← Back to Case StudiesCase Study: AI Sales Assistant

CRM Agent

Liberated high-value employees to focus on revenue-generating conversions.

80%

Manual Tasks Cut

24/7

Agent Uptime

Background

UNLMTD Digital Agency is a growing digital agency with an ambitious sales team juggling multiple tools daily. Their salespeople were constantly switching between HubSpot for CRM data, Google Sheets for lead lists, Google Calendar for scheduling, and various databases for prospect research. The fragmented toolstack created friction. Finding a contact's history meant opening HubSpot. Checking availability meant switching to Calendar. Researching a prospect meant digging through spreadsheets and databases. Every task required context switching, and context switching kills productivity. UNLMTD needed a way to bring all these tools together into a single, conversational interface their sales team already used every day: Slack.

The Challenge

The sales team faced a common but costly problem: death by a thousand tabs. Their daily workflow required searching HubSpot, managing lead lists, scheduling meetings, researching prospects, tracking activities, and accessing company knowledge—all in different applications. Each task meant opening a different app, remembering a different interface, and losing focus.

The core question: How do you give a sales team instant access to everything they need without forcing them to become experts in five different platforms?

The Solution

We built "Elliot Smith"—an AI-powered sales assistant that lives in Slack and acts as a single point of access to UNLMTD's entire sales infrastructure. This system is built around a central AI agent with persistent memory, connected to specialized tools and sub-agents for CRM, calendar, data access, and knowledge retrieval.

How It Works:

A salesperson sends a message in Slack: "Hey Elliot, what do we know about Acme Corp? And can you check if Sarah has any time tomorrow afternoon for a call?"

  1. The AI agent queries HubSpot for Acme Corp's history.
  2. It searches lead databases for additional intelligence.
  3. It checks Sarah's Google Calendar for availability.
  4. It combines all this information into a single response and sends it back to Slack within seconds.

Impact

The AI assistant eliminated the need for constant context switching, unifying the sales operations into a single conversational interface.

Unified Sales Operations

CapabilityBeforeAfter
Contact lookupOpen HubSpot, search, navigate"Elliot, what do we know about [company]?"
Calendar checkOpen Calendar, find person, scan times"Is [person] free tomorrow?"
Lead researchMultiple spreadsheets and databases"Show me leads in [industry]"
Meeting schedulingBack-and-forth across tools"Schedule a call with [contact] for [time]"
Process questionsSearch docs, ask colleagues"What's our pricing for [service]?"

What Makes This Different

Conversational, Not Command-Based: Reps talk to it like a colleague. Persistent Memory: It remembers past conversations for context. Extensible Architecture: New capabilities can be added easily. Grounded in Real Data: RAG integration ensures answers come from company documentation.

Next Steps

The foundation is built. The roadmap includes:

  • Enhanced reporting on agent usage to identify knowledge gaps.
  • Proactive notifications for at-risk deals or new ICP-fit leads.
  • Deeper CRM automation for deal stage management.

Next Steps

This implementation validated the approach. The roadmap for optimization includes:

  • Script Refinement: Analyzing successful call patterns to tune the AI agent's responses.
  • Time-of-Day Testing: Running calls at different times to identify optimal windows for each region.
  • A/B Testing: Experimenting with different AI agent personas and opening lines to improve answer rates.