About No-Hassle Digital
No-Hassle Digital is a paid-advertising agency founded in 2013, operating from Serbia and Dubai. Its focus is Google Ads, along with Microsoft, Meta and Amazon Ads. It is a Google Premier Partner and Amazon Verified Partner, and it serves more than 500 service-based, B2B, software and eCommerce clients in the US, UK and Australia.
The problem
Winning work on Upwork depends on speed and relevance. The best jobs get many proposals within hours, and generic proposals lose. Writing a strong, personalised proposal for every relevant job takes a senior specialist’s time: reading the brief, choosing the right approach, finding a matching case study and answering each of the client’s screening questions.
What the agent does
Solution design first. Before building anything, Amsby and No-Hassle Digital mapped the whole sales process together in a decision-flow design: how jobs come in, how clients are separated from agencies, which of three proposal paths applies, and when an industry reference, a technical-experience section or screening-question answers are needed. The agent was then built to follow that agreed flow exactly.
Pipeline 1: Job qualification and proposal drafting
- Daily job intake. New Upwork job posts arrive every day through the agency’s ExpandCrew job feed and land in a Google Sheet, with title, description, budget, experience level, category and up to six screening questions. Jobs qualified as a fit go forward.
- Client vs. agency routing. Gemini classifies whether the post comes from a direct client or another agency. Direct-client jobs go into the proposal pipeline. Agency posts are routed separately to a set-up-call track as potential partnerships.
- Template selection. The job is sorted into one of three proposal paths, each with its own template: clients with an existing Google Ads account (optimisation and management), lead-generation clients without a Google Ads account (setup) and audit requests.
- Deep job analysis. Gemini produces a structured JSON brief covering the client’s core problem, the required technical skills (Performance Max, Shopping, conversion tracking, GA4, GTM and so on), the industry, whether an industry reference is needed, every screening question and whether it needs a knowledge-base lookup, proposed first steps and a call to action.
- RAG-powered drafting. The proposal writer uses a Pinecone retrieval tool to fetch the agency’s own proposal template for that job type, keeping its exact structure, headings and tone; relevant past case studies from the same industry, when the job asks for them; and example answers to common screening questions, adapted to this job.
- Validation. An automated check confirms the subject line is there, every screening question is addressed and an industry reference is included where needed. Any gaps are logged.
- Delivery. The proposal is saved as a Google Doc and the link is written back to the job row in the sheet, ready for a specialist to review and submit.
Pipeline 2: Self-updating knowledge base
A Google Drive watcher picks up every new template, case study or answer document dropped into a shared folder. It splits the document into chunks, embeds them with Gemini embeddings and stores them in Pinecone. This keeps the agent’s knowledge current without any code changes.