SaaStr Annual 2024: Scaling a Category Defining AI Company from 0 to 100 with Lessons from Cohere, Together AI and Salesforce Ventures

The rapidly evolving AI landscape presents unique challenges for scaling AI-native companies compared to traditional SaaS. This discussion explores the differences in GTM strategies when it comes to scaling startups in this new era of generative AI, highlighting the importance of adaptability, differentiation, and forward-thinking approaches. We'll examine key metrics for evaluating GTM success, strategies for achieving growth through sales and marketing, and the critical role of category creation and positioning in the competitive AI market.

1. Rajan Sheth CMO - Together.ai, Cohere GP - Hypergrowth Partners Scaling a Category Defining AI Company from 0 to 100 with Lessons from Cohere, Together AI and Salesforce Ventures Emily Zhao Principal Salesforce Ventures

2. How AI is a different ball game and why you should care The “Scaling Laws” of B2B Software Gen AI B2B SaaS 3x → 3x → 2x → 2x or “Triple, triple, double, double” $1M ARR $100M ARR $100M ARR $100M ARR Traditional B2B SaaS 5-10x → 5x? → 3x?

3. Typical B2B SaaS cycle Product PMF Growth and Scaling Maturity & Expansion Company life cycle B2B Saas

4. ..doesn’t work for Gen AI Product PMF Growth and Scaling Maturity & Expansion Company life cycle … Technological Innovation New market entrants Gen AI B2B Saas

5. Part 1: The Changing Dynamics of Selling Generative AI

6. Multi-Front Battlefields in AI Cause Impact Examples Foundational models enabling platform shift in software Nascent market with fast moving underlying AI technologies Customers are eager to adopt AI but hesitant to make long term bets Cambrian explosion of startups in every category Exacerbated by AI driven productivity in software engineering + back office functionalities, etc. New startup ideas quickly become obsolete While other ideas previously impossible become possible Signing only short term contracts or small ACV contracts Using multiple vendors at once 1st year contract is effectively a “POV” Emergence of “Experimental Revenue” AI for the most niche use case (e.g., smart voice journal, auto-generate B roll videos for ads) OpenAI wrappers (e.g., summarize your emails) AI generated new molecules AI testing & monitoring AI security & governance

7. How many customers are through VC intros? Signs of “Experimental Revenue” Retention Usage / Use Case Sales Motion / Sales Metrics Others? High NDR but low GDR Continued high ARR / logo churn Shortened contract length, not truely annualized 1 2 3 4 Usage metrics fluctuate heavily High experimental / low production use case Smaller ACV vs. ability or willingness to pay Using multiple vendors at once for different use cases Unpredictable sales rep performance Inconsistent logo velocity Lumpy ARR growth Sustainable source of budget? How many customers are deployed / live? Customers skew towards high tech / startups

8. Some real life examples to dive in

9. Deep Dive #1: Retention (I/II) Example 1: AI Infrastructure Startup Product A Product B

10. Deep Dive #1: Retention (II/II) Other things to consider: cohort level performance, how many vendors are decided through competitive RFPs (for both new and expansion) Example 2: Legal Tech Startup

11. Deep Dive #2: Usage & Use Case (I/II) Example 1: AI Image/Video Generation New Feature Release New Model Release New Model Release New Feature Release

12. Example 2: AI Testing/Eval Startup Deep Dive #2: Usage & Use Case (II/II) Use Case Breakdown ACV by Customer Type

13. Deep Dive #3: Sales Motion (I/II) Example 1: AI Customer Support

14. Deep Dive #3: Sales Motion (II/II) 35% 70% 0% 0% 0% 91% 100% 0% Example 2: AI Testing/Eval Startup Attainment %

15. What does this mean?

16. Manufacturing Inventory mgmt, supply chain visibility New Platform Shift = New Moats to Acquire Real “ARR” Product / feature shipping velocity to adapt to changing customer demand & changing technology Capturing unique data that can create network effects or synergies that others are not positioned to capture Capturing industry specific or persona specific workflows super well that the product becomes entrenched in the user’s day-to-day MLOps vs. AIOps/LLMOps Potential differentiations when building your products (when you are not doing FM research)

17. GTM-Led Differentiation Network effects (give examples) Example: Hugging Face, Langchain User community / open source community (give examples) Example: Hugging Face, Protect AI Distribution model: inbound, outbound, PLG, channel partnership) Example: Anthropic, Cohere Marketing motion: branding and story Example: Runway Customer experience, customer love, Example: [TBD] Demonstrating time to value, ROI Example: Palantir - Ed Sim note from Aug 10th, Neuron7 Solving last mile problem Network Effects User community / open source community Marketing motion: branding & story Potential differentiations when building your GTM Distribution model: inbound, outbound, PLG, hybrid, etc. Customer experience, customer love Demonstrating time to value, ROI

18. Part 2: Differentiated GTM to Scale your AI Company from 0 to $100M

19. Heard someone mention “agents” this week?

20. Every company is becoming an agent company Source: CBInsights

21. The hyper-competitive AI market calls for differentiation Source: Sapphire Ventures

22. This is true for every part of the Gen AI stack Source: Maxime Labonne

23. So, how do you win this market?

24. Create a “movement”

25. Create a “movement” behind your vision and how you differentiate in the market AI safety and research company AI platform for enterprise

26. Change market perception and punch above your class

27. 2. Build credibility beyond the product

28. It's no longer enough to simply promote "what's on the truck today" Pre-announced our enterprise chat vision months before the product Became the biggest lead gen driver from enterprises we were targeting

29. Build a brand around your CEO & superstars

30. 3. Simultaneous GTM motions

31. Cannot wait for sequential GTM motions like traditional SaaS Enterprise Awareness “Movement” Bottoms Up PLG Motion Tops Down Sales Motion Partner and Channel Motion Developer and Research “Credibility”

32. 4. Unusual growth levers with strategic marketing

33. Shape the industry conversations, rather than just reacting to it. Your target audience is broader than your prospect list and they care about more than your product

34. Making long term bets on market evolution that will drive a sustainable long-term advantage “LLM University” — an initiative to upskill developers on the technology, even before the ChatGPT hype stemmed, from the conviction about the skilled resource gap in the AI market. Generated 40% of all Cohere traffic

35. Key Takeaways

36. The expectation for what is considered a good growth profile is different today vs. the previous generation of B2B software → it’s a different ball game A big part of it is due to the effects of “ experimental revenue ” and you can tell if your business is seeing signs of this from a few different aspects -- retention, usage/use case, sales motion / sales metrics, and other qualitative signals Founders in this new era of generative AI need to think of new ways to build moats / differentiation , for both product as well as GTM, but especially GTM Next generation of successful AI companies will be marketing-led as the product, technology and the market are in a constant flux Role of the CMO becomes crucial to drive the company vision and place a few long-term bets on how the market will evolve and what will drive a sustainable long-term advantage

37. THANK YOU!