Stepthrough Start free

n8n Showing Feedback Agent: 9 Mistakes to Avoid

Quick answer

The biggest mistake with an n8n showing feedback agent is letting it speak about price. One buyer's agent saying a home is "about $15,000 high" is an opinion, not a market verdict, and an AI summary that turns it into "consider a price reduction" puts words in your mouth in front of the seller. The agent in the free n8n showing feedback lesson is built to avoid that: it records only what was said, quotes price comments in the agent's own words and never recommends a price change. The other mistakes are about privacy, vague feedback and wiring.

Last updated

Take the free module →Module 11 of the free AI for Real Estate course, about 12 minutes. No account needed.

n8nReal EstateAI agentsGoogle SheetsTroubleshooting

The example is made up. Meridian Realty Group is a made-up brokerage, and 14 Orchard Lane and every agent and comment are made up too. The agent turns a week of feedback into one row per showing in a Google Sheet: date, buyer's agent, interest, price feedback and follow-up.

What are the most common showing feedback agent mistakes?

#MistakeFix
1Letting the agent recommend a price changeKeep "do not recommend a price change"
2Rewording price commentsQuote them, or "None given"
3Letting it guess offersKeep "do not guess what a buyer will offer"
4Pasting buyers' names and numbersPaste the agent and brokerage only
5Forcing vague feedback into a levelLet it use Unclear and say why
6Column headers that don't match the promptDate, Buyer's agent, Interest, Price feedback, Follow-up
7Leaving the Sheets node on reading rowsPick Append Row under Operation
8Skipping memoryAdd Simple Memory so "save those" works
9Sharing the sheet unreadRead every row before the seller sees it

Which mistakes can cost the seller's trust?

1. Letting the agent recommend a price change. The system prompt says not to recommend a price change, because the listing agent and the seller decide that. When you adapt the prompt for your brokerage, keep that line. An agent that offers pricing advice in a sheet the seller reads is speaking for you.

2. Rewording price comments. The prompt asks for price feedback in the buyer's agent's words, or "None given". So Dana Ruiz's comment goes into the log as "About $15,000 high." in quotes. Paraphrased, it can grow into "overpriced" and read like a consensus.

3. Letting it guess offers. The prompt says not to guess what a buyer will offer. Marcus Hill's buyers plan to write an offer this week, so the row says that, with High interest and "send the roof age today" as the follow-up. It doesn't predict a number.

What should stay out of the log?

4. Pasting buyers' names and numbers. The prompt says never to include buyers' names, phone numbers or personal details. The Buyer's agent column holds the agent's name and brokerage only, and the lesson's finish screen says to keep buyers' names and personal details out of what you paste. Personal details are PII; the fewer you paste, the fewer can leak.

5. Forcing vague feedback into a level. The interest levels are High, Some, None and Unclear. The prompt says that if the feedback is vague or mixed, use Unclear and say why. Don't edit that out to make the sheet look tidier. An honest Unclear tells you who to call back.

Try it free, step by step

Build an n8n AI Agent That Turns Showing Feedback into a Seller Update. Build an AI agent in n8n for Meridian Realty Group that turns buyer agents’ showing feedback into one row per showing in Google Sheets, with interest, price feedback and the next step, so the weekly… You do every step yourself in a practice copy of n8n, it checks your work as you go, and “Do it for me” finishes any step you get stuck on. Free, and nothing touches your real accounts.

Start the free lesson

Or read the full step-by-step guide

Why doesn't the log fill in?

6. Column headers that don't match. Type five headers in row 1: Date in A1, Buyer's agent in B1, Interest in C1, Price feedback in D1 and Follow-up in E1. They match exactly what the system prompt returns. Then in the Sheets tool, under Values to Send, click the ✦ button next to each column so the model fills it in.

7. Leaving the node on reading rows. n8n adds the Sheets tool as Get row(s) in sheet in Google Sheets, which reads rows by default. Open Operation and pick Append Row. In Document, pick Showing Feedback Log, not Listing Marketing Calendar next to it, then Sheet1 in Sheet. The n8n Google Sheets errors post covers sign-in and permission problems.

8. Skipping memory. Without memory, every message starts from nothing, and "save those" points at nothing. Add Simple Memory under Memory; the defaults keep the last 5 messages. Also check the key: a new OpenAI key with a zero balance just returns errors. Click Add to credit balance under Billing first, or pick Gateway credits in your own n8n to use its free credits. Copy the key when you make it; OpenAI shows it only once, so treat it like a password. See API key.

What has to happen before the seller update?

9. Sharing the sheet unread. The audience in the prompt is the listing agent, who reads every entry before sharing anything with the seller. The lesson's finish screen says the same: read every row before you share the sheet. Follow your brokerage's rules on what you share with sellers. That review is human in the loop, and it is what keeps the update one your clients trust.

Key takeaways

  • The agent never recommends a price change or guesses an offer. Keep those rules.
  • Quote price comments in the buyer's agent's words.
  • Log the agent and brokerage, never the buyers' names or numbers.
  • Let vague feedback be Unclear, with the reason.
  • Match the headers, use Append Row, add Simple Memory, and read every row.

Read the n8n showing feedback agent FAQ for setup answers, or try the buyer reply lesson in the same course.

Frequently asked questions

What if all the price comments say the home is high?

The log still quotes each one. Whether to change the price is a conversation between you and your seller, not something the agent decides.

Why did one showing get "No follow-up needed"?

Priya Shah's buyers need four bedrooms, so the home isn't a fit. The agent logs None interest and no follow-up.

Can I use it for every listing?

Yes. Paste each listing's feedback separately and name the listing in your message, as the lesson does with 14 Orchard Lane.

Where can I learn more about the nodes?

The n8n docs cover the AI Agent, memory and Google Sheets nodes.

Free lessons in this post

Plain-English definitions, each with an example from a free lesson.

The full AI glossary

  • Case study interview questions prompt: Plan a client interview that gets a case study's before, during and after story in their own words, with follow-up probes and…
  • Email subject lines prompt: Get ten honest subject lines for a finished email, each with preview text and a character count, grouped by approach so you can…
  • SEO meta description prompt: Write short search-result descriptions for your web pages: what the page offers, who it is for and one reason to click, inside…
  • Social post ideas prompt: A four-part ChatGPT prompt for social post ideas built on your services and real customer questions: platform, hook and call to…

All free prompt templates

Keep reading

All blog posts · All free tutorials