# Stripe Sessions 2023

## Opening Keynote

**Patrick Collison (Co-Founder & CEO, Stripe)**

Mission: Grow the GDP of the internet

Macro Trends:

1. Financial Services turbulence
2. Reconfiguration around ecommerce
3. New payment schemes
4. Generative AI
5. Inflation Challenges

Findings:
73% new-term revenue and cost effeciences
72% expect to grow revenue by 25%

stri.pe/go/2022letter

Payment Volume:
2019 - $216 billion
2023 - $1 trillion

Weekly new bussinesses:
2019 - 1.6k
2023 - 5.6k

Stripe structural changes:

1. Continued product expansion
2. Enterprise growth
3. Revenue uplift

---

**Will Gaybrick - President of Product**

Global Payments | Embedded Payments BAAS | Revenue And Finance Automation

#### Global Payments

1. Optimized Checkout Flows
2. Unified Omnichannel experiences
3. Advanced Payment Capabilities

_Optimized Checkout Flows_

1. Stripe Checkout & Stripe Elements

---

**Clara Liang - Business Lead, Stripe Network**

Introduction to Link

1. 7% more likely to make a purchase
2. Save bank accounts
3. Big companies part of link - Twitter, Uber, Airbnb

5,000 Payment element users | 5,000 control group users

1. 10.5% revenue uplift for payment element users
2. Easy to maintain
3. Better customer experience
4. Frictionless checkout

---

**Katie Dill, Head of Design**

STRIPE TERMINAL

Introduction to Stripe Reader - Tapping card

- Customizable checkout
- all-in-one POS
- 50% greater battery life
- Offline mode

Amazon is moving to Stripe

1. Network tokens
2. Overcapture
3. Multicapture
4. Access to PINless debit netwokrs
5. Local acquiring

Choose between a PAN & Network token

Enhanced Issuer Network - Businesses connected to Issuing Bank

**Dave Nicole, Business Development, Global Payments, Uber**

---

**Tanya Khakbaz - Head of Marketing**

**Alex - Platform Lead**

Why Stripe Connect?

1. Get to market faster
2. Make more money
3. Improve user retention

Introduction to Connect Embedded Components

1. Can be integrated with Stripe Apps
2. Monetization - announcing brand new pricing rules.
- Buy now, pay later options can be added
3. Instant payouts
4. Tap to pay using Stripe Reader

---

**Vivek Sharma - Business Lead**

Revenue & Finance Automation

Process:

1. Acquire
2. Collect
3. Report
4. Grow

Cloudly (DEMO)
Stripe Products used - Billing, Invoincing, Tax, Revenue Recognition, Data Pipeline, Sigma

SIGMA

- Get data using SQL queries
- NLP model that understands english to create the query
- Will be released later this year

**Ritchie Bros (Ranbir Chawla) + Stripe - Partnership discussion**

---

**Emily Sands, Head of Information**

Introduction to Docs AI - asking questions in english instead of searching docs
Introduction to Workbench - Dev console (Beta - workbench.stripe.dev)

---

## 2. Panel discussion - What’s happening in fintech—banking-as-a-service, compliance,and new revenue models

1. With the failure of Banks, FinTech companies have realized their potential and taken this as an opportunity to introduce new banking services and revolutionize industry
2. Country regulations and compliance policies need to be redefined for more FinTech companies to be given a chance.
3. Revenue models should re-defined keeping user trust in mind. Monetize transactions.
4. High focus on reducing the inconvenience in user journey.

## 3. Breakout talk - Why payments experience is customer experience

Problems:

1. Customers are looking for faster medium of payments
- Under 10 seconds checkout
- Auto-fill
- Multiple mediums/gateways available for checkout
2. Customers often don't carry wallets so it should support Virtual wallets like Apple Pay

Integration of Link with Stripe

Improvement in overall lead --> customer conversion metrics.

## 4. Panel discussion - Accelerating product development in the age of AI

**Github + ServiceNow**

Learnings:

1. AI must always be suggestive and not conclusive
2. AI should be used to analyze customer data/feedback and improve user experience.
3. AI should be used in tandem with compliance policies & ensure security standards are met.
4. AI should be used to increase productivity.

## 5. Fireside Chat with Sam Altman

**Sam Altman - CEO of OpenAI**

A lot of people are worried about AI. Creating systems that learn complex problems

Finally gotten to a new computer interface. New way to interact with computers is going to go super far.

_Favourite GPT use-cases?_

- Unanswered

_Conversational bots_
Earlier the technology couldn't deliver or fulfill user needs. Now the user has more control & desires are being fulfilled.

_What's next after GPT 4_

- Probably 5 or 4.5?
- A system that can interact with more than text. Audio, Images, Video
- A lot of intelligence is not well distributed. So the focus will be on that.
- Adding new knowledge.

_AI Hallucination_

- On the path to fixing that?
- Very little forgiveness of a user to a computer making mistake. The tolerance level is different. So it's going to take a while.

_Tracking Model Performance_

- Looking at the model and making them harder, increasing robustness.

_Research + Product Split_

- We're still in a vry steep part of technology discovery.
- Still understanding if the product to make is possible
- Focus on making things to work but may not be right
- Certain approaches should work rather than certain technologies to work
- Difficult to predict so definitely focus on improving the algorithm

_Compute power_

- Inelegant to add more compute but can be helpful

_Truth seeking research project_

- All about people you hire
- Being dilligent about them and focus on if they fit in.
- Have an open-mind and not be focused on one singular team/mindset.

_Taking real-time feedback_

- Accept much longer time horizon
- People who don't depend too much on external validation
- Need people who are open to exploration and not focus on instantly hitting dopamine levels but instead gradually enhancing it.

_Learnings_

- People apply lessons from previous experience.
- Look at problems from a fresh pair of eyes

_Failures_

- Journey wasn't easy
- People didn't believe in the product
- Not a lot of technical progress
- Not very clear how to make AI cause it was fairly new
- People said Deep-mind is non-touchable
- Constantly not able to find enough money, people
- But you keep going to get something that works

_What kept you going_

- Very less customer feedback in the beginning due to internal research + niche market (very less people with that knowledge)

_Competition_

- Like early days of computing (Apple vs. Microsoft)
- Very chill now
- People want to be faster in launching products + also focus on collaborating to get the right product out.

_Regulations_

- Computing systems above certain thresholds are kind of powerful.
- Global regulatory authority should exist to pass certain test
- Standards for audits should exist
- Don't cap compute power
- There is a need to increase regulatory demands depending on the use-cases - humans protecting humans, robots protecting robots

_AI Startups OpenAI is looking to invest in_

- Anything in the direction of designing a new interface model
- We underestimate the creative energy of the world

_Software platforms & Productivity gains through AI_

_Advice for non-AI companies_

- Internal productivity
- Embrace AI tools to figure out how you can increase productivity

_AGI_

- AGI is a system that can dramatically increase the progress of scientific research
- Working in collaboration & becoming part of society and focuses on scientific progress

_Why invest in Stripe_

- One of the largest makes.
- Taking friction out of economic transactions
- Better, faster, cheaper, more accessible
- Stripe has outperformed expectations
- Big market idea, great founders
- Wrote a cheque for Stripe (ironic)!

_Product feedback for Stripe_

- Stripe does a good job with customer support & documentation
- OpenAI will focus on that as well
- Technical documentation vs Business documentation (creating a company that does blah blah)
- GPT 4 vs. Stripe AI - Will people ask questions to Stripe AI or GPT 4? Depends on user experience and not filled it out yet

_Data & usage problems_

- Don't train data submitted through the API
- Understanding how sensitive it is?
- Continue making models through internal knowledge
- Make them long-term stable
- Focus on how much intelligence can be gained from how little money

_OpenAI integrating with real-world_

- Long time until GPT-4 would come on iPhone

_Advice for start-ups trying to integrate AI_

- Everybody can go full remote with no loss of creativity
- Technology isn't good enough that we can go fully remote
- Feels like start-up needs more time to understand ideas & nuances of these things
- Slack and zoom work but not fully create productivity

_Expert advice vs AI_

- Sometimes experts can be wrong
- Avoid their twitter
- Experience built to creation of AI