Use cases by function7 use cases

Sales automation that does the research, not the spamming

The cold-outbound wave taught everyone what AI in sales should not be. Tools that send personalized-sounding email at scale burned domains and goodwill, and buyers now recognize it in one line. The work that pays is quieter: answering inbound interest quickly and correctly, doing the research a rep would otherwise do by hand, turning a request for quote into a draft quote, and keeping the CRM true without asking anyone to type.

Sales and marketing is already where businesses use AI most. Census Bureau research published in April 2026 found that 52% of US firms using AI apply it in sales and marketing, ahead of every other function. Most of that is writing help. The builds here go further: they read from and write to the CRM, the ERP and the calendar, and they act inside rules someone on your team owns.

In practice that means every inbound lead researched and routed before a rep opens it, RFQ emails with drawings turned into draft quotes, freight quote requests answered in minutes, a CRM that updates after every call, and trades that stop losing jobs to voicemail.

Sales and lead handling systems I can build

Where AI earns its place in a sales process

  • Speed on inbound, not volume on outbound

    The value is in answering the people who already asked, quickly and with the right information, rather than in contacting more people who did not.

  • Research a rep would do anyway

    Company size, systems, recent news, an existing account, an open support ticket: gathered before the first reply, so the rep's first minute is spent selling rather than searching.

  • In B2B, the quote is the bottleneck

    RFQs, freight quotes and tenders are won by whoever answers first with a correct number. The model reads the request; pricing rules stay in code; a person sets the final price.

  • The CRM as a by-product

    Field updates proposed from calls and emails, each with the sentence that justifies it, approved in one click. Pipeline reports start telling the truth because nobody has to remember to update them.

  • What I would not automate

    Negotiation, pricing exceptions, anything that commits the company, and cold outreach at volume. The first three need judgment; the last one mostly damages your domain.

Frequently asked questions

Can AI qualify leads automatically?

Yes, if qualification is written down. The model extracts signals from the form, the email and enrichment data; a rubric in code scores them; the lead is routed with the reasons attached. What should not be automated is the rubric itself. Sales leadership owns it, and it is reviewed against won and lost deals every quarter.

Does this work with HubSpot and Salesforce?

Yes. Both have mature APIs, and HubSpot now has an official MCP server with read and write access to CRM records. The builds read contacts, companies and deals, write back proposed updates with an audit trail, and respect the fields your team already uses rather than inventing new ones.

Is lead enrichment GDPR compliant?

It can be. For business contacts, legitimate interest is usually the lawful basis, which means a documented balancing test, enrichment limited to business information, a clear privacy notice and an easy way to object. Scraping personal social profiles is where most enrichment setups cross the line.

How do I automate lead follow-up without annoying people?

Follow up on what the person actually asked about, stop the moment they reply, cap the number of touches, and hand warm replies to a person immediately. Personalization that is only a first name and a company name reads as automation. Personalization that references their actual request reads as service.

Also browse: HubSpot · Salesforce · SaaS · Manufacturing and distribution · all use cases