DTM E54. The Future of User Feedbacks, Customer Acquisitions, and Moats in an AI First World- Rob May, BrandGuard.ai
From the Deep Tech Musings Podcast - Get actionable and tactical insights to take your Deep Tech startup from 0 to 1 [Idea to Traction]
Listen now on - Spotify, Apple, Google
Rob stands at the intersection of AI innovation and business strategy. He is the insightful author of the 'Investing in AI' newsletter, providing invaluable perspectives to investors and tech enthusiasts alike. He previously co-founded Backupify, then Talla, then was a General Partner at PJC. Rob has an intricate understanding of the startup ecosystem and the role of AI in shaping the future of industries. Currently steering the ship as the CEO of BrandGuard.ai, an AI tool to protect & boost your brand, and also a Partner at the AI Operator's Fund, which makes pre-seed investments in AI startups. He continues to be a driving force in AI's integration into the business world. He is an angel in 100+ early-stage companies, mostly AI companies. On the show we discuss,
How Rob Comes Up With Amazing Content for 20 Years
Synthetic Users and Why Real User Feedback Matters Less in an AI World
Why Horizontal Will LLMs Will Loose
The New Age Customer Acquisition Channels and Achieving “Agent-Market Fit”
Why AI Will Kill Most Forms of Long-Term Competitive Advantage
Listen to the episode here,
or on the below platforms
Links:
Rob May - robmay (LinkedIn), @robmay (Twitter), rob@aiinnovators.com (Email)
BrandGuard.ai - Website
Investing in AI Newsletter - Substack
AI Innovator’s Podcast - Apple, Spotify
Rob’s Articles Referred to in the Episode
Pronojit Saha, DTM Podcast - pronojitsaha (LinkedIn), @pronojits (Twitter)
Episode Tweet - Link
Show Notes & Summary:
(0:53) How does Rob come up with great quality content every week, year after year?
1. I write because my brain works differently, allowing me to see things in technology before others.
2. Started writing in 2003 as a way to clarify and think through my scattered ideas.
3. Writing helps me figure out what I think rather than seeking traffic or popularity.
4. Not motivated by traffic stats or clickbait; it's a personal journey.
5. My entrepreneurial ventures are also driven by my unique perspective on the world.
(2:03) The Inefficiencies in the Current AI Development Process: User Feedback
The efficiency of incorporating user feedback is a complex issue. Over my two decades in tech, I've observed how the landscape has shifted. In the past, startups faced skepticism from larger companies and could spend considerable time interviewing users, building prototypes, and refining products. However, my own experience led me to question the value of user feedback, as it can sometimes be misleading. Users may request features that end up being underutilized or irrelevant in practice. With the rapid acceleration of technology adoption cycles, user feedback's significance is diminishing. We're entering an era where technology evolves faster than human adaptability. If you're consistently ahead of the market and introducing innovations beyond users' current comprehension, extensive user feedback may not align with your pace of development. While there are scenarios where user feedback remains valuable, the gap between technology advancements and user understanding will continue to widen, potentially accelerated by AI, presenting a new set of challenges for product development.
(7:14) The Role of Product Development in a World of AI
I find it really intriguing how public companies are now resembling startups, especially since Twitter's acquisition by Elon Musk. The way they experiment with features, like the blue check marks, and swiftly roll them out, only to reverse course within days, mirrors the daily chaos of startups. It's a shift in dynamic that many might not realize. Another interesting concept I stumbled upon is the idea of synthetic users. It all began when someone with a domain like syntheticusers.com subscribed to my newsletter. Intrigued, I delved deeper and discovered the concept's essence. Essentially, instead of relying solely on real user feedback, some companies are using specifically engineered prompts to interact with AI models like chat GPT, effectively creating synthetic users. Surprisingly, this approach often yields valuable and coherent feedback, even though it may sound counterintuitive. These AI models, such as LLMs, have a vast knowledge base, making them capable of mimicking average user sentiment. Although human input remains crucial for edge cases, this approach presents a unique and promising perspective for innovation in various domains, incorporating both product categories and user profiles into AI models. The potential for innovation in this realm is vast, and we may witness more companies adopting these approaches in the future.
(12:08) The implications of horizontal LLMs losing in the AI industry.
1. I believe that foundation model companies will not dominate AI in the long run, despite their large investments.
2. The cost of training models is decreasing rapidly, making it more economically valuable to use specialized vertical LLMs.
3. I see potential competition between OpenAI and other big players like Amazon, Google, and Microsoft.
4. Advancements in compute architectures and chips could drastically improve efficiency and change the game.
5. There are rumors of OpenAI going bankrupt but I find them unlikely.
(16:54) The Role of Agents as New Customer Acquisition Channel: Achieving “Agent-Market Fit”
The most intriguing aspect of LLMs, in my opinion, is their potential to perform tasks through a series of steps, essentially building agent architectures. These agents can be tailored to various roles, like a CEO agent, which can draft reports, send emails, and more. This concept could revolutionize how work is managed, with individuals overseeing these digital agents rather than traditional teams. I often enjoy extrapolating technological trends to extremes, as I did with Clubhouse's growth. While it was hailed for its intimacy, I foresaw scalability issues. Now, applying this to Auto GPT and agent architectures, these agents could reshape how we choose software and vendors, given their ability to process vast amounts of data quickly. If, in the future, most of us rely on agents for such decisions, it will transform software marketing and acquisition. Startups should recognize the potential, as it could become a significant distribution channel, reshaping the B2B industry's sales and marketing practices.
(24:01) The Impact of AI on Long-Term Competitive Advantage
Exploring the concept of synthetic users and their impact on network effects, I see how the landscape of competitive advantage is evolving. In a world where LLMs can simulate users effectively, the traditional advantage of having more real users might diminish. This leads me to ponder the potential disruption of industries with dominant players like LinkedIn and Facebook, especially if agent architectures streamline our interactions across various platforms. It reminds me of a discussion in business school about competitive advantage in the internet age, where information availability transformed the business landscape. Now, with AI's capabilities, competitive advantage could hinge on infrastructure and unique patents, potentially reducing the number of traditional competitive advantages from seven to a smaller number. While I remain cautiously skeptical about this shift, I stress on the importance of stability in customer relationships, adaptability in teams, and other factors for startup founders aiming to establish defensible and competitive positions in this evolving landscape.
(28:15) The True Competitive Advantage in an AI First World: Customer Relationships and Brand
The emphasis on customer relationships is becoming increasingly crucial in a landscape where smaller companies can leverage agents to compete with larger ones efficiently. As technology makes tasks easier and more accessible, it prompts us to consider the ripple effects. When intelligence becomes pervasive, it's applied in various ways, from manipulating consumers with marketing gimmicks to genuinely solving their needs with innovative products. In the generative AI era, content and marketing will surge, making trust a scarce commodity. Knowing your vendors and their values will become more critical than ever. This aligns with the concept I've been working on in my current role, where we emphasize the importance of a strong brand that conveys our identity, values, and commitments to customers, as it's becoming a cornerstone in a world where AI commoditizes many aspects of business. Your point about nurturing customer relationships and consistent branding is spot on.
(30:50) Using BrandGuard.ai to Protect the Identity of Brands
1. My current startup Brandguard.ai has developed a brand governance platform that ensures consistency and adherence to brand guidelines across all channels.
2. The platform uses machine learning models trained on style guides and historical content to ensure everything produced is in confirmation with brand values.
3. The platform can be used by companies to score and review content, as well as by celebrities to protect their brand.
4. In the future, brands and celebrities may have their own customized models that are integrated with brand governance platforms to continually train and govern content.
5. This future scenario would involve loaning or giving access to models when collaborating with other companies.
(34:33) AI Innovators Community
1. I started focusing on AI in 2015 after selling my first company in 2014.
2. I began organizing dinners with AI professionals, which grew over the years into larger gatherings.
3. I formed the AI Innovators community to bring together individuals from different backgrounds to share knowledge and foster growth in the industry.
4. We try to keep sponsors non-VCs and non-companies to maintain independence.
5. We host monthly social and intellectual events, exclusive dinners with speakers, and occasional conferences.
(38:02) Crypto: What Gives?
1. I had the idea to use blockchain for identity verification in a crypto project I tried to launch in 2017.
2. I got subpoenaed by the SEC and their boss said all crypto is a security, so I had to go talk to people in DC.
3. I don't think blockchain and tokens solve many useful problems, and they are slower and not as flexible as traditional financial systems.
4. The use cases for blockchain are limited, and I don't understand why so much money is being invested in the industry.
5. VCs are investing in crypto because they can quickly sell tokens and earn a return, even though the specific use cases are not strong.
If you enjoyed this episode, please leave us a rating on Spotify, Apple, or wherever you listen to podcasts. It helps us reach more people who are interested in deep tech & grow the community.
Also, don’t forget to subscribe to Deep Tech Musings Podcast on pronojits.substack.com so you never miss an episode.
Thank you for listening! See you next time!