A majority of organizations now use AI in at least one function — 88%, according to McKinsey — but only 6% report significant enterprise-wide impact. This isn’t a failure of AI adoption. It’s a reflection of how organizations use AI.

To take an analogy from the past, the first cars used horse carriages and simply added an engine — the same frame, seating, and roads. It took a long time for the chassis to be redesigned. The technology arrived before the thinking caught up, and cars were reimagined.

Something similar is happening with AI. Companies are optimizing tasks without rethinking how they create value. According to the same study, only 23% of organizations that use generative AI have redesigned their workflows for the new technology. The rest are building very fast carriages and haven’t yet learned how to adopt a new business model.

The biggest impact of AI may not come from doing existing work faster, but from discovering entirely new ways to create value and generate revenue.

The four stages of AI value

Peter Drucker famously defined efficiency as “doing things right” and effectiveness as “doing the right things.”

Efficiency saves money — working faster and using less expensive products for an existing pie — while effectiveness makes money by growing the whole pie. Both matter, but they require different organizational muscles.

Stages one and two (the first two columns) in the above graphic of AI value are like factory work, which focuses on scalability, predictability, and high performance. These are cost-driven and measurable.

Stages three and four (the last two columns) are like laboratory work, which is built for experimentation, agility, and flexibility, and where new, unproven journeys are tested.

The factory mindset often wins in internal budgeting because it’s easier to see and quantify efficiency gains. It’s harder to see gains in effectiveness — the laboratory mindset — until an experiment succeeds.

The success of experimentation

Here’s an example of how experimentation can work: Tech entrepreneur Pieter Levels thought the only way to find out whether a company would work was to ship it — to experiment. Many projects later, several generate more than $250,000 per month combined. 

In another example, IKEA deployed a chatbot “Billie” in 2021 to handle customer service. It resolved 47% of all customer inquiries, or 3.2 million interactions. Costs dropped, a classic stage one outcome.

But 53% of inquiries were questions Billie couldn’t answer. IKEA saw this as an opportunity, not a failure. The company reskilled 8,500 call center workers as remote interior design advisers and built an entirely new sales channel.

The result: €1.3 billion in new revenue in 2022 from a channel that didn’t exist before the experiment.

Marketing versus the four horsemen

Advertising executive Rory Sutherland puts it bluntly in “The 4 Corporate Enemies of Innovation.” puts it bluntly. Most large organizations are concerned with cost-cutting and regulatory paranoia, not innovation. 

Finance, compliance, procurement, and HR departments — what he calls the “four horsemen of the bureaucratic apocalypse” — are disproportionately punished when things go wrong and therefore disincentivized from trying anything new.

Experimentation mandates should come from the marketing department, specifically marketing ops, because it’s responsible for future revenue, not the four horsemen departments.

Marketing ops already works at the intersection of data, technology, customer signals, and commercial outcomes, and can run experiments quickly and inexpensively.

In the IKEA example above, solutions surfaced in a customer interaction log and through experimentation, not in the boardroom. The people equipped to read that log and act on it were in marketing.

Get MarTech Insights That Matter

Platform news, strategy analysis, and industry trends. Trusted by 40,000+ marketing professionals.

How to create AI value for your company

If you recently adopted AI, your organization is likely in stages one or two of using AI to create value, using the factory mindset, and pleasing shareholders with efficiency. The efficiency wave is a necessary precondition for going further and creating more value with AI.

Stage three and four AI value can’t be planned. It must be discovered through deliberate, fast, and inexpensive experimentation. A planned AI roadmap isn’t necessarily the answer — building the muscle to experiment at volume and follow the right signals is.

Source link