How Rankup automated its prospecting without losing control
Rankup wanted to prospect more efficiently, without letting AI act on its own on the agency’s behalf. Together with the team, we designed a semi-autonomous prospecting system: AI handles the repetitive work, people keep control of the decisions.
- Client
- Rankup
- Industry
- SEO & prospecting agency
- Outcome
- Up to 80% of prospecting time saved
Context
Rankup is an agency specialising in online visibility: ranking in traditional search engines (SEO), and being cited by AI-powered answer engines (GEO).
Like any agency, its growth depends on finding new clients. Prospecting is demanding work: identifying the right companies, understanding their situation, then writing each one a genuinely relevant message.
The challenge
Rankup wanted to automate part of its prospecting operations, but not at any cost: letting AI send generic messages on the agency’s behalf, or act without supervision, was out of the question.
The challenge was twofold: save time on repetitive tasks while retaining human control over the actions taken by AI.
Our approach
We started from a simple principle: AI only writes a good message when it has the right context. A prospecting email is only relevant if it relies on accurate information about both Rankup and the company being contacted.
We then drew a clear line between what AI can do alone and what must remain a human decision. This is the “Human-in-the-Loop” principle: AI prepares, people approve.
What we built
The data foundation
We first structured the knowledge needed for personalisation: what Rankup does, its offers and arguments, and information about prospects collected automatically from Google.
This knowledge base powers a RAG retrieval system: before writing, the AI looks up the relevant information for each prospect instead of relying on its general knowledge.
The AI layer
On this foundation, we built a semi-autonomous prospecting system made of several AI agents, each with a specific role, all the way to generating personalised marketing emails.
Key steps go through human approval: the team can review, edit or discard what the AI has prepared.
In parallel, we designed an AIOS for Rankup: a set of AI tools that automates various SEO and GEO operations carried out for its clients.
Results
Up to 80% of the time spent on prospecting is now saved, thanks to the automation of tasks handled by AI.
The time freed up can go to what genuinely needs a person: approving, talking to interested prospects, closing.
« I strongly recommend their services, which are very professional, with particular attention to our needs and outstanding speed. Such quality of service is rare. I actually intend to continue working with them. »
TechnologiesRAG · LLM · AI Agents · Human-in-the-Loop · AIOS · SEO/GEO