Podcast Image: Discussing Mystic AI with Max Tee and Chris Wang

Discussing Mystic AI with Max Tee and Chris Wang

Demystifying Mystic AI: How a Startup is Navigating the B2B AI Infrastructure Landscape

Host

Rod Rivera

๐Ÿ‡ฌ๐Ÿ‡ง Chapter

Guest

Max Tee

VC Expert, AI Investor, BNY Mellon

Discussing Mystic AI with Max Tee and Chris Wang

The conversation with Mystic AI covered various topics related to the company's role in the AI infrastructure market and the challenges they face in selling to corporates. The discussion highlighted the importance of building credibility, establishing strategic partnerships, and differentiating themselves in a crowded market. The convergence of infrastructure players and the need for strong branding were also discussed. Overall, the conversation provided valuable insights into the opportunities and threats in the AI infrastructure space.

Takeaways

  • Building credibility and establishing strategic partnerships are crucial for startups targeting the enterprise market.
  • Differentiation and strong branding are key to standing out in a crowded market.
  • The convergence of infrastructure players presents both opportunities and threats.
  • Understanding the challenges of selling to corporates and navigating their procurement processes is essential for B2B startups.

Episode Transcript

Initial Impressions of Mystic AI

Rod Rivera: We recently spoke with Oscar from Mystic AI, which was quite insightful. Christine, what were your thoughts on Mystic?

Christine: I was very impressed with both the company and Oscar himself. He came across as humble and deeply passionate about helping AI engineers design for the world. I thoroughly enjoyed our conversation.

Rod Rivera: Max, although you weren't present for our interview, you've encountered Mystic in the London startup scene. Any impressions based on that and our interview?

Max: From my perspective, Mystic AI is entering a relatively crowded market. It's encouraging to hear how Oscar is thinking about positioning Mystic AI moving forward. They're tackling a significant problem, though the scale of it is still unclear to me. I'm eager to delve deeper into this during our discussion.

Understanding Mystic AI's Business Model

Rod Rivera: Let's take a step back and interpret what Mystic is doing. Here's my understanding:

Mystic started before the generative AI wave, when most data science teams were developing in-house models for tasks like credit scoring or fraud detection. The challenge was deploying these models from local environments to production systems that banks could use. This was difficult because data scientists aren't typically developers, so their code doesn't always suit robust production environments.

Mystic stepped in to offer a solution: "Put your model in our infrastructure. We'll make it robust and easy to query. You only pay as you go based on the number of requests." This allowed teams to focus on their core strengths without worrying about infrastructure or deployment.

The landscape changed with the advent of generative AI. Now, instead of building everything from scratch, teams often download pre-existing models from repositories like Hugging Face. The challenge shifted to how to use these models in their own infrastructure.

Enterprises also became wary of solutions like OpenAI's GPT due to data privacy concerns and unpredictable costs. They wanted more control, preferring to host models on their own hardware.

Mystic adapted to this new landscape by offering software that can be installed on a company's own infrastructure, providing the benefits of managed deployment without the need to interact with a cloud API.

Max: That's a good summary, Rod. Mystic's motto seems to be about helping users run machine learning and AI models quicker and better. They're taking care of the infrastructure so that machine learning engineers can focus solely on the models. It's a noble mission and quite exciting from an opportunity perspective. As a machine learning engineer myself, having someone handle everything else would be incredibly valuable.

Rod Rivera: Christine, you've been involved with companies that Mystic would likely target - organizations where customer data is crucial and legacy infrastructure might not be cloud-friendly. If Oscar approached you with his product, what would you need to see to consider using it or even just having a discussion?

Challenges in Enterprise Adoption

Christine: When talking to Oscar, I felt his mission to simplify infrastructure setup was admirable. However, in the B2B space, especially with larger corporations, there are significant security considerations. Corporates tend to trust established brands like Microsoft. As a startup, it's challenging to differentiate yourself.

There are two aspects to consider:

  1. Technical and infrastructural convincing: Addressing privacy and security concerns is crucial.
  2. Business side convincing: There's often a gap in understanding between what Mystic does and what infrastructure is actually needed. This necessitates significant education.

The question is, how much of this education should be done in-house by the startup? It helps not only you but potentially your competitors as well. How can you differentiate yourself both technically and from a business perspective?

One way to gain credibility is to convince investors and advisors who can then vouch for you in the B2B world.

Rod Rivera: You raise an interesting point about competition with established organizations like Microsoft. The challenge for many infrastructure-focused startups is that while their user might be an individual contributor like a data scientist, the buying decision often comes from a CIO or CTO. These executives might prefer the convenience of a single bill from a provider like Microsoft Azure or AWS, even if Mystic offers a superior product.

Max: That's absolutely true. Having worked in several corporations, I've observed that convenience often trumps performance. As long as a solution is "good enough," that suffices for many corporates.

In a large corporate, decision-making tends to follow a bell curve when considering probability and impact. They're more concerned about avoiding mistakes than getting the absolute best out of everything they buy. This is where credibility, consolidated billing, and similar factors play well for big corporates.

The question then becomes, how can startups like Mystic lend themselves credibility to get their foot in the door? Because once they're in, they're in.

Strategies for B2B Sales and Market Penetration

Christine: This is a common challenge for B2B SaaS startups, not just in AI. Selling to corporates typically involves long sales cycles, procurement processes, and tender procedures.

One effective strategy has been to build credibility by targeting growth companies and startups first. Oscar mentioned this approach as well.

Rod Rivera: There's a saying that to succeed in B2B enterprise sales, you must have prior experience. This is challenging for many startups because B2B enterprise sales is heavily relationship-based. If you've spent years at a company like Salesforce, you have the connections to approach decision-makers directly.

Starting from scratch in the startup world makes this an uphill battle. It's remarkable when companies without this enterprise background achieve enterprise penetration, but they're few and far between. Most startups end up in the mid-market, using social media to access developers who might have the authority to make smaller purchases, resulting in deals of 200โˆ’500permonthratherthan200-500 per month rather than 100,000+.

Christine: This is why the go-to-market strategy is crucial. Mystic needs to focus on this and be prepared for trial and error. However, many corporations are aware of the difficulties startups face in working with them. This is why innovation hubs and technology centers exist - to facilitate technology transfer and allow for smaller proof-of-concepts in a lower-risk environment.

Another approach to consider is strategic partnerships. If you can partner with an organization that has a large corporate client base, like a bank or insurance company, you can gain access to a significant customer pool.

Rod Rivera: Those are excellent points, Christine. While I've expressed skepticism about innovation units, I do see value in using them to build relationships and potentially co-host events. For instance, if Mystic gets into Citibank's innovation unit or acceleration program and has a positive experience, they could leverage that to host joint events. Bringing in Citibank's CIO for a fireside chat on AI challenges could lead to valuable leads.

Marketing Materials and Buyer Preferences

Rod Rivera: Christine, imagine you're the buyer for Mystic in an enterprise setting. If Mystic approaches you with their product, what marketing materials or mechanisms would you want to see to start a discussion? A white paper? A case study? A demo?

Christine: There are multiple approaches, but from an investor's perspective, I'd want to see a pitch deck first. I need to quickly understand the company's differentiators - you need to be able to convey what's unique about your company in about 20 seconds.

Once interest is piqued, the next step would be a call and a product demo. You want to be wowed. This helps people who aren't directly in the business advocate for the startup within the organization.

The third step is finding the relevant person in the enterprise to understand their current projects and business problems. Is there a match? You'd then have a joint call to reiterate the pitch and demo, tailoring it to their specific business problem.

Alternatively, startups can participate in corporate-driven conferences, events, or hackathons to get their foot in the door.

Max: I have a slightly different view. The best way to engage with corporates is when they start reaching out to you. This happens when you've built a strong brand. If Mystic AI becomes known as the go-to place for easy machine learning model deployment, corporates will think of them when they encounter related problems.

Increasingly, it's not just procurement teams reaching out to startups. Business owners with specific problems are also starting to look for solutions. That's why it's crucial to differentiate yourself in both messaging and product within the ecosystem.

Competitor Analysis and Feature Comparison

Rod Rivera: How relevant do you find it for startups to mention their peers or competitors in their marketing materials? I'm ambivalent about this. Many companies include feature comparison matrices in their decks and on their websites, often portraying themselves as superior in every aspect. I suspect most people discount these comparisons.

However, I also think it's important to anchor your product in the market by referencing your closest competitors. This helps buyers understand whether you're an enterprise solution or a $500/month product. How do you advise companies to bridge this gap?

Christine: Ah, the feature comparison slide! You know the one - where everyone's crowded in one quadrant, and you're the only one in the top right. Personally, I never look at these because they're often arbitrary.

I would emphasize storytelling and the overall narrative of the company. Focus on your value proposition and the features you bring to the table. If a company is serious about buying a certain product or infrastructure layer, they'll do their homework anyway.

Max: I agree. As a corporate buyer, we care about what you bring to the corporate. Sometimes we don't even care about the features - we care about how you're going to solve our problem. How does your solution translate to money saved or revenue gained? This translation from your solution to business impact is crucial for any B2B software, especially when talking to decision-makers and buyers.

Christine: That's a great point. It's also worth noting that different pitch decks serve different purposes. An investor pitch deck looks different from a sales pitch deck. For selling to corporates, it's less about competitor analysis and more about demonstrating use cases, especially for infrastructure layer startups.

I actually really appreciated Mystic's walkthrough of their "bring your own cloud" feature. I found it very illustrative and easy to understand, even for those without a technical background.

SWOT Analysis of Mystic AI

Rod Rivera: Let's sum up our discussion with a SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis of Mystic:

Strengths:

  • Technically competent team
  • Accepted into Y Combinator, which is a significant achievement

Weaknesses:

  • Limited brand awareness
  • Shifting focus from cloud endpoint to "bring your own cloud" model, which may lead to lagging brand perception

Opportunities:

  • Growing market with no clearly defined leader
  • Continued growth in the AI infrastructure space

Threats:

  • Unable to undercut cloud provider infrastructure costs
  • Potential for established players to become more aggressive if Mystic gains traction

What are your thoughts? What's missing from this analysis?

Max: From an opportunity perspective, this is definitely a big market. The big players can't cover everyone, even in the mid-market. As software continues to eat the world, some form of AI will be integrated into almost everything, creating a huge opportunity.

The threat, of course, is that other players are trying to enter this area from multiple angles. That's why differentiation or targeting a specific market is crucial, even as an infrastructure layer.

Christine: I think that's a good summary. I'd add that as a young startup, it's unlikely they can immediately approach the biggest corporations. It will take time, and that's okay. Focusing on growth and scale-ups is a good strategy.

It's also important to think about team composition. Once the product is established and they're ready to scale up, having the right team on the business, sales, and marketing side becomes crucial.

Final Recommendations

Rod Rivera: What would be your one key recommendation for Mystic?

My recommendation would be to focus on branding, especially as they target enterprises. Their current branding feels very startup-oriented, targeting individual developers. They need to work on positioning themselves for enterprise clients.

Christine: I'd say they should continue what they're doing, but focus more on customer education and awareness. It's important to keep in mind the long-term goal of B2B sales if that's the game they want to play. Also, as they scale up, thinking about the right team composition on the business, sales, and marketing side will be crucial.

Max: My advice is a bit of wordplay: Don't be mystic. The word "mystic" has an air of secrecy, but because they're doing such great work, a lot of education is required for larger corporates. Being out there, doing interviews, appearing on podcasts - all of this will be incredibly helpful. They're addressing a significant pain point in large corporates, so they should keep getting their message out there. Keep calm and carry on, as they say.

Rod Rivera: That's a great way to wrap up - Mystic vs. being mystic. Thank you everyone for this fantastic discussion. To our listeners, tune in next week for our next discussion on exciting companies in the AI space and how we can build cool AI products.