Designing an AI Assistant for Smarter Sales Workflows

Design a modular AI platform to empower teams, starting with CGX’s sales team, by automating repetitive tasks, surfacing timely insights, and supporting strategic decision-making.

Client

Clear Grain Exchange

Services

- UX/UI Design - Design System Setup - User Story & Documentation - Product Strategy Support

Industries

Agritech, Supply, Wholesale, Retain SaaS

Overview

Design a modular AI platform to empower teams, starting with CGX’s sales team, by automating repetitive tasks, surfacing timely insights, and supporting strategic decision-making.

Problem

Sales reps at CGX are responsible for managing hundreds of grower relationships. Key challenges included:

  • Manually tracking who to contact and when

  • Following up across inconsistent channels (calls, SMS, email, Slack)

  • Drafting repetitive emails and messages

  • Lack of shared visibility across the team

  • Missed opportunities due to timing, pricing, or offer gaps

Goals

To build an Agentic AI that acts as a central, intelligent assistant, proactive, context-aware, and deeply integrated, helping teams identify opportunities, prioritize follow-ups, and understand market dynamics.

This AI empowers users to:

  • Gain instant visibility into key sales and support conversations

  • Receive smart follow-up recommendations and timely nudges

  • Automatically draft emails, SMS, and call scripts personalized to clients

  • Surface trends from offer activity, grower behavior, and pricing shifts

  • Automate repetitive tasks and route action items to the right tools

  • Collaborate better through unified insights across teams

Solution


We set out to build an Agentic AI, a central, intelligent assistant designed to transform how sales teams work.

Instead of just reacting, this AI acts: listening, learning, and proactively surfacing what matters most.

Whether it’s spotting a sales opportunity hidden in yesterday’s calls, reminding you to follow up with a silent lead, or flagging a pricing trend across grower conversations, the AI becomes your extra pair of eyes, ears, and hands.

The solution enables teams to:

  • Stay ahead of their pipeline with real-time nudges and summaries

  • Draft follow-ups, SMS, and call scripts tailored to client sentiment

  • Organize chaos from Slack, CRM, and calls into a single, actionable feed

  • Detect deal risks, market shifts, and customer intent before it’s too late

  • Reduce manual admin, so reps can focus on building relationships

Collaborate better across sales, support, and marketing with shared AI insights

Key Features

l. Action-oriented widgets

  • Prioritized tasks (Urgent vs Reminder)

  • Embedded context (who to contact, why)


ll. Summarized insights & metrics

  • Daily summary based on Slack logs

  • KPIs: Calls, SMS sent, urgent replies, trades closed, etc.

  • Helps sales team prioritize daily work


lll. Proactive AI Advice

  • Pre-filled email templates based on recent interactions

  • AI-suggested email drafts and contextual advice

  • Personalized content (buyer name, topic, bid, etc.)


lV. Dashboard & navigation

  • Centralize access to assistant-driven tools

    • Activity Log

    • Grower Lookup

    • Drafts & Templates

    • Tasks & Automations

    • Knowledge Assistant

    • Settings & Support

Process

Discovery & Research

I conducted interviews with CGX sales reps and analyzed:

  • Slack message flows

  • Call transcripts

  • CRM usage patterns

  • Email follow-ups

1. Discover
I analyzed real Slack interactions, interviewed users, and tried to identify key pain points like: missed follow-ups, scattered data, and time-consuming comms by analyzing:

  • Slack message flows

  • Call transcripts

  • CRM usage patterns

  • Email follow-ups

2. Define
I mapped user flows, built the information architecture, and prioritized MVP featured, focusing on assistant suggestions, summaries, and Slack integration.

3. Ideate
Early wireframes and card-based UI explored how AI could proactively surface insights and draft follow-ups in a dashboard-first model.

4. Design
I developed high-fidelity mockups, agent creation flows, and an interactive prototype simulating daily workflows.

5. Test
User feedback sessions focused on clarity, usefulness, and trust. I refined the information layout and assistant suggestions based on real sales team input.


Onboarding Flow

New users complete a smart onboarding flow:

  • Define role and goals (e.g., “I want to automate follow-ups”)

  • Connect Slack and email

  • Upload tone of voice docs (emails, brand book)

  • Answer key business context questions

Based on this, personalized agents are activated.

Results

Time saved

  • 2.5 hrs/week saved per rep

  • Measured via time comparisons between manual vs. AI-generated follow-up messages.

Follow-up completion

  • +22% more follow-ups sent within 48h

  • Tracked via Slack/CRM logs pre- and post-assistant usage.

Collected insights

  • 3× more actionable buyer signals identified

  • AI flagged 9 insights/week vs. 3 manually; validated in user feedback.

Draft adoption

  • 68% of messages started from AI drafts

  • Based on in-app analytics of “Use Draft” button engagement.


User trust & Usefulness

  • 86% found the dashboard useful,

  • 73% trusted AI suggestions

  • Results from post-testing surveys.

Faster lead identification

  • 30–50% reduction in time to decide who to contact next

  • Measured during timed usability tests using the AI dashboard vs. Slack review.


Conclusion

The AI assistant helped the sales team stay on top of follow-ups, spot key signals from conversations, and save time on writing and planning. It fit into their existing tools and actually made their work easier. The early results show it's solving real problems, and there’s clear potential to scale it further.