Skip to content
Amsby

AI agents & automation

From three form fields to validated leads and a personalised email.

Amsby built an n8n and Google Gemini agent that researches a target account, finds and checks decision-makers, and writes a tailored cold email in Google Sheets.

Client
Control of Work software company
Sector
Enterprise software (Control of Work / LOTO)
Type
AI agent for sales prospecting
Engagement
04 · Project-based

About the client

We built this agent for a client of one of our partners, so we don’t name them here.

The client is an enterprise software company. It provides Control of Work solutions, mainly Lockout/Tagout (LOTO) and isolation management, for high-risk industries such as oil and gas, utilities, mining, chemicals, manufacturing, construction and data centres. Its buyers are safety, operations, maintenance and facilities leaders. That is why the agent targets exactly those roles.

The problem

Researching a target account by hand takes a lot of time. You have to understand the company, find the right decision-makers, check they still work there and write a relevant first message. Done manually, this slows down outbound sales, and the quality depends on who does the research.

What the agent does

The user enters just three fields (company name, country and city), and the agent runs the full research-to-outreach process:

  1. Company research. It runs a targeted web search on the company through the Serper Google Search API, localised to its country and city, to collect background information.
  2. Decision-maker discovery. It searches public LinkedIn profiles for people at that company in the target roles. In this deployment those were C-suite, Safety/EHS, Operations, Plant/Maintenance and Facilities leadership.
  3. AI validation and deduplication. Google Gemini reviews every candidate profile against the company and its locations. It removes duplicates and implausible or mismatched profiles, keeps only the most credible person when several claim the same title, and filters for relevant roles.
  4. Structured lead output. Verified contacts (name, role, LinkedIn URL, company) are parsed into structured data and added to a lead list in Google Sheets.
  5. Personalised outreach. Gemini combines the company research with the client’s own base sales message to write a tailored cold email: subject, body and closing, under 150 words, referring to specific details about that company.
  6. Status tracking. If no verified contacts are found, the agent flags the account in the sheet, so nobody wastes effort on it.

At a glance

input fields (company, country, city)
3
automated steps from research to outreach
6
words per personalised cold email
<150

Business value

  • Minutes of manual research per account become an automated run that starts from a single form.
  • Every account gets the same research and validation logic, so quality is consistent.
  • AI-checked contact data reduces time wasted on wrong or duplicate profiles.
  • Messages are personalised at scale while keeping the client's own value proposition and tone.
  • Everything lands in Google Sheets, so sales teams keep working in tools they already use.

Technologies

  • n8n
  • Google Gemini 2.5 Flash
  • Serper (Google Search API)
  • Google Sheets API
  • JavaScript
  • Prompt engineering

Hand repetitive work to an AI agent?

Built on your own data and tools, with a human approving anything important.