Using (Off-The-Shelf) AI to Build a Scalable GTM Playbook: Lessons from MedScout

How MedScout built a scalable GTM motion using plug-and-play AI tools

I meet with startups every day and have looked at hundreds of GTM AI tools - it's a crowded space. At Stage 2 Capital, we've invested in a few, like Momentum, which helps enterprise teams turn conversations into insights. But for early-stage startups running lean, the journey often starts with getting creative using basic, readily available tools.

That’s what I loved about MedScout’s approach. As part of our 2023 Catalyst program, they built a solid sales playbook using the Science of Scaling framework. But they didn't want it gathering dust - they used AI to turn it into something their team could actually use day-to-day to work more effectively and efficiently.

I had Brian Aoyama, their Head of Marketing, share how they're implementing AI to do this. Their goal was simple, let AI handle the heavy lifting so their team can focus on what humans do best: building genuine connections and closing deals.

The best part? MedScout didn’t overcomplicate it.

They started with a simple use case, enhancing the call discovery process, and then kept tweaking and improving based on real results. What started as a few practical experiments has turned into a powerful AI system driving their entire GTM motion.

Ok - Let's get into how they did it.

Why MedScout Focused on Discovery First

Discovery is hands down my favorite part of the sales process. It's where great sales teams set themselves apart – playing detective to uncover what's driving their prospects' needs. At Stage 2, we've consistently seen that companies who excel at discovery close larger deals faster and build more predictable revenue engines. Many sales teams rush to showcase their product, thinking that's all it takes to win the deal. But in today's market, that approach falls flat.

Great discovery requires balancing three critical elements:

1. Depth of understanding. This means getting the full picture of your customers' situation – their business objectives, challenges, and what's driving their needs. Your reps need to play detective, asking the right questions and reading between the lines to uncover what matters most.
   

2. Consistency of execution (across calls and reps). It's not enough to have one or two reps who excel at discovery. Every conversation needs to maintain the same high standards, following proven frameworks that surface the insights your team needs to be successful.
   

3. Speed of follow-through. The insights from discovery calls are most valuable when they're fresh. Waiting days to document key learnings or send follow-up materials dramatically reduces their impact.

The challenge is that in a startup with limited resources, these three elements are often in tension. The more thorough your discovery and documentation, the slower your follow-up. This is what led MedScout to embrace AI. Brian explained, "It takes a lot of time and effort to execute our discovery process well. We needed a way to make cycles faster and more reliable without sacrificing quality."

Three Stages to AI-Enhanced GTM

Stage 1 (Crawl): Start with GTM Playbooks

Outcome: First, nail your foundation.

In my conversations with startups, I often see teams excited about the promise of AI and eager to jump straight into implementing AI tools. While the enthusiasm is understandable, this approach typically leads to frustration – overlapping tool capabilities, mounting costs, and low team adoption. Without the right foundation, AI tools can generate a lot of extra "noise" that ends up gumming up your GTM motion rather than improving it.

MedScout took a different path, one that I believe every early-stage company should follow: they invested first in building a strong GTM playbook.

A good playbook doesn't just align your team - it's the foundation for growth. At Stage 2, we've seen that companies who document their GTM process can iterate faster and spot what works. MedScout built their playbook during our Catalyst program and kept refining it, using Winning by Design's SPICED framework - though whether you use SPICED, MEDDIC, Challenger, or another approach, what matters is having a systematic way to understand your prospects and capture key information.

Example of SPICED Framework - Credit: Winning by Design

"You don't want to upend your process to fit off-the-shelf AI tools," Brian explains. "We wanted the process to serve us, not the other way around. So we used our GTM playbooks as a starting point, then layered in AI to execute those playbooks more consistently and effectively than we could before."

When you take the time to document your GTM fundamentals properly, it's much easier to start layering AI into your workflows. Let's walk through how MedScout used AI to eliminate the biggest bottlenecks in their discovery process.

Watch Brian Aoyama talk through MedScout's strategy for creating an AI playbook.

Stage 2 (Walk): Basic GTM AI Implementation

Once you’ve documented your GTM fundamentals, it’s fast and easy to start enhancing your process with AI.

We'll look at two approaches:

1. Quick Start: Simple steps you can implement right now.
   

2. Optimize & Expand: The step-by-step process that MedScout followed to level up their entire discovery process.

Quick Start: Put Your Playbooks to Work

Outcome: Create an AI-powered GTM coach for your team in under 5 minutes.

The quickest way to boost your GTM with AI is straightforward: feed your existing playbook into tools like Claude, ChatGPT, or Perplexity. Just this simple step helps these AI tools understand your company and approach, leading to much better outputs.

"When we first started, it helped to think about our AI tools as new team members – extremely productive but lacking context or direction," Brian explains. "The more context you can give them about your company, objectives, and expectations, the better your results will be."

Claude AI, trained on MedScout’s GTM playbooks, can analyze discovery calls and provide guidance and coaching based on your business and sales framework.

When you take advantage of this capability, you'll see immediate improvements in everything from prospect research to follow-up materials.

Quick tip: If you haven't completed your GTM playbooks yet, use ChatGPT to create a simple one-page document capturing critical information about your company, target customers, and value props. Some tools will even analyze your website to draft this for you. While not comprehensive, this gives you a foundation to build upon.

Optimize & Expand: Build a Custom Discovery Prompt

Outcome: Deliver thorough, personalized follow-up within minutes of every call.

After getting familiar with general-purpose AI tools, the next step is creating specific prompts that strengthen key parts of your GTM process.

MedScout focused on the discovery phase as their biggest opportunity. They built a custom prompt that combined their playbook and discovery frameworks, starting with a basic goal: create better sales call summaries than their automated software.

Want to build better AI prompts? Let AI help. "We used AI to help us build better AI," Brian explains. "We just told it what we wanted to achieve and what information we needed, and asked it to create a template. The key was being specific about our goals, what data to capture, and how we wanted it structured."

Their first implementation was intentionally simple and manual. By using actual discovery calls as their testing ground, the team could evaluate and refine their approach multiple times each day. After each discovery call, reps would:

1. Copy and paste the sales call transcript into Claude or ChatGPT
   

2. Drag and drop the latest version of the shared discovery prompt into the chat
   

3. Review the AI output and provide feedback. Where did it succeed? Where did it fall short?
   

4. Paste the final notes into the CRM

The team revised the prompt every day for 2 weeks. Eventually, the prompt grew to include sophisticated components like:

• Prospect Voice: How to handle prospect quotes: cleaning up filler words while preserving the client’s distinct voice, terminology and key phrases

• Reading Between the Lines: How to make informed inferences based on unstated or implied topics (while clearly marking them with confidence levels)

• Customer First: Improving focus on the prospect's perspective rather than the sales rep's talk track

• Rep Coaching: Scoring the rep's discovery performance and providing coaching feedback

This collaborative process showed two significant benefits beyond just improving the prompt.

1. It built the team's trust in AI-generated outputs. While early results weren't impressive, the rapid improvement cycle helped overcome initial skepticism.
   

2. It strengthened the team's discovery skills. The team became more attuned to what good discovery looks like and how to capture key insights systematically.

Here’s how MedScout structures its prompt to transform its sales playbook and team insights into a powerful AI tool. Use this template as inspiration for your own sales process.

INTRO & ROLE

Open your prompt by stating exactly what outcome you want—such as a SPICED discovery call analysis—and why it matters. This immediate clarity helps the AI target your core needs from the start. Then define the AI’s role (e.g., “You are a senior sales strategist”) so responses match the level of expertise you expect.

COMPANY & CONTEXT

Provide brief details on your product, target customers, and market positioning so the AI can interpret transcript references in the right context. Tying the analysis to your unique value proposition and ICP avoids generic output and ensures recommendations align with what truly sets you apart.

ANALYSIS FRAMEWORK

Introduce your chosen framework—whether SPICED, MEDDPICC, or another—and outline how each element should be examined. Then detail a simple step-by-step process (like extracting key facts, grouping them by SPICED element, and suggesting next steps) so the AI covers every important point without skipping or merging crucial details.

GUIDELINES

Clarify how you want the AI to handle uncertain data, flag missing information, and optionally coach the sales rep. This maintains consistency and trustworthiness: the AI can highlight where the conversation falls short, rate completeness, and provide targeted feedback for improvement.

OUTPUT FORMAT

Finally, specify how the results should look, whether it’s a bullet-point summary or a detailed SPICED breakdown with quotes. If you can share an example of the final format, the AI will know exactly how to structure its response for seamless handoff to your CRM or other team members.

Successful Prompts Aren’t Without Challenges

The custom prompt approach proved highly successful. When tested against specialized products in the market, MedScout's homegrown tool consistently outperformed more expensive alternatives. More importantly, the sales team had fully embraced it – every rep was using the prompt daily and trusted its output.

This success, however, highlighted new challenges. The prompt had grown to 35 pages after months of refinement, making updates increasingly difficult. The manual copy-and-paste workflow, while valuable for early learning, now felt cumbersome. MedScout was ready for their next evolution: automating these proven processes to create a truly AI-native GTM motion.

Stage 3 (Run): Evolving into an AI-native GTM team

Outcome: Build a GTM system that gets smarter with every customer conversation.

Quick sidebar: I love the crawl-walk-run approach. It gives you time to work out the bugs and build trust with your team (and we all know how skeptical sales teams can be). My advice? Don't jump straight to automating workflows until you've earned your team's trust and you're consistently getting high-quality output.

Ready to level up? Here’s how MedScout did it.

Automate AI Workflows

The first thing MedScout did was break its extensive prompt into discrete, specialized components. Then, using accessible tools like Zapier and Clay, they connected these components into automated workflows. And they didn't need expensive specialized software – they used the same tools as before, just connected intelligently.

"Our manual process had served its purpose, helping us refine and validate our approach," Brian explains. "But the repetitive steps were taking time away from actual customer interactions. This friction helped us identify exactly which elements to automate first."

This automation dramatically improved their discovery process. Sales reps could now focus entirely on customer conversations, knowing that detailed call summaries, SPICED framework analysis, and personalized follow-up materials would be ready in moments. This consistent, rapid follow-up accelerated their pipeline velocity.

The Next Iteration: Augmenting the Team with AI Agents

Traditional automation platforms like Zapier require pre-programmed sequential steps, following rigid "if this, then that" logic. But GTM work requires flexibility, which is why MedScout is now exploring AI agents as their next evolution.

AI agents function as digital team members who understand your GTM playbooks, think contextually, and work directly within your tech stack. Unlike fixed automation workflows, agents can determine the best approach based on context and collaborate with team members to achieve objectives.

MedScout is now using Toolflow.ai to build AI agents trained specifically on their specific GTM playbooks. These agents access the team's tech stack to handle sophisticated tasks: researching prospects, analyzing discovery calls through the SPICED framework, enriching CRM data, and creating tailored sales enablement materials.

"We're not looking for agents to replace any functions on our team," Brian explains. "Human judgment remains critical. Instead, we want to augment our team with AI assistants that help execute our playbooks more consistently and at a greater scale than previously possible."

The impact of this AI-native approach is already visible in three areas:

1. Customer Intelligence: A smarter customer data model. Each interaction adds context, helping MedScout engage more effectively throughout the customer journey.
   

2. Cross-Team Learning: Insights flow automatically between teams. Product learns about pain points, Customer Success gets rich context for handoffs, and Marketing sees what resonates in discovery calls.
   

3. Scaling: AI handles routine tasks, freeing the team to focus on work that needs human judgment.

Key Principles for GTM AI Success

What's exciting about MedScout's story is how they built sophisticated AI capabilities starting with basic tools. Through quick iterations, they got there faster than you might think - and this is huge unlock startups.

The market's creating a perfect opportunity: while big companies are stuck wrestling with legacy systems and resistance to change, smaller teams can build AI into their DNA from the start. We're seeing this play out across our Stage 2 portfolio - companies taking a thoughtful approach to AI are actually outmaneuvering much larger competitors.

Five principles that enabled MedScout's success:

1. Start where you can learn fastest. MedScout chose discovery calls because they could test improvements multiple times per day and see immediate impact on deal progression. This rapid feedback loop accelerated their learning and helped them avoid common implementation pitfalls.
   
   

2. Build team trust through visible iteration. By starting with manual processes, MedScout's team could see and shape how their AI tools evolved. This transparency was crucial – by the time they moved to automation, the team trusted the system because they had helped build it.
   

3. Keep humans at the center. AI should handle mechanical tasks like note-taking, framework analysis, and quote extraction, freeing your team to focus on what matters most: building relationships and making strategic decisions. The goal isn't to replace human judgment; it's to amplify it.
   

4. Focus on cross-functional value creation. The real power of AI in GTM comes from building systems that help your entire organization learn and improve. When MedScout automated their discovery process, they didn't just make individual reps more productive – they created a systematic way to capture insights that benefit their product, marketing, and customer success teams.
   

5. Establish a strong foundation before laying in AI. Start with clear GTM playbooks and processes, then layer in AI thoughtfully. This foundation ensures your AI tools enhance rather than complicate your GTM motion, enabling you to iterate quickly toward more sophisticated capabilities.

   
   Good Luck!
   

A huge thank you to Brian Aoyama, head of marketing at Medscout, a Stage 2 Portfolio Company, for sharing his insights and learnings with us. We’re all figuring this out together, and I can’t wait to see what comes next!