Senior investor: AI has passed the "too cool" stage and is moving towards the "what's the use" stage, and it is necessary to learn from the experience of the Internet

As the funding machine for AI startups continues to roar, one of the questions we keep hearing is: **How big can this company get? **

Going one level deeper creates another set of questions:

• How sticky is this product? Early acceptance of many AI applications soars, only to plummet once the “newness” wears off.
• How is this app different from many similar apps? As the barriers to development for new applications come down, competitors are sure to flood the space.
• Is the ROI sufficient for new users? Replacing an existing application means learning new habits, investing time in training, and freeing up room in your budget.

As venture capitalists, we are constantly evaluating startups to determine if they can become the next big company that will change the world, or become a company valued at tens of billions of dollars. While serendipitous timing and luck play a large role in any major success, there are plenty of success stories to prove that there are often some core characteristics that companies must possess if they want to be track disruptors.

In this week’s article, we dig into what lessons we can learn from the previous generation of app development, and how these insights can guide new generative AI startups.

According to the latest report “Businesses at Work 2023” from cloud security company Okta, let’s take a look at the five most used and prolific B2B products in the world today: Microsoft 365, Salesforce, Slack, Zoom and Atlassian.

Microsoft 365 (Office)

Originally announced by Bill Gates at COMDEX in 1988, Microsoft Office was originally packaged as a software suite designed for a new wave of workers who were just becoming familiar with desk computers. Initially included three applications: Excel, PowerPoint, and Word. Today, Office has grown into Microsoft 365, with more than 350 million paying customers and more than a dozen widely used applications.

• How did they succeed? Our opinion:*

• Early Innovation: Office was designed at a time when computing was democratizing. Not only were office workers becoming more familiar with desk computers, but personal computers became more popular (a host of new machines started appearing in the late 1970s and 1980s).

• Workflow Solutions: Office consists of Word, Excel, and PowerPoint, applications designed to work together seamlessly to provide a comprehensive suite for a variety of office tasks such as word processing, spreadsheet analysis, and presentation creation . This allows users to switch between tasks without leaving the suite.

• **Become the “de-facto” industry standard: ** Vigorously marketed to become the de-facto standard in office productivity software, making it easier for businesses, organizations and individuals to collaborate and share documents. Microsoft is also aggressively bundling its solution with new computers and working directly with computer manufacturers to have it preinstalled on computers. ***Microsoft has also become the de facto leader in personal and professional computing.

• Continuous Innovation: Office is constantly innovating with the times (although perhaps not all users would agree!). With the advent of cloud computing, Office transitioned to a subscription-based model. They also extend beyond the core applications, continuing to dominate user workflows (such as Outlook, OneNote, Publisher, and Access). Microsoft also encourages third-party developers to create new plug-ins and extensions within Office applications, further enhancing their full functionality.

Salesforce

Founded in 1999, Salesforce is currently the largest cloud-based customer relationship management (CRM) software with a market cap of over $200B and annual revenue of over 30B+. The Salesforce suite has grown from an initial CRM product to a multi-product suite of sales and productivity tools, built on a vast repository of customer and sales data.

• How did they succeed? Our opinion:*

• **Innovative Cloud and SaaS Model: **Salesforce is an important pioneer in delivering software as a service through the cloud. This model enables businesses to access and use software over the Internet and eliminates the need for complex installations and lengthy setup times. They were able to support this with a subscription-based pricing model, ultimately changing the game for how most software companies deliver value.

• Simple and streamlined: Competitors in this space, such as Siebel and SAP, have interfaces that are clunky and difficult to use. Salesforce emphasizes user experience, ease of use, and a clean product interface. This ultimately improves stickiness as users come back to the product on a daily basis.

• **Comprehensive suite of solutions: **Salesforce started out as a CRM, but over time offers a variety of cloud-based solutions beyond CRM (e.g., marketing, automation, customer success, analytics wait). As a result, Salesforce has been deeply integrated into various business operations and has incurred high switching costs (meaning that the product alone cannot help new CRM startups win).

• Developer friendly: Builds a platform where developers can build custom applications on top of Salesforce and creates a central community where Salesforce users can build and sell their applications.

** **

Slack

Slack was founded in 2013 as a communication tool for a gaming company called Tiny Speck. Over the past decade, the company has gained tremendous popularity and has become one of the most widely used employee communication tools. They were acquired by Salesforce for $28 billion in 2020.

• How did they succeed? Our opinion:*

• Simple and user-friendly interface: Slack has always been praised for its user-friendly interface, which is easy to navigate and makes it easy for anyone to start using the platform. Slack also introduced channel organizations, allowing teams to create channels for topics, projects, departments, and more.

• Powerful Application Integration Ecosystem: Slack allows users to integrate third-party applications into their workspace, enabling users to increase productivity and customize and streamline their workflow. Centralizing work tools and the ability to seamlessly share data and files keeps users coming back to Slack, increasing stickiness.

• Product-led growth and strong user virality: Slack was a strong early supporter of the freemium model, which made initial adoption easy for individual users and eventually increased virality and large enterprise adoption.

Zoom

Founded in 2011, Zoom has become the number one video conferencing platform, beating out legacy players like Logitech and Webex, as well as a host of “next-gen” tools and products in product suites from the likes of Microsoft and Google. Even though today’s growth has plummeted relative to the pandemic, Zoom remains firmly on the desktops of millions of knowledge workers.

How did they succeed? Our opinion:

• **Best-in-Class Infrastructure and Reliability: **Zoom is truly the first video communication tool to offer a consistently reliable meeting experience. The infrastructure is designed to handle high-quality connections, international dial-in, and large numbers of participants. Depending on your location, there will be no communication delays or degradation in quality. For something as sensitive as video chatting, reliability is a top priority.
• Ease of Use: The product’s simplicity, intuitive “how to share” screen, and easy setup are key factors in Zoom’s meteoric rise. The ability to easily schedule Zoom calls and join meetings with just one click on any device (mobile, PC, iPad) is a key advantage.
• Collaboration Platform: Zoom offers a “Share Screen” feature that enables participants to collaborate on documents, presentations, and other materials. Participants can be anywhere in the world and don’t need to be paying users of Zoom (but “link sharing” is a key element of its virality).

Atlassian

The founding story of Atlassian, which was founded in 2022 by two Australians trying to use software to help other support teams, is perhaps the most unconventional of any other leading business app. Atlassian has since grown from a bootstrapped business with a single product (bug tracking) to a $40B+ global behemoth spanning multiple products and 10M+ monthly active users.

How did they win? Our opinion

• Pioneering self-service model: For most of Atlassian’s history, they didn’t have a sales force. Instead, they focus on developing a loyal market by building best-in-class project management tools for engineering teams. They then strategically expanded their product offerings through acquisitions, extending their customer base to teams around the development team*. * This allows them to keep sales & marketing costs low and accumulate more gross profit to make the product better.
• Deliberate and connected product extension: Atlassian initially offered a product: Jira for bug tracking. Over the years, they have strategically expanded into knowledge bases (Confluence), Trello (project management), and Bitbucket (CI/CD). Throughout their expansion, they have kept developers at the center of the customer experience, meaning products integrate seamlessly and create a delightful end-user experience.
• Strategic Acquisition: Atlassian’s growth is not purely organic. Over the years, they’ve made about 20 acquisitions (eg Trello, OpsGenie, Chartio, etc.) that enhanced their existing offerings and allowed them to enter new markets, effectively raising the switching costs of ditching the Atlassian suite.

What are the common themes?

Looking at these business applications used by millions of people, we can see many common themes emerging that apply to AI startups.

• The democratization of new innovative technologies enables the building of disruptive applications: Each of these companies has undergone a technological shift that makes the final application more accessible to a variety of users (in many cases, This is the transition to the cloud). When the underlying enabling technology is democratized, it creates a huge opportunity for a new set of applications to take hold. *** Lesson for Startups: Technological change doesn’t happen often, but when it does, it can cause a tsunami of results. Don’t take this for granted! ***
• Using stickiness contributes to “ultimate success”: The product must exhibit compelling stickiness and must become the platform of choice for the end user. Each of the products detailed above is typically used on a daily basis (often multiple times a day). Products that integrate well into existing workflows help make this happen. In other words, if you remove these applications from your employees’ desktops, they simply won’t work. ***Lesson for startups: Using models to generate novel or unique content is great, but in the end it won’t lead to huge results unless the application becomes critical to the user’s daily workflow. ***
• Integration, Compatibility, and a Developer-Friendly Approach → As the poet John Donne wrote “No man is an island”, likewise no very successful B2B application is an island. In this modern age, productivity apps don’t just need to integrate with each other, but that’s the only way they can sustain and thrive. Users may find it interesting to try a new product during the initial stages of the hype cycle, but unless the product integrates seamlessly with the rest of their workflow, we expect usage to drop significantly after the novelty period ends. Ensuring apps are compatible with each other (and maintaining the ultimate developer experience) is key to app success. ***Lesson for Startups: Don’t be an island. Learn how to integrate with other productivity tools to make your application easier for users to access. ***
• Powerful UI/UX and more pleasant user experience: Creating a pleasant experience for the end user cannot be underestimated. Whether it’s Zoom’s “one-click-to-join” or Slack’s innovative approach to organizing synchronous communications, a great experience can lead to viral growth and sustained engagement. *Lessons for startups: What is your UX charm? How do you ensure that customers are deeply disappointed if they stop using you? *

What can artificial intelligence startups consider?

We are at an interesting point in the hype cycle for generative artificial intelligence. As we detailed in a previous article, We believe we are at the “end of the beginning” and the initial excitement about generative AI is shifting from “This is cool!” to “What does this do for me?” help".

The slowdown in traffic to AI tools bears this out:

During this part of the cycle, it is helpful to look at time-tested applications to determine the path forward. To this end, we believe AI companies should consider the following questions when generating:

• **What functions can be realized by artificial intelligence and basic model technology that cannot be realized by traditional SaaS? **One of the great things about LLM and FM and related techniques is the ability of the model to reason about and create up-to-date augmented outputs. ***These models can think more like humans and eventually take over the workflow.
• **How to cover the entire workflow and improve user stickiness? **As budgets tighten, the ability to purchase another tool is challenged. Having a “cool” product that doesn’t increase productivity or show ROI can be challenging. AI tools that replace and enhance existing workflows while reducing the time and resources required to complete tasks will win.
• **Are there new business models for deploying AI compared to traditional SaaS? We’ve seen a lot of innovation in the open source investment model and API-as-a-service. Is there a way to take advantage of the switching costs associated with building the model? What new business models might emerge?
• **Can we move to value-based pricing? **The shift from a seat-based pricing model to a consumption-based pricing model is driven by customers finding value driven by usage and headcount. Given that many GenAI applications will drive value from the quality of output they generate, should we move to a value-based pricing model? Perhaps this will better incentivize applications to generate more accurate, relevant and high-quality output, rather than an “all you can eat” consumption model.
• **How do I integrate and collaborate with the surrounding ecosystem? **In short, AI applications isolated from the rest of the knowledge worker toolkit will die. Part of every business app’s success in the top 10 by usage is that it “plays well” with other productivity tools. As an AI application, what’s the best way to create links to both legacy and novel applications that my target audience might be using?

While the initial excitement about new AI applications in the workforce is starting to slow, we believe this is ultimately a good thing for the ecosystem as powerful, useful, and powerful new enterprise-level companies emerge. It’s useful to understand how some of the best B2B applications on the market today are built.

The question is: **How many applications will be replaced by AI tools? **

The author of this article is venture capital firm Madrona investor Sabrina Wu and partner Vivek Ramaswami, compiled by Wall Street News.