AI summary by TwinMind
How AI is Reinventing Software Business Models ft. Bret Taylor of Sierra
Sequoia Capital
33m 49s
Brett Taylor Background & Introduction
- Former Google APM who rewrote Google Maps in a weekend
- Founded company, became Facebook CTO, founded Sierra (AI agents for businesses)
- Described as "10,000x engineer" who also excels at enterprise sales
- Currently Chair of OpenAI while building application layer company
Overcoming Single-Issue Voting as Entrepreneur
- Entrepreneurs often become "single issue voters" - defaulting to their comfort zone
- Product people always try to fix problems with product changes ("famed redesign like dead cat bounce")
- At 29 as Facebook CTO, struggled by trying to do everything himself
- Sheryl Sandberg intervention: "hold team to high standards, stop doing work yourself, replace underperformers"
- Shifted from "how do I conform this job to me" to "what's most important thing to do today"
- Became "chameleon" - Facebook sees him as engineer, Salesforce as suit
- Key lesson: don't let something you're good at become who you are
Sierra's Mission & Vision
- Helps companies build customer-facing AI agents
- Hypothesis: every company's main digital interface will be AI agent in 5 years
- Evolution: Yahoo directory → website → social/mobile → AI agents
- Customers include ADT home security, Sirius XM
- Goal: if you encounter branded AI agent, hope it's powered by Sierra platform
Three-Tier AI Market Structure
Foundation Models
- Will consolidate heavily due to capital intensity (like data centers)
- Handful of players, relatively low margin, very high scale
- Collect "taxes" from everyone in AI ecosystem
Tools/Infrastructure ("Pickaxes")
- Companies like Databricks, Snowflake
- Threatened by foundation model companies (AWS/Azure equivalents)
- Challenging but high scale potential
- "Closer to sun" = more risk
Applied AI/Agents Market
- "New software as a service" - purchasing agents that do jobs
- Companies like Harvey (legal), Sierra (customer experience)
- Most exciting opportunity - selling outcomes vs productivity enhancement
- Potential for first trillion-dollar applied enterprise software company
- Larger addressable market than traditional SaaS
Outcome-Based Pricing Model
- Sierra charges when AI agent resolves issue autonomously
- Free if escalation to human required
- Aligns with customer business model
- Evolution: box software → SaaS subscriptions → outcome-based pricing
- "Sales people get paid commission, why not AI?"
- Disrupts traditional arms-length vendor-customer relationship
- Eliminates 6-12 month systems integrator implementations
Business Model Advantages for Startups
- Changing business models harder than closing technology gaps
- Examples: Microsoft Windows→Azure transition, Adobe→subscription
- "Graveyard of CEOs" fired for failed business model transitions
- Startups not encumbered by existing business models
- New delivery models, technology models, and business models emerging
- Historical advantage: Salesforce beating Siebel, ServiceNow beating ITSM incumbents
Infrastructure Evolution & Focus
- Past: built own servers, brought to colo, tripped over power wires
- Now: virtualized infrastructure, cloud services, GitHub, open source libraries
- Virtual infrastructure support: Ramp, Rippling
- Much less intellectual overhead for starting companies
- Can focus purely on differentiation vs operational scaffolding
- Mistake: spending resources on non-core activities instead of customer acquisition/differentiation
- Metric: "what will make us happy a year from now"
Enterprise Customer Success Strategy
- Established companies (ADT = American District Telegraph) facing competitive pressure
- AI drives productivity in previously underserved departments
- 20,000-person contact centers costing hundreds of millions annually
- Converting opex to business investment enables price reduction and growth investment
- Similar to internet era: Walmart succeeded, Blockbuster failed
- Current executives lived through browser, smartphone, Instacart/DoorDash eras
Enterprise Sales Approach
- Speak customer's language, show empathy
- Ask questions first, listen, understand problems
- Reflect technology value through their described problems
- 20 AI vendors pitch daily, all sound the same
- Stand out through deep customer research using ChatGPT
- Show up as partner solving acute business problems
- Find intersection of differentiation and business value
Vertical vs Horizontal AI Strategy
- Strong belief in verticals over horizontal platforms
- Each industry has specialized workflows: telecom vs commercial bank vs residential bank vs insurance vs health insurance
- Skeptical of horizontal AI platforms (exception: LangChain due to open source)
- Enterprise: hard to win with "slightly better mousetrap"
- "Closer to solving company problem = more successful business"
- Marc Benioff lesson: 1 customer → 2 → 10 → 100
- Tactical excellence beats "ivory tower cleverness"
- Customer success reveals product direction
Pricing Strategy Insights
- Both cost savings and revenue growth matter
- Private equity companies focus on EBITDA/cost savings
- Competitive market CEOs prefer growth over cost savings
- Cost savings typically reinvested in growth (capitalism principle)
- Topline outcomes more valuable than cost savings
- Cost savings temporary - compressed over time as AI becomes standard
- LinkedIn recruiting example: HR departments need subscription vs usage-based for procurement
- Must understand decision makers, approval processes, budgeting cycles
- Sometimes easier for companies to prepay than pay on-demand
