For two decades, search engines have been the primary gateway to the internet. Their business model rested on a simple formula:

1. Aggregate attention by being the starting point for information discovery.
2. Monetize visibility through advertising, with sponsored results and keyword auctions.
3. Lock in users through habitual use and incremental improvements in relevance.

This model created some of the most profitable businesses in history. Yet generative AI now strikes at the very foundation of this model.

From Queries to Answers

Traditionally, search engines returned lists of links. Users had to sift through them to find what they needed. Advertisers paid to appear prominently in those lists.

Generative AI reframes this interaction. Instead of links, users receive direct, synthesized answers. The entire logic of the search engine — page ranking, sponsored slots, click-through funnels — is undermined when the answer bypasses the list.

Erosion of the Advertising Core

This shift has two profound consequences:

• Fewer impressions – If the user receives their answer in the AI output, they have less reason to scroll through search results or click ads.
• Weaker targeting – AI intermediates the query, stripping out much of the context advertisers once used to optimize placements.

The economic engine of search is paid visibility and weakens as AI compresses the user journey from query → links → clicks to simply query → answer. Continue reading →

For nearly two decades social platforms like Facebook, Twitter (X), YouTube, and TikTok have defined the way information is consumed online. Their business model has been remarkably consistent:

1. Capture attention through feeds optimised by algorithms.
2. Monetise discovery by inserting ads into those feeds.
3. Lock in creators and audiences by making the platform the essential intermediary between the two.

This model relies on one central feature: control over content discovery. The feed is the bottleneck and the algorithm is the gatekeeper.

Generative AI is shifting that balance of power.

From Algorithmic Feeds to User-Directed Discovery

In the old model users consumed what the platform’s algorithms decided was relevant. Discovery was not organic, rather it was curated by the system to maximise engagement.

Generative AI enables a different interaction. Users can now summarise, filter, or reorganise content themselves without depending on the platform’s feed. For example:

• Instead of scrolling through thousands of tweets, a user can ask an AI agent to summarise trending discussions.
• Instead of relying on YouTube’s “up next” suggestions, a user can instruct AI to curate a learning path across multiple creators.

The power shifts from platform-driven feeds to consumer-driven control. Continue reading →

Marketplaces have been one of the internet’s most successful business models. From Amazon to eBay to Alibaba they thrive by owning the consumer entry point, controlling search and discovery, and taking a cut of every transaction.

But generative AI is beginning to erode this model as well.

How Marketplaces Work Today

The traditional marketplace model depends on three pillars:

1. Traffic Control: Consumers start their journey on the platform.
2. Search & Discovery: The marketplace determines what products are surfaced thus shaping demand.
3. Transaction Capture: Every sale flows through the platform, securing fees and data.

The platform’s strength lies in controlling both demand aggregation and supply access.

The AI Shopping Agent Shift

Generative AI agents change the consumer journey. Instead of searching inside Amazon or eBay, a consumer can now tell an AI: “Find me the best mid-range laptop for graphic design under $1,500.”

The agent can then:

• Search across multiple marketplaces and independent sellers.
• Compare prices, reviews, and delivery times.
• Present the user with a shortlist — often with a single “best option.”

In this world, the marketplace is no longer the starting point. It becomes just one of many suppliers feeding the AI layer.

Continue reading →

Banks have long enjoyed dual advantages: they not only owned the balance sheet but also owned the customer relationship. Depositors and borrowers came for safety, trust, and capital; and then stayed because switching was costly and alternatives were limited.

Fintech challenged the second advantage, but not the first. Generative AI is now pushing the erosion further.

The Irreducible Moat: Balance Sheet Trust

Despite waves of fintech innovation, large regulated banks remain irreplaceable because they own balance sheets that are:

• Large and diversified: able to absorb shocks.
• Equity-buffered: giving depositors confidence.
• Regulator-backed: often explicitly by central banks.

Depositors, savers, and institutional investors continue to prefer this combination. For all the user experience innovation in fintech the fundamental preference for safety and stability means banks retain their role as systemic anchors.

Portability: Banks as Utilities

The true disruption is not replacing banks but commoditizing them.

Continue reading →

Consulting has long thrived on a simple premise: firms bring external perspective, proprietary knowledge, and structured problem-solving to help organizations address complex challenges. The value of the consulting model rests on three pillars: the brand premium of trusted firms, access to proprietary databases and benchmarks, and the ability to mobilize armies of analysts to process information quickly.

But generative AI may be eroding all three.

AI as a Force-Multiplier for Clients

Consultants have justified their fees by framing problems, gathering information, and producing structured recommendations. Yet senior and mid-level managers already hold most of the operating experience that consultants spend weeks “discovering.”

With generative AI managers can now:

• Frame insights directly: into prompts enriched with company data.
• Test hypotheses instantly: cross-checking with public benchmarks or market analysis.
• Generate strategic options: without needing external researchers.

In this model, AI helps clients convert tacit knowledge into structured intelligence. What once required an outside team can now be done internally faster, cheaper, and sometimes better.

Continue reading →