In this newsletter, I have never had the chance to zoom in on using AI for sales. Yet there are plenty of time-consuming, repetitive tasks, like researching prospects, entering data, preparing a meeting, or generating personalized content after a call...
It is one of the departments that gains the most from adopting AI and letting teams focus on meetings (rather than generating them).
I will try to share the classic use cases with you. Some are large automations with a touch of AI and others are full AI agents. I will distinguish uses in “copilot” and “autopilot” mode. The latter is more valuable if you have a high volume of tasks.
(If you have questions, you can drop a comment at the end of the edition or reply to this email)
AI copilot
These examples work for any company (mainly B2B) and mostly save time on small tasks while improving the quality of the output.
Several examples:
Automatic note-taking + action(s)
Whether meetings are in person or remote, note-taking tools have become essential to capture all the information shared during a meeting. These tools can be invisible via the extension; you don't even notice the recording is happening. The cheapest ones start at $10/month. If you have few meetings, you can stay on the free version.
But where it gets powerful is when you automate the repetitive actions a salesperson or account manager would do.
In 99% of cases, there is CRM synchronization: filling in the variables, creating a summary note, linking the full transcript to the account, and so on.
Then other actions can be considered: generating a post-meeting email from a template directly in the salesperson's inbox, creating a document (quote, report, invoice, …) or an internal notification.
Meeting preparation
Every day, AI can organize meeting preparation for the team. For this, there is no need to reinvent the wheel, just to reproduce the internal business logic.
This can mean pulling in CRM information + external information about the company + the latest news, and generating a complete Google Doc per meeting with all the information.
By anticipating, AI can even identify objections or friction points relative to the company's value proposition and generate suggested answers.
Here, for example, the “DailyMeetingPrep” assistant is triggered every business-day morning to list the meetings and prepare the ones that are sales meetings. To do this, it calls another assistant, “PrepMeetingHubspot”, which retrieves internal information (CRM, HubSpot here + internal docs) and external information (web, social media) to generate a complete report.
Automatic message processing
In some of the companies we support, salespeople are swamped with emails, often customer requests. They may handle both prospecting and client portfolio management.
As with support, we set up automatic processing of incoming emails to, at the very least, categorize and prioritize requests: a question about a product does not have the same priority as a meeting request for the end of the week.
And to handle some tickets autonomously: archiving, automatic processing of the simplest tickets, and drafting replies for more complex requests by retrieving the context of the request.
Here is a preview of categorized emails + draft generation for each email that requires a reply:
When we set up multi-channel lead generation systems, we add automatic processing of replies. No need to spend time reading the negative answers “I'm not interested” or the bad-timing ones “we don't have this need for now, but get back to me in 6 months” → we analyze the replies to update the lead status to “lost” or “bad timing”, for example, automatically.
Here, as soon as we get a reply from a campaign on lemlist, we analyze the reply to update the status automatically with a simple Make automation.
AI autopilot
For the next examples, we go deeper into automation: part of the sales process can be automated. This applies to most companies: those with a broad market (high lead volume), or those with a niche market (where volume quality matters, so a lot of research / enrichment / monitoring work …)
Lead generation, enrichment & CRM synchronization
I shared on LinkedIn this workflow that describes the steps from lead generation through to enrichment and CRM synchronization.
One of the most powerful tools in this workflow is Clay, a kind of Google Sheets that lets you generate leads from sources:
Let me zoom in a bit to explain why this tool is relevant and includes really useful AI building blocks.
Step 1: Activate sources to generate leads
The idea is to activate sources that automatically generate leads into a table, from a source such as companies or people from filters, or leads from Google Maps. Here is an example:
Once the tool has created a list of prospects, that is where AI can step in, on the enrichment.
Step 2: enrich the leads relative to your business
It has become relatively simple to enrich variables with AI from a company or a prospect.
AI is very effective at information research, analysis and content generation (new variables) all at once.
A preview of variables you can enrich with AI:
Note that on the enrichment side, you can go very far, but each enrichment has a cost (an AI model API call or a scraper), so you need to find the right balance between gathering enough information to qualify the lead without going too far, to keep the enrichment cost low.
You can also add custom API calls to retrieve other information or use a custom AI agent for a specific goal.
Here, for example, the idea is to combine web search + AI (via ClayAgent) to automatically generate the following variables: a company summary, 2 icebreakers for an HR manager and 2 icebreakers for a finance manager. You can see the prompt on the right, the fastest Claude model (Haiku 4.5).
The distinctive feature of the agent is that it searches for information until the goal is reached (an iterative loop), unlike a simple API call.
Intelligent lead scoring
Once you have leads that are enriched and synchronized with your CRM, it is time to score them with AI. The big advantage is going beyond a simple comparison that adds or subtracts points, such as “if the company has more than 50 employees then 5 extra points”. With AI, you can analyze the company's activity to find information and adjust the scoring accordingly.
AI will take into account the qualitative information (nature of the business, persona, ..) and the quantitative information (headcount, revenue, founding date, age of the executive, ..) of the prospect as a whole.
Here is an example where we categorize the prospect into one of our personas (for personalizing outreach later on) with a score that takes into account all of the company's variables.
Each persona has different templates in a multi-channel campaign that we trigger.
Personalized content creation
During the enrichment work, I didn't get the chance to zoom in on personalized content creation with AI.
The idea is similar to what we saw above: use the variables and the models to generate a personalized message from the template you initially use.
This naturally adds a touch of authenticity to content that already performs well.
We just slightly change the template to add personalization while keeping the format that maximizes your reply rates (and avoid changing the content too much, otherwise it is hard to track results because you end up comparing apples and oranges).
Here, we have different icebreakers for each point of contact. These variables are added into the prospecting sequence (LinkedIn + email). We personalize only a few variables while keeping the template.
We personalize every exchange, not just the first interaction.
For example, we will slip a specific resource into the 2nd email of the sequence rather than another one because a certain value proposition has been identified.
A preview of a multi-channel LinkedIn + email campaign. We regularly add phone calls (a manual task for the salesperson to carry out).
And it works! Of course, when you target qualified prospects, personalize the approach and multiply the touchpoints across different formats, the results follow: 27% reply rate on this segment.
Managing your sales pipeline and nurturing
AI can also be very useful for follow-ups (pipeline management) or nurturing. The approach is to identify weak signals from the prospect in order to use a follow-up that is relevant to the company's context.
Concretely, across all the opportunities in the pipeline, a weekly search is run to pull the company's latest news, analyze it and select the item that is relevant to follow up with the prospect on their need “I saw that you just launched a marketing campaign in the Paris metro! Well done, I hope that …”.
Behind the scenes, this content can be added directly to the CRM with an alert to the account owner for manual validation of the follow-up, or fully automated in the case of nurturing.
Other use cases
I am sharing a few more examples we have put in place, in no particular order, which may spark ideas:
- a voice agent to call back leads from forms and reduce the “time to call” → often relevant for B2C companies running ads. The voice agent qualifies the lead (and checks the number). Everything is synchronized in the CRM so the salesperson can take over.
- a voice agent to handle inbound calls for the sales team when no one picks up. It makes sure you never miss a lead (especially if you get inbound calls where timing matters). The agent books a meeting on the salesperson's calendar and creates a task in your CRM.
- a sales coach to help the team easily find answers to technical questions (during a meeting or for post-meeting follow-up). It makes teams more autonomous and eases onboarding for new salespeople.
- a welcome video for each customer using Eleven Labs for the voice and HeyGen for avatar creation. Each customer receives a personalized video to onboard them onto the platform: a better experience + higher platform usage. With VideoAsk, it also lets you gather content from the customer.
