AI for Sales Webinar January 22nd
This webinar shows how to use AI across the sales funnel: from free deep research in ChatGPT or Perplexity to Claude Code pipelines that find leads with Apify, clean them, research intent signals and turn meeting transcripts into proposals.
Pour qui: For business owners and sales leads who spend hours finding, cleaning and researching prospects and want to hand that work to AI while keeping the actual conversations human.
À retenir
Sales is a whole funnel, not only the meeting
I split sales into lead sourcing, data cleaning, research, contacting, the meeting, the business proposal and closing. Most of the time goes into the first three steps, so that is where AI saves the most. By the end of the webinar I estimate about 80% of the funnel is covered, with the contact step left to people.
One deep research can find a company's decision makers
Paste a company URL into ChatGPT deep research and ask for the names of its decision makers. If it finds even one email, company emails follow the same structure, so you can ask it to extrapolate that pattern for everyone else. Claude, Perplexity and Gemini have the same research feature. For a $20 subscription it works for around 10 to 20 companies.
Connect AI to tools instead of expecting AI to do everything
AI will not solve everything by itself. The results come from knowing which tools to use and connecting them to AI. I have Claude Code run an Apify Leads Finder search (C-level, small and mid companies, El Salvador, professional services), download the results and remove duplicates and people with no email: 99 leads, 23 removed.
Timing matters: look for intent signals
Sending the right message to the right person is not enough; it has to arrive at the right time. People who comment on a busy LinkedIn post, or who appeared in the media, raised funds or were acquired recently, are showing intent. Apify scrapes comments, reactions and reposts, and Perplexity or a model can sort out who is really interested.
Save the steps as a workflow and view the results in a dashboard
Once sourcing, cleaning and research work, ask Claude Code for an HTML dashboard, because rows of CSV are hard to read. Then tell it to save the workflow so every future lead request runs all the steps. I put the cost of 1,000 leads who interacted with a post yesterday at about $10.
Your campaign replies hold free data
After each Instantly campaign, a Claude Code workflow moves non-repliers to a list to contact again in three months. It adds out-of-office replies to the CRM with a follow-up date for when they return, and pulls phone numbers from email signatures (60 to 70% of cases). Positive replies go to the CRM with next steps, and opt-outs and bounces go to separate lists.
Automate everything except the human contact
I am against automating LinkedIn messages because everyone is already flooded with AI-written messages and people want human contact. Cold email is the exception, since you send the same email to many people. After the meeting, have AI analyze the transcript for pain points and write the proposal in the prospect's own words.
Outils présentés
- ChatGPTDeep research on companies and people, finding decision makers and email patterns, and agent mode to pull local businesses from Google Maps.
- PerplexityDeep research in the app, and through its API to check which leads appeared in recent media coverage.
- ClaudeResearch feature for finding decision makers; the Anthropic API was used to sort LinkedIn commenters by interest.
- GeminiAlternative deep research tool that works as well as the others.
- ApifyScrapers (actors) for leads: Leads Finder with natural-language search, LinkedIn comment scraping and a cookieless LinkedIn profile scraper.
- InstantlyCold email campaigns, plus lead lists (block list, no reply, bounced) that the AI analysis workflow maintains.
- Claude CodeThe central agent that runs Apify, cleans data, calls APIs, builds dashboards, analyzes campaigns and transcripts, and drafts proposals.
- FirefliesRecording meeting transcripts that Claude Code analyzes for pain points.
- NotionCRM where positive and polite leads from Instantly are imported.
- TypstFree formatting language for producing clean-looking PDF business proposals.
- Google AntigravityThe editor I run Claude Code in, and a budget alternative with a generous free tier.
- Gemini CLIBudget alternative to Claude Code for people on a $20 plan.
- AnyMailFinderFinding emails from LinkedIn profile links.
Questions du public
How do I keep my Claude Code setup organized so I can reuse it as a service for other companies?
I keep my working rules in a CLAUDE.md file that I improve with Claude Code whenever something goes wrong. Build a separate, specialized framework for each kind of work (SEO, web development, workflow automation), since one brain does not fit every job. To share a framework, hand over the CLAUDE.md file. To share a workflow, ask Claude Code to turn it into an application with a visual interface, deploy it with basic authentication, and the other person can use it without touching Claude Code.
I run business development and have a tech team. Should I build this myself or with them?
It depends on how complex the use case is and how fast you need it. Your developers may be busy or not yet used to building internal tools with AI. Check the level of AI adoption in your organization, then give them a product requirement document (PRD) and tell them you don't need it to be extremely robust, you need it to work within about a week. Building apps is cheap now, and simple commercial use cases should take a couple of days at most.
I'm on a $20 Claude plan. Which AI tools should I use to save money?
On a budget, I recommend Google Antigravity or Gemini CLI instead of Claude Code, because Google's plans are very generous while it tries to win users. Gemini CLI works just like Claude Code. I pay $100 a month for Claude Code out of habit and switch to Antigravity for a few hours when I hit limits. For a student or someone watching costs, paying for Claude Code doesn't make sense.
I exported my LinkedIn contacts. Can Claude Code clean them and qualify them?
Yes. The export gives you profile links but not the profile details, so first have Claude Code enrich the list with the Apify LinkedIn Profile Scraper (No Cookie, around $3 per 1,000 profiles) using your Apify API key. Then ask it to qualify everyone against your ICP, which needs an API key for an AI model. For emails, run the profile links through a service like AnyMailFinder. It takes me about fifteen minutes.
