Sales development with AI is the part of selling that happens before the first conversation, built as a pipeline: find the companies that fit, learn enough about each one to have a reason to write, contact them, and pass the ones that answer to a person. A company needs it when its salespeople build lists by hand, when nobody can say why a given account was contacted, or when interest arrives through a contact form and goes cold in an inbox.
What I build
- A prospecting engine. Four stages: discover, enrich, score and reach out. It starts from a written description of the ideal customer.
- Account dossiers. One versioned record per company, under a declared schema, where every field keeps its source.
- A scoring rubric. A written scale with bands that decide who gets outreach, who waits and who is discarded.
- Outreach. Segmented sequences by email and LinkedIn, written from the dossier, with consent and opt-out rules in the path.
- Quotation funnels. Pages for the searches buyers actually make, and a quote request that produces a structured lead instead of a loose message.
- The handoff. Each real reply delivered to a salesperson with the dossier and the activity of the prospect.
Questions that show whether the base is there
- Can you describe your ideal customer in writing: kind of company, size, place, and what has to be true for a conversation to be worth the time? The engine starts from that document.
- Do you know who your best salespeople are, and can you measure it? Without that, there is no pattern for a system to learn from.
- Is the path to a sale divided into steps that someone can learn? When one representative does everything, the result depends on that person.
- When a buyer shows interest, does it become a record with fields, or an email in a shared inbox?
- Is there a domain ready to send? Outreach from a cold domain reaches filters, not people.
The answers decide the order of the work. A company with a clear profile and no data starts at the engine. One with demand and no structure starts at the funnel.
How the work runs
- Write the thesis. Who buys, why now, and what evidence would show it. One client needed suppliers likely to require industrial space near a corridor where manufacturers were installing plants. That sentence is a thesis.
- Discover and enrich. Agents and data sources find companies that match. Each dossier fills with contacts, roles, signals from the site of the company and fit data, with a source for each field.
- Score against the rubric. Each company gets a number and a band. Only the top band moves to outreach.
- Reach out by segment. Each thesis has its own sequence. A personalization that the dossier cannot back does not get written.
- Catch and hand off. Replies and quote requests become structured leads, and the salesperson sees what the prospect did before picking up the phone.
- Trace back. When a prospect converts, we reconstruct which evidence produced the conversation, and the rubric changes with what we learn.
What you receive
The engine and its queue, the dossiers in versioned storage, the rubric and the sequences as documents, the funnel pages, and a pipeline that can be inspected after the fact. Your team can open any conversation and read where it came from.
What this does not include
I do not sell lists, and I do not send volume for its own sake. The system does not negotiate, set prices or close: that judgment stays with your salespeople. And I do not report signed contracts as a result of mine, because the sales team owns that step.
Where the evidence is
I have run outbound since 2015. As part of a team of two in Monterrey, I helped run a remote sales operation of more than 60 people across Latin America for a company in the United States, with the customer journey divided into roles and standard scripts for the first contact. The engine described here started inside one of my own products and has been built for clients since, most recently for an industrial real-estate developer in northern Mexico. The quotation funnel was built for a distributor of industrial supplies, also in northern Mexico. The Cotizera case study describes the next piece, an agent that prepares the quote while the seller keeps the pricing judgment.