Jun 2, 2017 · 25m · saastr
AI: The New Platform for SaaS
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Tomasz Tunguz of Redpoint Ventures and Ludo Ulrich of Salesforce examine the transformative impact of machine learning on enterprise SaaS, providing actionable strategies for startup defensibility, product design, and vertical market expansion.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Jason as informed peer 2.0 Guest teaching 1.7 Guest disagreement 0.1 Jason pushing back 0.0
speaking balance: gold is Jason, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 17:55 Deconstructing chatbot overpromising
Tom forcefully warns against claiming an AI chatbot can answer any question, dismissing overhyped marketing as a direct cause of user abandonment.
Hardest push from Jason ▶ 19:05 Redirecting conversation to practical founder adviceLudo presses Tom for actionable, tactical advice on how founders can avoid being crushed by cloud monopolies and scarce talent.
Biggest teaching moment ▶ 21:03 Educating on proprietary data sourcing frameworksTom outlines the two concrete strategies for acquiring proprietary enterprise datasets, illustrating with real industrial IoT and recruiting company case studies.
Jason holds their own ▶ 10:45 Demonstrating Salesforce incubator AI portfolio depthLudo demonstrates internal domain expertise by citing Salesforce acquisitions and incubator trends in unconscious bias and team composition AI.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Jason as informed peer | Guest teaching | Guest disagreement | Jason pushing back | Why |
|---|---|---|---|---|---|---|
| Defining Machine Learning and Core Business Functions | 0 | 0 | 0 | 0 | This is a solo presentation monologue by Tom Tunguz introducing machine learning fundamentals and categorization. Because there is no host interaction during this prepared keynote segment, all dynamic scores are zero. | |
| The Deep Learning Revolution and Technical Drivers | 0 | 0 | 0 | 0 | Tom Tunguz continues his solo keynote presentation detailing the shift from manual feature selection to deep learning automation. As a monologue segment, all interactive pushback and schooling scores are set to zero. | |
| Vertical SaaS Transformation Across Industry Sectors | 0 | 0 | 0 | 0 | Tom finishes his keynote presentation outlining Redpoint's five investment criteria and vertical SaaS transformation. Without active host dialogue, all dynamic scoring scales remain at baseline zero. | |
| Enterprise Demand and Salesforce Ecosystem Trends | 4 | 2 | 0 | 0 | Ludo transitions into the fireside chat, sharing Salesforce's ecosystem perspective on enterprise AI adoption and asking Tom how portfolio companies manage hype. The exchange is highly collaborative and aligned. | |
| Seamless User Experiences and Value-Driven Pitching | 4 | 2 | 0 | 0 | Tom shares his experience with Nuance Dragon and the film Her, illustrating how AI must remain invisible and value-focused. Ludo readily validates this, emphasizing the rule of pitching value without relying on AI buzzwords. | |
| Human-Robot Interaction, User Expectations, and Trust | 3 | 4 | 1 | 0 | Tom details the transition from human-computer interaction to human-robot interaction and cautions founders on setting low user expectations to preserve trust. Ludo agrees with practical examples from sales email tools. | |
| Defending Against Cloud Monopolies and Sourcing Data | 3 | 4 | 0 | 0 | Ludo asks how startups can compete against major cloud monopolies and secure proprietary data. Tom breaks down the two viable strategies: building workflow apps to generate proprietary data pipes or partnering directly with Fortune 500 enterprises. |
Statements from this episode (15)
Tunguz: Machine learning simplifies to just four core capabilities
“Machine learning is really simple. All it does is it teaches the machine to find patterns and data, and there are four things that you can teach a machine to do.”
Tunguz: Deep learning's breakthrough is automating feature selection and tuning
“The big advance in deep learning, basically, is that deep learning automates both processes.”
Tunguz: Google's WaveNet generates speech indistinguishable from human voices
“Google has a product called WaveNet. It's a computer that speaks so well that no human can tell it's a computer.”
Tunguz: Microsoft achieved human parity in speech recognition in 2016
“Microsoft released some research in 2016 that a computer can understand human speech as well as another human.”
Tunguz: Google released a zero-shot translation model for unseen languages
“And Google released a machine translation algorithm that will translate from English to any other language even if it's never seen that language before.”
Tunguz: AI will transform vertical SaaS at cloud computing scale
“The big theme about machine learning is, it's going to change the world of SAS because all of a sudden, it's not just, The machine learning innovations will happen in the core categories that we've seen in the past. Those horizontal companies selling to custom…”
Tunguz: AI algorithms are commoditized; startups must build proprietary datasets
“The algorithms themselves are not that innovative. They're not that close source. So you, we're going, the company that's going to really innovate is going to need its own data set to be able to train its own models.”
Tunguz: Google, Microsoft, and Amazon will dominate foundational AI APIs
“Google, Microsoft, Amazon, they will be releasing the APIs that lots of companies will be building on, competing with them. You may want to compete with them, but we will likely not invest in that because it's a very challenging go to market strategy.”
Tunguz: AI innovation alone fails without a strong go-to-market advantage
“A machine learning innovation, in, in and of itself, is not enough to generate substantial customer demand. It has to be able to change the go-to-market.”
Tunguz: Off-the-shelf AI models only hit 80% of required enterprise performance
“You can take a generic NLP or speech learning model and you can get 80% of the way there, but in order to really have a fundamentally new experience, you need to get it 95% of the way there, and you can't do that with just off-the-shelf algorithms. You're goin…”
Tunguz: Machine learning products succeed when they operate invisibly in the background
“That's a great, ah, analogy for how machine learning succeeds. And what I mean by that is, it's not in your face. It's hidden in the background. It's doing the work behind the scenes.”
Tunguz: AI startups must pitch customer value without mentioning machine learning
“Can you pitch your startup without saying machine learning? And what that means is you're not focused on the technology. You're focused on the value proposition. You're focused on why the customer is going to care about it. Why the buyer is going to be promote…”
Tunguz: AI interfaces must set user expectations deliberately low to build trust
“What it really comes down to in order to get people comfortable with UI, particularly interactive UI, is being able to set the expectation of the user appropriately, and then and that means much lower than you think it typically ought to be. And then the secon…”
Tunguz: The first ML SaaS startups will hit $15M ARR by 2018
“I think we're in the world of, like, there are very few SaaS companies that use machine learning at scale, and I think maybe next year we'll be talking about the first handful of them are, that are 10 or fifteen million in ARR, and we're going to be trying to …”
Tunguz: AI technology advantages do not automatically lower customer acquisition costs
“Like, let's say you know, you create the next version of Zendesk that's got a whole bunch of ML. It's not necessarily that that technology, you know, that Intel inside technology is going to all of a sudden reduce your cost of customer acquisition relative to …”