Drove a 45% increase in conversions with AI sales agents

OLIV AI

Oliv AI is a sales tool that helps sales teams close deals faster by automating the tiresome tasks that slow them down, such as researching potential customers, tracking leads, and providing data-driven coaching.

Oliv AI is a sales tool that helps sales teams close deals faster by automating the tiresome tasks that slow them down, such as researching potential customers, tracking leads, and providing data-driven coaching.

Oliv AI is a sales tool that helps sales teams close deals faster by automating the tiresome tasks that slow them down, such as researching potential customers, tracking leads, and providing data-driven coaching.

This case study details how I contributed to Oliv's vision by defining workflows for its AI agents to assist Sales Managers and Sales Reps in their day-to-day work.

Image: AI reports for sales team to track every deal in the pipeline

Impact

53% increase in sales demo bookings.

~45% increase in successful signups within the first two weeks of launch.

Secured a competitive advantage by positioning Oliv as an early innovator in AI for sales.

Users & their needs

SALES MANAGER

Pain: Manually review hours of sales calls to track deal progress and forecast sales.

Consequence: Key deals slip through the cracks due to a lack of real-time visibility into the pipeline.

Need: An intelligence platform for objective data on deal health and coaching.

SALES REPS

Pain: Losing selling time to manual work like research and post-call deal updates.

Consequence: Inconsistent and delayed follow-ups, leading directly to lost deals and a leaky sales funnel.

Need: An AI assistant to automate tasks and help them focus on the sale.

Team & my role

As the founding product designer working with a small team, I worked closely with co-founders, engineers, product manager, and product marketer.

Defined AI agent tasks and their integration into user workflows.

Designed user journeys and AI interactions.

Partnered with product marketing on the go-to-market strategy and narrative.

Constraints & roadblocks

Limited budget: As a seed-funded startup, all projects required strict prioritization.

Aggressive timeline: A pressing need to launch quickly to secure a first-mover advantage.

User trust barrier: Overcoming user hesitation to grant AI agents access to sensitive sales data.

Key decisions

Create dedicated AI agents for key sales functions: deal tracking, forecasting, coaching, and task automation.

Prioritize seamless integration of the AI agents into the team's existing daily tools like Gmail, Zoom, and Slack.

Strategically reposition the product from a simple meeting recorder to an "Intelligent Deal Hub".

Design solution

Our journey to launch AI agents for Sales was a 3 phased evolution.

Our journey to launch AI agents for Sales was a 3 phased evolution.

Our journey to launch AI agents for Sales was a 3 phased evolution.

PHASE 1

Realizing we solved the wrong problem

To understand why Oliv wasn't becoming a daily-use tool for our users, we began by analyzing our churn data.

"I'm still blind. Oliv gives me another recording to watch, but I still have to dig through the entire call to figure out why a deal is stalled."

- Sales Manager

"We're already using Gong. It has its own issues, but asking us to switch to another meeting recorder is a huge hassle for little additional value."

- Account Executive

"I tried the prospect research feature, but the data quality just isn't there. I have to double-check everything anyway."

- Sales Rep

"Oliv doesn't touch real time-sucking manual work; updating the CRM and deal tracking, so it doesn't really change my day."

- Sales Rep

Image: Few feedbacks from lost deals and users that helped to see the bigger picture.

My key contribution here was pushing the conversation beyond the surface level. In brainstorms, I helped the team realize that even a "perfect" meeting tool would fail if it didn't help a sales rep with their ultimate goal: moving a deal forward.

Image: An overall comparison I called out during an ideation session. It contrasts the old model (meeting focus) with the new model (deal focus), illustrating our strategic shift.

PHASE 2

The new strategy – AI agents for sales

With our new focus set, I partnered with leadership to design the Agentic Framework, a blueprint for how our AI agents collaborate.

Image: This flow map shows the Agentic Framework in action, how our AI agents collaborate to automate workflows and deliver critical sales insights.

For sales reps

Vision

We aimed to automate the reps' most tedious administrative tasks, they would see Oliv as an indispensable assistant rather than a burden, leading to consistent daily use.

Action

We designed dedicated agents for sales reps:

Meeting Assistant: To help in preparing for meetings and running them hands free and automate post meeting tasks.

Researcher: To gather vital company and contacts insights, fill in gaps and analyze any source material to create account maps in 15 minutes.

Deal Assistant: To automatically tracks deal data, loop sales managers in at key moments.

CRM Manager: To maintain accurate CRM data and keep evolving with the sales process through real-time feedback from sales reps.

Image: Breakdown of specific tasks for each AI agent for sales reps.

To help the Engineering team write better prompts, I provided variants for each example message, ensuring they would deliver quality insights to users.

Image: A few automated messages from AI agents such as pre-call research and post-call summaries, that are delivered directly to reps in Slack.

Reps get an instant health check on any deal with AI-powered insights and recommended next actions.

Team can see how qualified the deal is with an AI-generated MEDDIC scorecard and summary.

Reps can understand who's who in the deal, mapping out every contact and their role in the sales process.

Images: These are the 3 key MVP screens designed for our sales demos to sales reps.

For sales reps

Vision

We aimed to automate the reps' most tedious administrative tasks, they would see Oliv as an indispensable assistant rather than a burden, leading to consistent daily use.

Action

We designed dedicated agents for sales reps:

Meeting Assistant: To help in preparing for meetings and running them hands free and automate post meeting tasks.

Researcher: To gather vital company and contacts insights, fill in gaps and analyze any source material to create account maps in 15 minutes.

Deal Assistant: To automatically tracks deal data, loop sales managers in at key moments.

CRM Manager: To maintain accurate CRM data and keep evolving with the sales process through real-time feedback from sales reps.

Image: Breakdown of specific tasks for each AI agent for sales reps.

To help the Engineering team write better prompts, I provided variants for each example message, ensuring they would deliver quality insights to users.

Image: A few automated messages from AI agents such as pre-call research and post-call summaries, that are delivered directly to reps in Slack.

Reps get an instant health check on any deal with AI-powered insights and recommended next actions.

Team can see how qualified the deal is with an AI-generated MEDDIC scorecard and summary.

Reps can understand who's who in the deal, mapping out every contact and their role in the sales process.

Images: These are the 3 key MVP screens designed for our sales demos to sales reps.

For sales reps

Vision

We aimed to automate the reps' most tedious administrative tasks, they would see Oliv as an indispensable assistant rather than a burden, leading to consistent daily use.

Action

We designed dedicated agents for sales reps:

Meeting Assistant: To help in preparing for meetings and running them hands free and automate post meeting tasks.

Researcher: To gather vital company and contacts insights, fill in gaps and analyze any source material to create account maps in 15 minutes.

Deal Assistant: To automatically tracks deal data, loop sales managers in at key moments.

CRM Manager: To maintain accurate CRM data and keep evolving with the sales process through real-time feedback from sales reps.

Image: Breakdown of specific tasks for each AI agent for sales reps.

To help the Engineering team write better prompts, I provided variants for each example message, ensuring they would deliver quality insights to users.

Image: A few automated messages from AI agents such as pre-call research and post-call summaries, that are delivered directly to reps in Slack.

Reps get an instant health check on any deal with AI-powered insights and recommended next actions.

Team can see how qualified the deal is with an AI-generated MEDDIC scorecard and summary.

Reps can understand who's who in the deal, mapping out every contact and their role in the sales process.

Images: These are the 3 key MVP screens designed for our sales demos to sales reps.

For sales managers
For sales managers
For sales managers

PHASE 3

Go-to-market – Shaping the narrative

To launch AI agents, we needed a compelling story. I helped shape the narrative for our launch videos and created low-fidelity storyboards to guide the animation team. This ensured the final videos reflected the true user experience we aimed to build.

Video: It presents the Oliv as a personal assistant for sales reps to automate notes and eliminate admin work.

Video: It presents Oliv as an intelligence hub for sales managers to coach, forecast, and enforce their sales playbook

Outcome

Doubled the sales close rate for key customers, with one team reporting an increase from 25% to 50%.

Drove significant market share growth and solidified Oliv's position as a key innovator in the competitive sales tech market.

Streamlined post-call workflows leading to "less deal tracking chaos" and more time for "real, human connection".

Learnings & growth

This project was a period of immense growth, teaching me three key lessons:

The most impactful design work happens when you move beyond the UI to diagnose and solve the core business problem.

To succeed in B2B, you must solve a critical, expensive problem for the person with purchasing power (the manager), by creating a tool their team (the reps) loves to use.

Trust in AI is not a given; it must be consciously designed through transparency, clear communication, and delivering on promises.

Explore more case studies

Create a free website with Framer, the website builder loved by startups, designers and agencies.