New technology affords a new way to win, but we need new words to set strategies, writes Ash Fontana, author of The AI-First Company.

    The imperative to build artificial intelligence (AI) is clear, but how does a business win using AI? Now that we know more about what’s possible, we can settle on a vocabulary that defines the output of AI in terms of business strategy.

    Competitive strategy in this AI-first century centres on data learning effects (DLEs) — a new type of competitive advantage that compounds more quickly than any we know. DLEs start when new data makes existing data more valuable. This involves collating data, capably processing it into information and building AI. Previously, machines merely calculated things and stored the output in a database. Machines can now take that output, run more calculations to turn it into information, and learn new information to generate predictions. This learning compounds fast — generating even more predictions at an increasing rate — with the addition of new data.

    No-one has articulated this new type of competitive advantage or why it matters. Rather, existing books about AI focus on explaining the technology and its potential impact in the far future; or the theory behind how it may affect labour markets. People know that AI- First companies such as Google are worth a trillion dollars and are formidable competition, but they don’t know why or how they got to be so dominant, so fast. An understanding of DLEs is essential to start thinking about how AI-First companies such as Google were created.

    Data learning effects (DLEs) start when new data makes existing data more valuable.

    Ash Fontana

    Competitive advantage

    We’ve fallen short in defining this new type of competitive advantage because DLEs are a counterintuitive combination of three different categories of competitive advantage — learning effects, scale effects and network effects. However, DLEs don’t fit into any single one of these categories. Three effects differentiate DLEs:

    1. Learning effects

    Learning effects related to the knowledge economy, the era that started after the industrial revolution, when we moved from manufacturing to services. Consulting firms benefit from learning effects.

    DLEs are for today’s economy. This is the era in which we’ve made several leaps — distributing information digitally instead of verbally; sharing information automatically rather than manually; learning across minds instead of learning manually; learning on hardware (computers) as opposed to “wetware” (brains); and learning not on one node, but across nodes in a network. AI-First consulting firms automatically generate new insights from client data.

    2. Scale effects

    Scale effects were relevant to the pastoral and industrial eras when investment in key inputs could be leveraged into generating a large number of outputs. Manufacturing firms benefited. There are scale effects with data and having a lot of data can be a competitive advantage. However, more data only makes a product useful up to a point, after which it has less utility because it’s effectively the same data. That’s why DLEs aren’t just scale effects. AI-First manufacturing firms automatically optimise production to get more from less.

    3. Network effects

    Network effects applied to the software era, when lots of people and businesses could cheaply connect and online marketplaces could benefit. The difference between a network effect and a data network effect is what’s added to the network. With a network effect, something becomes more useful through the addition of communicative nodes to the network. With a data network effect, usefulness is enhanced by the addition of data to the network. DLEs can automatically grow because AI can create its own data. AI-First marketplaces automatically optimise pricing, which attracts more buyers and sells more.

    Seize the power

    None of these categories accurately or completely capture what DLEs and AI can do for business. Machines don’t learn the same way we do, they have a different type of intelligence, so “learning effects” isn’t an accurate term. We thought more data was always better — that “data is the new oil” — but not all data can be processed into valuable information.

    We thought AIs were a type of network effect after seeing social networks boom, but neural networks are only one type of AI. DLEs are more powerful than learning, scale or network effects — and they’ve helped companies such as Amazon and Facebook leapfrog the competition.

    Ash Fontana is the author of The AI-First Company, published by Penguin.

    Latest news

    This is of of your complimentary pieces of content

    This is exclusive content.

    You have reached your limit for guest contents. The content you are trying to access is exclusive for AICD members. Please become a member for unlimited access.