My service manager needs to look up commercial property information to build proposals. We are subscribed to CoStar for $600/month for one seat. I built an app for him instead. It started out as a simple lookup tool using Attom Data. A large dataset for commercial properties (and residential). As the app was coming together, I added our standard services which are saved in a table for easy recall. He can build a quote right in the app. This will evolve into a full proposal by sending all the data to PandaDoc template (like DocuSign) with AI building a proper proposal based on the services selected. The app has further evolved to be a tool for any small business that sends quotes to commercial accounts. So I added a Stripe integration to charge for 3 different subscription levels allowing more lookups per month, and some other features. Now it is also useful for real estate agents. When they talk to someone about selling a house, they need to come up with a price. So I also have a lookup tool another platform called RentCast which is strong with residential property data, sales data, market analysis, and even likely rent for a house. An agent can now lookup a house, generate a report that gives them an estimated price (like Kelly Blue Book for a car) with a low, medium and high valuation based on recent sales in the area of the same type (single family, condo, etc) and number of bedrooms. It also provides an estimate of rent that could be charged if the house is being purchased as a rental. All of this has been done using Lovable and Claude, using , n8n and Stripe. I’ve got six n8n workflows for various tasks. It’s not fully baked yet. I still need to build out all the fluff required for a website, but if you want to see it go to cresearchai.com I’d like your feedback or suggestions.

This video is a complete walkthrough and in-depth tutorial on how to build advanced conversational AI voice agents using the ElevenLabs platform, taking you from beginner to expert. You will learn to master every feature, including the new Workflows canvas for building complex, multi-agent systems with conditional logic, selecting the best Large Language Models (LLMs) like GPT-4 and Claude, and connecting external tools like n8n or Zapier with webhooks to book appointments and update CRMs. Check it out 👇

Hi guys, I’m not experiencing a technical issue as such, but I’m attempting to get vapi to capture the address and postcode of the user. This seems to be very difficult as it hears it wrong address and postcode a lot of the time and notes it down incorrectly in make.com as addresses technically aren’t real words. I’ve tried getting the user to spell it out and vapi to spell it out back in confirmation, however this seems very inefficient as I fear most people will get annoyed by having to spell out a whole postcode and address, and then have it thrown back at them pretty fast to confirm that it is correct. Any suggestions? Would be grateful, thanks!