#32- ImAIgine: GRW.AI Ceo Interview, Klarna axes Salesforce and Monday for AI, AI Research paper, Scaling the State of Play in AI, and Runway.

The newsletter and podcast are sponsored by GTM AI Academy AND AI Powered GTM.

Business Impact> Learning Tools

1. We have LAUNCHED the FREE Slack Community, come hang with your fellow GTM professionals, get insights, and content free to help in your teams and roles.
2. The GTM AI Demo Tools Library is a free resource for you to view AI tools and demos to see what is out there to help your team or to help you in your own workflow

Now with all that being said, lets move forward with todays newsletter which is:

1. We have #32 Podcast episode where I had the pleasure of interviewing Alex McNaughten the CEO and CoFounder of Grw AI
2. Klarna shaking up the industry with cutting Salesforce and Monday and replacing with AI.
3. Research Paper, "Let's Verify step by step" on how to improve the reliability and performance of large language models.
4. Scaling: The State of Play in AI by Ethan Mollick
5. AI Tool of the week: Runway

Some AI posts from this last week in case you missed it:

Advice from OpenAI on prompting o1-preview

Using 4o vs o1 on Competency model building

Asking AI what did it know about me with crazy results

Now with that, lets rolllllllllll

You can go to Youtube, Apple, Spotify, or here on Linkedin as well as a whole other host of locations to hear the podcast or see the video interview.

"AI-Powered Sales Coaching: How GRW AI is Revolutionizing Revenue Teams"

I recently sat down with Alex McNaughten the CEO and Cofounder of Grw AI to discuss how his company is reshaping sales enablement through AI-powered coaching. Alex walked me through his career path, starting as a guest columnist and evolving into a tech sales leader before founding GRW AI.

We dove deep into the problems plaguing sales teams today: minimal coaching, missed quotas, and high turnover rates. Alex introduced me to Taylor, GRW AI's AI-powered deal coach, explaining how it provides tailored, context-aware guidance to salespeople. He talked to me about how Taylor works with CRM systems, adapts to company-specific sales methods, and offers real-time advice throughout the sales process. We also covered the tech behind GRW AI, including its use of multiple large language models and ability to incorporate new AI advancements quickly.

Alex shared a specific customer success story, where a company achieved record-breaking results even after losing their sales manager, thanks to GRW AI. We addressed common concerns about AI trust and hallucination, with Alex explaining their approach to training the AI on company-specific playbooks for accuracy.

Our conversation then shifted to the future of sales enablement, predicting a move towards more targeted, data-driven strategies enabled by AI. We explored AI's potential to enhance rather than replace salespeople, emphasizing that human relationships remain crucial in sales. We wrapped up by discussing how AI might reshape sales processes, with Alex suggesting that while core buying and selling dynamics may stay the same, AI will greatly improve efficiency and preparation in sales activities.

Highlights

- GRW AI addresses the "sales execution gap" with an AI-powered deal coach named Taylor

- Taylor provides personalized, context-aware guidance throughout the sales process

- GRW AI integrates with CRM systems and is trained on company-specific sales methodologies

- The platform has helped customers achieve record-breaking sales results

- AI in sales enablement is predicted to focus on performance improvement and data-driven approaches

- The future of sales is seen as AI-augmented rather than AI-replaced

- AI is expected to increase the time salespeople spend with customers by reducing administrative tasks

5 Key Quotes from Alex McNaughten :

1. "Imagine if every single one of your salespeople in your organization had a deal coach available to them that had intimate knowledge of who you are, what you sell, who you sell to and your way of selling."

2. "Taylor played the part of expert manager, but the difference is Taylor can be in 10 places at once, 100 places at once and a manager can't."

3. "I think AI can, in our case anyway, certainly help sales enablement allocate time, effort, and resource towards the right things. And then also attribute that what they're doing is working."

4. "I see a world where being in sales, you're spending more time focused on the customer and less time doing the admin and that side of things."

5. "AI augmented selling is to me... like you put on the Ironman suit and I'm able to do things that I couldn't, like I'm faster, I'm stronger, I'm smarter. I'm able to do things I couldn't do before, or I'm able to do things 10, a hundred times faster and better than I could do them before."

Take a listen and tell me how you think of it!

@klarna the Swedish fintech giant, is making a bold and potentially industry-shaking move by severing ties with enterprise software stalwarts Salesforce and Workday in favor of AI-powered, internally developed solutions. This decision is part of a larger strategy to leverage artificial intelligence across their operations, aiming to dramatically increase efficiency and reduce costs. The company has already implemented an AI assistant for customer service that reportedly performs the work of 700 human agents, and they're now setting their sights on replacing core business systems.

This shift is not just about technology replacement – it represents a fundamental rethinking of how large companies can operate in the AI era. By building their own AI-powered applications, Klarna is betting that they can achieve greater flexibility, cost-efficiency, and competitive advantage compared to using off-the-shelf enterprise solutions. This move aligns with their broader cost-cutting measures, including significant workforce reductions, as they prepare for a potential IPO.

However, this strategy is not without its skeptics. Industry experts question whether Klarna can effectively replicate the complex functionalities of established enterprise systems, particularly in areas like HR and payroll management. There's also debate about whether the resources required to build and maintain these in-house systems will truly result in long-term cost savings and efficiency gains.

For GTM teams and leaders across industries, Klarna's initiative serves as a provocative case study in the potential of AI to reshape not just individual processes, but entire organizational structures and technology stacks. It raises important questions about the future of enterprise software, the changing nature of workforce skills and composition, and the potential for AI to drive significant competitive advantages in go-to-market strategies.

As this situation unfolds, it will be crucial for business leaders to closely monitor the outcomes. If successful, Klarna's approach could inspire a wave of similar initiatives across the corporate world, potentially disrupting the enterprise software market and forcing a reevaluation of how companies build and maintain their core business systems. On the other hand, if Klarna struggles to replicate the functionality of established enterprise solutions, it could serve as a cautionary tale about the limits of AI and the risks of abandoning proven technologies too quickly.

What's Happening?

1. Ditching Enterprise Giants: Klarna has stopped using Salesforce for sales and marketing data management and Workday for HR and hiring processes.
2. AI-Powered Replacements: The company plans to replace these systems with internally built applications, likely leveraging OpenAI's infrastructure.
3. Customer Service Automation: Klarna has already implemented an AI assistant that reportedly does the work of 700 customer service agents, handling 2.3 million interactions in its first month.
4. Workforce Reduction: The company has cut 1,200 jobs over the past year and hints at potentially reducing its global workforce from 3,800 to 2,000 employees.
5. OpenAI Partnership: Klarna was one of the first enterprise clients for OpenAI's ChatGPT, with 90% of its workforce using the tool daily to automate various processes.

Why This Matters to GTM Teams and Leaders

1. Potential Disruption of Enterprise Software Market: If successful, Klarna's move could inspire other companies to follow suit, potentially disrupting the enterprise software market dominated by giants like Salesforce and Workday.
2. AI-Driven Efficiency: GTM teams may need to reevaluate their tech stacks and consider how AI can streamline operations and reduce reliance on multiple SaaS providers.
3. Changing Workforce Dynamics: The potential for AI to replace not just customer service roles but also functions traditionally handled by enterprise software could lead to significant changes in team structures and skill requirements.
4. Data Management and Integration Challenges: Teams will need to consider how to effectively manage and integrate data across AI-powered systems, especially if moving away from established enterprise solutions.
5. Cost-Benefit Analysis: GTM leaders should carefully assess the long-term costs and benefits of building in-house AI solutions versus using established SaaS providers.
6. Skill Set Evolution: There may be a growing need for team members who can effectively work with and optimize AI systems, rather than just operate within traditional SaaS environments.
7. Customer Experience Impact: GTM strategies may need to evolve to address how AI-driven automation affects customer interactions and overall experience.
8. Competitive Advantage: Early adopters of AI-driven solutions may gain a significant edge in efficiency and cost-saving, potentially allowing for more aggressive market strategies.
9. PR and Marketing Opportunities: Klarna's move demonstrates how embracing AI can generate significant media attention and position a company as innovative, which could be a valuable PR strategy for GTM teams.
10. Vendor Relationships: GTM teams may need to reevaluate their relationships with software vendors and consider how to negotiate contracts in an environment where AI alternatives are becoming more viable.

While Klarna's strategy is bold and attention-grabbing, it's important to note that there's skepticism in the tech community about the feasibility of fully replacing complex enterprise systems with in-house AI solutions. HR technology analyst Josh Bersin points out the challenges of replicating decades of workflows and complex data structures built into systems like Workday.

GTM teams and leaders should closely monitor Klarna's progress and results. This move could signal a significant shift in how companies approach their technology infrastructure, potentially leading to new opportunities and challenges in go-to-market strategies. However, it's crucial to approach such drastic changes with caution, carefully weighing the risks and rewards before making similar moves.

"Let's Verify Step by Step: Improving Mathematical Reasoning through Process Supervision"

"Let's Verify Step by Step" is a groundbreaking research paper authored by a team of researchers from OpenAI, including Hunter Lightman, Vineet Kosaraju, Yura Burda, and several others. Published in 2023, this study tackles a critical challenge in artificial intelligence: improving the reliability of large language models in complex reasoning tasks, particularly in mathematics. The researchers were motivated by the fact that even state-of-the-art AI models frequently make logical mistakes when solving multi-step problems. To address this, they compared two methods of training AI systems: outcome supervision, which only judges the final answer, and process supervision, which provides feedback on each step of the problem-solving process. Their work aimed to determine which method produces more reliable and accurate AI models, with potential implications for AI safety and alignment. The study is particularly significant as it pushes the boundaries of AI capabilities in mathematical reasoning and offers insights that could be applied to improve AI performance across various complex tasks.

1. The authors compare two methods of training reward models for language models: outcome supervision and process supervision.

2. Process supervision, which provides feedback on each step of problem-solving, is found to significantly outperform outcome supervision, which only judges the final answer.

3. The researchers created a large dataset called PRM800K, containing 800,000 step-level human feedback labels for mathematical problem-solving.

4. They demonstrate that their process-supervised model can solve 78% of problems from a subset of the challenging MATH dataset.

5. The paper also discusses the benefits of active learning in improving data efficiency for process supervision.

6. The authors argue that process supervision has advantages for AI alignment, as it encourages models to follow human-endorsed reasoning processes.

7. The research includes experiments with both large-scale and small-scale models, and examines out-of-distribution generalization.

For GTM Professionals:

1. Step-by-step reasoning: Encourage the AI to break down complex problems into smaller steps. This approach, similar to the "chain-of-thought" mentioned in the paper, can lead to more accurate and reliable outputs.

2. Verification prompts: After getting an initial response, ask the AI to verify its reasoning or check for errors in its process. This mimics the "process supervision" concept from the paper.

3. Multiple attempts: The paper shows that considering multiple solutions (best-of-N) improves performance. You could ask the AI to generate multiple approaches to a problem and then synthesize or choose the best one.

4. Explicit instructions: Provide clear, detailed instructions in your prompts. The paper emphasizes the importance of guiding the model's reasoning process.

5. Active learning: While you can't directly apply this, the concept suggests that focusing on challenging or ambiguous cases can lead to better results. Consider refining your prompts based on where the AI struggles.

6. Domain-specific knowledge: The paper uses math-specific pretraining. For Go-to-Market tasks, consider providing relevant industry or market context in your prompts.

7. Output verification: Ask the AI to explain its confidence in its answers or to highlight any assumptions it made. This can help identify potential errors or biases.

8. Iterative refinement: Break complex tasks into stages, reviewing and refining the AI's output at each stage before moving to the next.

9. Combine AI and human insight: The paper shows the value of human feedback. Consider using AI outputs as a starting point, then applying your professional judgment to refine and improve the results.

10. Awareness of limitations: The paper discusses model hallucinations. Remember that while AI can be a powerful tool, it's not infallible, especially in highly specialized or rapidly changing fields.

By applying these concepts, Go-to-Market professionals can potentially improve the quality and reliability of the AI-generated insights and strategies they use in their work.

A recent article by Ethan Mollick dives into the current state and future trajectory of AI, focusing on Large Language Models (LLMs) that power chatbots like ChatGPT and Gemini. It explains how model size and computing power drive AI capabilities, introducing a generational framework for understanding model progression. The piece covers the leading "Gen2" models, their unique features, and introduces a new scaling approach based on inference compute or "thinking time." This dual scaling in both training and inference suggests significant AI advancements are on the horizon, with implications for various sectors.

Key highlights:

1. AI capability is largely driven by model size, which follows a generational approach requiring 10x increases in data and computing power.

2. The author proposes a simplified generational framework for frontier models:

- Gen1 (2022): ChatGPT-3.5 level, <10^25 FLOPs, ~$10M training cost

- Gen2 (2023-2024): GPT-4 level, 10^25-10^26 FLOPs, ~$100M+ training cost

- Gen3 (2025?-2026?): Upcoming, 10^26-10^27 FLOPs, ~$1B+ training cost

- Gen4 and beyond: Expected in coming years, potentially $10B+ training cost

3. Five leading Gen2 models are discussed: GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro, Grok 2, and Llama 3.1 405B.

4. A new scaling approach based on inference compute or "thinking time" has been introduced by OpenAI's o1 models, offering another avenue for AI improvement.

5. The combination of training scale and inference scale suggests significant AI capability improvements in the coming years.

Why this matters to GTM professionals:

1. Understanding AI capabilities: GTM pros need to grasp current and future AI capabilities to effectively position products and services in an AI-enhanced market.

2. Product differentiation: Knowledge of different AI models and their strengths helps in highlighting unique selling points of AI-powered products or services.

3. Market timing: The generational framework provides insights into when more advanced AI capabilities might become available, informing product roadmaps and market entry strategies.

4. Resource allocation: Understanding the costs and computing power required for different AI generations helps in budgeting and planning for AI integration or development.

5. Competitive landscape: Awareness of leading models and their creators (e.g., OpenAI, Google, Meta) provides context for the competitive environment in AI-related markets.

6. Future-proofing strategies: Insights into AI's trajectory allow GTM professionals to anticipate market shifts and adapt strategies accordingly.

7. Customer education: GTM pros can use this knowledge to educate customers about AI capabilities, addressing both opportunities and potential concerns.

8. Use case development: Understanding model capabilities helps in identifying and developing compelling use cases for AI-enhanced products or services.

9. Pricing strategies: Knowledge of AI model costs and capabilities can inform pricing strategies for AI-powered offerings.

10. Partnership opportunities: Insights into different models and their strengths can guide decisions on potential AI partnerships or integrations.

By staying informed about AI advancements, GTM professionals can better navigate the rapidly evolving tech landscape, identify new opportunities, and effectively communicate the value of AI-enhanced solutions to their target markets.

GTM AI Tool of the week, RunwayML: The Future of Creative AI

RunwayML is an AI-powered creative suite designed to empower creators with advanced tools for video editing, image generation, and more. With features like real-time video editing, generative image models, and customizable AI workflows, it’s a powerful tool for filmmakers, designers, and artists. It allows users to experiment with cutting-edge AI tools, providing unparalleled flexibility and creative potential.

Why GTM Professionals Should Pay Attention

For GTM professionals, RunwayML opens new avenues for marketing, content creation, and brand storytelling. Visual content plays a crucial role in capturing audience attention and building brand identity. RunwayML enables teams to produce high-quality, unique visual content at scale, leveraging AI for tasks such as video editing, image generation, and content creation, all while reducing manual effort.

RunwayML’s suite is particularly useful for teams that rely heavily on visual marketing and digital media. It allows rapid iteration, seamless experimentation, and highly customized visuals tailored to different campaigns, audiences, and platforms.

Practical Applications of RunwayML for GTM Professionals

1. Automated Video and Image Production: Use AI to automate video editing and image creation, producing professional-grade visual content at scale. RunwayML can help brands create highly engaging marketing materials, from video ads to social media posts, without the typical time investment.

2. Brand-Specific Content: RunwayML allows marketing teams to generate branded assets tailored to specific audiences or campaigns, ensuring that the visual messaging remains consistent and on point.

3. Real-Time Editing and Iteration: Teams can leverage RunwayML’s real-time tools to quickly iterate on creative concepts, enabling faster decision-making and adaptation to market trends.

How Different Teams Can Use RunwayML

1. Sales

- Use Case: Sales teams can create engaging presentations and multimedia content that capture attention. RunwayML can quickly generate promotional videos or product demos that sales reps can present to prospects.

- In Depth: For example, sales teams can use AI-generated visuals in product showcases, helping bring complex features to life and facilitating clearer communication during pitches.

2. Customer Success

- Use Case: Customer Success teams can leverage RunwayML to create instructional videos, tutorials, and FAQs, enhancing customer onboarding and providing effective self-service resources.

- In Depth: AI-generated visuals can help explain product functionalities or solutions to common issues, making the customer experience smoother and reducing support tickets.

3. Marketing

- Use Case: Marketing teams can use RunwayML to generate custom video ads, social media content, and landing page visuals, all optimized for specific campaigns and audience segments.

- In Depth: RunwayML’s ability to create and edit videos in real-time allows marketers to experiment with different creative strategies quickly, helping them discover what resonates most with their audience.

4. Enablement

- Use Case: Enablement teams can produce training videos, role-play simulations, and visual aids that help upskill employees quickly and efficiently.

- In Depth: Teams can create tailored training materials for different departments, ensuring that employees receive engaging, relevant content that supports continuous learning and skill development.

5. Business Development

- Use Case: Business Development teams can use AI-generated visual content to craft compelling partnership presentations or pitch decks that stand out in competitive markets.

- In Depth: RunwayML’s ability to produce stunning visuals enables business development teams to highlight key value propositions, making their presentations more persuasive and visually impactful.

6. HR

- Use Case: HR teams can use RunwayML to create engaging recruitment videos and internal communication materials, helping build a strong employer brand and improve employee engagement.

- In Depth: With AI-generated content, HR can produce high-quality videos that promote company culture or explain policies, ensuring employees are aligned with the company’s goals and values.

RunwayML is a game-changer for creative teams across sales, marketing, customer success, and business development. Its AI-driven tools streamline the production of visually captivating content, allowing GTM professionals to focus on strategic initiatives rather than manual design work. By integrating RunwayML into your creative workflow, you can ensure that your brand remains innovative, engaging, and responsive to the needs of your audience. Whether for visual storytelling, brand engagement, or training, RunwayML provides the tools needed to stay ahead in today’s content-driven market.

Let me know what stands out to you this week!