Unique Research · Insight / a16z Connect/Enterprise · October 18, 2023 · Editorial rewrite of ~2,265 CJK — Atlassian + Navan session highlights preserved
a16z’s annual Connect/Enterprise put generative AI and LLMs at the center. Below: distilled remarks from Atlassian President Anu Bharadwaj (with GP Kristina Shen) and Navan co-founder & CEO Ariel Cohen (with Ben Horowitz). Source: a16z — Building Products with Generative AI.
Atlassian: rewriting teamwork and operating leverage
Why is Atlassian so interested in AI? The answer is not unique to one company — the whole industry is pulled in for many reasons. Atlassian is a company that takes teamwork seriously; under AI, the definition and practice of teamwork change. A team is no longer only humans in different locations — it also includes AI virtual teammates that assist and amplify people.
The challenge and opportunity: how to integrate those virtual teammates, and how to change collaboration patterns for a new era. Atlassian has rich data on application development roles and workflows — and on how connected data can make each role more effective.
Interaction models and how applications are built are shifting fast. Even tightly regulated customers are adopting quickly. Customers know the value of data they already hold on the Atlassian platform and want to put it to work.
Core philosophy: an open toolchain. Atlassian does not want to be the company that supplies everything; different roles need different generative AI to enhance or redefine their work. The company already holds substantial context on where leverage is highest.
On Atlassian Intelligence, early work includes a collaborative editor that shows role-specific use cases — summarize, change tone, search in natural language — so people find information faster, act faster, and automate more easily. Potentially disruptive pieces such as virtual agents for customer service and help desks change the game: large headcount is no longer required for many of those tasks. Jira Service Management’s virtual agent has already saved substantial time.
Focus: combine those disruptive capabilities with additional actions the company can deploy to create more value and efficiency — and keep innovating as the landscape moves.
Redefining “developer”: widening who gets to create
Today “developer” is often limited to people who code and script. Keeping that narrow definition can unintentionally block others from becoming creators. Application lifecycles are shorter than before, which actually opens the door for more people to ship ideas as working apps — faster and more efficiently.
That trend will keep expanding. For companies that only serve coding specialists, this is a medium- to long-term strategic question. Atlassian’s view is different: it is not only for professional developers. The interest is helping more people turn ideas into reality even when they are not coding experts.
Collaboration stays core. Even as technology advances, human-to-human collaboration on projects remains essential. The path is inclusive expansion, not exclusion — new methods and tools so more people can realize ideas in multi-person environments.
Navan: generative AI for intelligent travel booking
Cohen agrees travel is an ideal generative-AI scenario. From Navan’s founding, the question was: why do people still call a travel agent? Machine learning was the answer — from day one, ML re-ranked search so users typically pick one of the top three results.
Generative AI adds a new possibility: predict and meet needs so users have time for other things. With frequent-traveler data, Navan can build precise models of preferences. Planning a trip to Paris, the system can anticipate services and book proactively — but people may not be ready for full automation yet. So the UI is hybrid: virtual assistant Ava interacts by need and gradually surfaces navigation.
Navan will watch feedback and behavior closely; which interaction style wins is still unclear. The travel use case is clear. Few people realize how large the infrastructure behind booking a ticket is — combining generative AI with that stack unlocks major potential.
Adapt and learn
Cohen has lived through internet, cloud, SaaS, and mobile waves; generative AI is next. He entered the industry in 1999. Last November he still was not sure what “generative AI” meant — his co-founder knew a little more, but not fully. That is the craft: endless learning. Whether you are a developer or in sales, fast learning and adaptation are required; if you do not invest in continuous learning, the world moves on without you.