Know Who You're Talking To: AI-Powered Contact Role Classification from Email Conversations in Salesforce
Quick glimpse
A fast-growing B2B technology company needed its sales team to understand who they were really talking to inside each account. ABSYZ built an AI-powered solution in Salesforce that reads every logged email, classifies the sender’s role based on the conversation itself, and automatically updates the Contact record. Sales teams now have accurate, up-to-date stakeholder insights across every account, without anyone having to type in a single role.
Impact Delivered
- Zero Manual Data Entry for contact role updates.
- Every Logged Email automatically analyzed for stakeholder role signals.
- Up to 70% Less Time spent by sales reps researching and updating contact roles.
- 5 Stakeholder Roles classified consistently across every account.
Tech Stack
- Salesforce Sales Cloud
- Salesforce Email Logging
- AI-Powered Prompt (LLM Classification)
- Salesforce Automation
Client
- A fast-growing B2B technology company selling to multiple stakeholders inside each customer account, from leadership and finance to technical and engineering teams.
Project Highlights
- Email-Triggered AI Analysis: Every email received and logged in Salesforce is automatically analyzed by an AI-powered prompt that reads the message's intent and context.
- Role Detection Without Job Titles: The AI classifies contacts from the language they use, even when the sender's job title never appears in the email.
- 5 Stakeholder Categories: Contacts are sorted into Leadership Team, Technical Team, Product Engineering Team, Operations, or Finance Team.
- Automatic Contact Updates: The classified role is written straight to the Contact record on the related Account, with no manual data entry.
Industry
B2B Technology / SaaS
Challenge
In B2B sales, deals are won by engaging the right people: the leaders who sign off, the technical teams who evaluate, and the finance teams who approve budgets, but knowing who is who inside an account depended on reps keeping CRM data up to date by hand.
The core obstacles included:
- Missing or Outdated Contact Roles: Many Contact records had no role or an old one, leaving reps unsure who they were dealing with.
- Job Titles Rarely Tell the Full Story: Senders often don’t include a title, and titles alone don’t show whether someone is a decision-maker, evaluator, or approver.
- Manual Data Entry: Reps had to read emails, judge each sender’s role, and update records by hand, taking time away from selling.
- Blind Spots in Stakeholder Mapping: Without accurate roles, sales teams struggled to see which leadership, technical, or finance stakeholders were engaged in each deal.
How we helped
- Email-Triggered AI Analysis
We configured an automation that fires whenever an email is received and logged in Salesforce. The email body is passed to an AI-powered prompt that reads the message to understand its intent and context.
- LLM-Based Role Classification
The AI evaluates language patterns, such as references to budgets, architecture, roadmaps, or strategy, to classify the sender into one of five roles: Leadership Team, Technical Team, Product Engineering Team, Operations, or Finance Team. It works even when the sender’s job title is never mentioned.
- Automatic Contact Record Updates
The classified role is written directly to the Contact record associated with the Account. Sales teams see accurate, current stakeholder roles in their CRM without any manual data entry.
The Outcome
The company turned everyday email conversations into a continuous, automatic source of stakeholder intelligence.
- Always-Accurate Stakeholder Data: Contact roles stay up to date as conversations occur, without relying on reps to update records.
- Smarter Account Engagement: Sales teams can see which leadership, technical, product, operations, and finance contacts are involved in each account, and tailor outreach accordingly.
- More Time Selling: Reps no longer spend time reading emails to guess roles and keying them into Salesforce.
