Pick a depth. Each prompt opens in your AI pre-loaded with the lesson. Click a row to preview the prompt.
Research is the stage where AI genuinely changed the economics of selling: what used to be an hour of tab-hopping per prospect is now a 20-minute sprint, and the output - three facts a generic sender would not know - is the difference between an email that reads like spam and one that reads like it was written by someone paying attention. Buyers now receive AI-generated cold email daily and have learned to delete anything that could have been sent to anyone; specificity is the only remaining signal of effort. The discipline matters as much as the tools: unbounded research becomes procrastination with a professional face, which is why this task is timed. Twenty minutes, four sources, one half-page brief. Master this sprint and you can prepare ten genuinely personal emails in an afternoon - the leverage that used to require a research assistant.
The sprint, with a timer running. Minutes 0-5, company: open the company site, read the homepage, the pricing page, and the two most recent blog or changelog posts; paste anything interesting into a scratch doc. Minutes 5-10, person: their LinkedIn profile - role, tenure, last three posts or comments, anything they have written elsewhere. Minutes 10-15, situation: search Google News for the company name, check their careers page for open roles, skim their latest product announcement. Minutes 15-20, synthesis: paste your scratch doc into Claude or ChatGPT with this exact prompt: 'Below are my raw research notes on [name], [role] at [company]. Write a half-page brief with three sections: (1) What this company does and how it makes money, in two sentences. (2) What is probably on this person's mind this quarter, based only on evidence in my notes - cite which note each inference comes from. (3) The three most specific, non-obvious facts here that a cold email could reference. Flag anything you are guessing at as GUESS.'
Here is what a real output looks like for the Priya example from the last task: '1) HelpDeskCo sells help-desk software to mid-market e-commerce companies, priced per agent per month. 2) Likely on her mind: onboarding - three open support roles against a team of ten (careers page) means ~30% headcount growth this quarter; and self-serve deflection - her one LinkedIn post last month asked how other teams cut repeat tickets. 3) Three specific facts: she was promoted from senior agent to head of support 14 months ago (LinkedIn); the company launched a Shopify integration three weeks ago (changelog), which typically spikes new-user questions; her post about repeat tickets got 40+ comments, so the pain is public and current.' Two warnings that keep this honest: never let the AI research from its own memory - it will confidently invent funding rounds and job changes; you feed it only what you gathered from live sources. And verify the load-bearing fact: if your whole email leans on the Shopify launch, spend sixty seconds confirming it is real and recent. One hallucinated 'congrats on the funding' destroys you with that buyer permanently.