A Case for the Web in the Early AI Age

Not long after some prognosticators said generative AI will be the end of websites, the generative AI pioneer OpenAI is directing users to its newly acquired domain name.

OpenAI paid over $15 million to buy chat.com from Dharmesh Shah, the founder and CTO of HubSpot, according to a LinkedIn post yesterday by Shah.

Today, OpenAI Co-Founder and CEO Sam Altman shared the website with users in a post on X with only the domain name — which re-directs to ChatGPT.com.

The premium purchase by OpenAI, albeit minor among its billions in revenue, is an indicator of the web's clear connection to innovation in AI.

Despite the decades-old web perhaps being a legacy technology, the AI and technology industries are deeply entangled in the web — and the web is a critical business pillar in the growing AI market for several reasons:

AI Leaders are on the Web

The customer-facing chatbots that define the AI category are located on the web and function there.

ChatGPT, Gemini, Claude, Copilot, Grok, Poe and others are web-based tools.

Users sign-up for the products online and chat within their web user interfaces, with mobile apps being an alternative.

Modality-specific products by GenAI leaders are also web-based offerings, such as Midjourney for images, Runway for videos and ElevenLabs for audio.

The web is a core element of the pioneering go-to-market and growth strategies in AI.

AI Models Depend on the Web

Foundation AI models by AI leaders are trained on web data. Without the web, there are no foundation AI models as we know them.

Foundation AI models can also use search to tap into real-time web data beyond their fixed training data. Real-time web data was an early differentiator for foundation AI models, but it's fast becoming an expectation.

Notably, the ongoing relevance of AI models depends on new data, which is in finite supply. AI leaders are signing licensing deals with major web publishers to access their data.

In other words, the web is feeding AI.

AI Agents are Built Around the Web

The AI agents being built and released into the market are designed to autonomously complete tasks in employee workflows.

Those proposed administrative, repetitive and time-saving tasks asked are tied to web-based tools, such as email, websites and web applications.

Minimally, the AI agent promise of autonomous work depends on the agent interacting with web-based data supplied by the agent provider.

As AI agents mature, the best AI agents will interact with the provider's tools, customer tools and third-party tools to complete end-to-end tasks — all on the web.

AI Software is Being Built into Web Applications

Software-as-a-service (SaaS) products that are delivered via the web are mission-critical tools for companies across industries.

These web applications are synonymous with commerce and work — from back-end infrastructure, such as cloud computing management, to front-end tools, such as workplace software.

As companies race to invest in AI and create AI products in their categories, those products are mostly being integrated into their existing web platforms and in some cases, offered as stand-alone AI web applications.

In most cases, AI R&D and product development are focused on releasing an underlying technology, feature or module of a broader web application that serves a specific use case.

Learning Opportunities

Webinar