our company produced. That third-party credibility source eliminates the "you're just trying to scare me into buying" objection cleanly.

After the demo (follow-up): For appointments that do not close on the day, a follow-up sequence that includes the contamination report as a PDF keeps the health concern present in the homeowner's mind. Most sales orgs are sending "just checking in" follow-ups that get ignored. A follow-up that says "here is your water quality data from the Environmental Working Group" is a reason to re-engage.

Why This Changes Your Numbers

We track this across our territories and the pattern is consistent.

When a homeowner receives a personalized water quality report before the demo, three things happen:

Show rate improves. The homeowner has a reason to show up that is separate from trusting your rep or believing your ad. They have already started their own research. Cancellation is harder when you are halfway through reading about what is in your water.

Early objection volume drops. The "I don't know if my water is bad" objection disappears almost entirely. The rep starts the demo with a homeowner who has already accepted the premise. The conversation moves faster to the product.

Close rate on shown appointments increases. We see close rates 15–25 percentage points higher on leads who received and engaged with the pre-demo report versus leads who did not. The homeowner who walks in skeptical requires more selling time and closes at a lower rate than the homeowner who walked in pre-sold on the problem.

This matters at scale. If you are running a five-rep sales org with each rep running 12 demos a week, moving from a 30% close rate to a 45% close rate on shown appointments is not a marginal improvement. It is roughly nine additional deals per week across the team.

How to Replicate This for Your Sales Org

The components are straightforward. The EWG data is publicly available — you can pull it for any territory you operate in. The AI formatting layer requires prompt engineering and a connection to a large language model API, but the logic is not complex. The delivery mechanism connects to whatever CRM or follow-up system you already run.

What you need:

1. EWG data for your territory. Pull the contaminant list for the zip codes you operate in. Identify the top 5–8 contaminants that exceed EWG health guidelines and their multiples above the threshold. This is your city-specific content library.

2. An AI formatting prompt. The prompt takes raw contaminant data and instructs the model to explain each finding in plain English, ranked by severity, with health context and citation to EWG. The output should read like a concerned expert explaining the data to a non-technical homeowner.

3. A trigger in your lead flow. When a new lead is created — from your Facebook form, your website, or your CRM — an automation pulls their zip code, matches it to the EWG data for that utility, runs the AI format, and sends the output to the homeowner. This can run in GoHighLevel, N8N, Make, or any automation platform.

4. A PDF template. The formatted report drops into a one-page PDF with your brand and a clear call-to-action for the scheduled demo. Reps carry this on their tablets.

The goal is that by the time a rep arrives at the door, the homeowner has already read a factual document — government-sourced, third-party validated — explaining exactly what is in their water. The rep walks in to confirm what the homeowner already suspects and present the solution.

For details on how we drive the leads that feed this system using territory-specific Facebook ads, see How to Get More Water Treatment Leads with Facebook Ads.

What Else the AI Handles in Our Sales Process

The water quality proposal is one piece. The broader system automates most of the gap between ad click and rep showing up at the door.

AI-generated video ad scripts pull from the same EWG data — the Facebook ad for Dallas references chromium-6 at 14 times safe levels because the ad script was written using that specific data point. The same territory intelligence feeds the ad creative, the pre-demo report, and the rep's talking points.

Speed-to-lead is handled by an AI calling system that contacts every new lead within 60 seconds of form submission, qualifies them, and books the demo automatically. The AI calling system is covered in depth in a separate post — but the short version is that 60-second response time drives a 38–45% contact rate on fresh leads, compared to 10–20% when calls happen manually hours later.

The combination — AI-generated local ad creative, AI contamination reports pre-selling the homeowner, AI calling booking the demo, rep arriving to a pre-sold appointment — is the system that produces $10–$20 CPL and 35–45% close rates on shown demos.

No single piece of this is magic. Each one closes a specific gap in the funnel. The EWG proposal layer closes the gap between "lead who filled out a form" and "homeowner who is genuinely motivated to buy before the rep arrives."

---

If you run a water treatment sales org and want to see what the EWG data looks like for your territory — and how this system would fit into your current lead flow — book a free strategy call. We will pull the contamination profile for your market and show you exactly what a pre-demo report would look like for your homeowners.