Turn your texts, PPTs, PDFs or URLs to video - in minutes.
The original article was published on weforum.org.
- Experts promise much economic upside from generative artificial intelligence (GenAI) but the response from businesses on the ground has been more muted.
- Businesses must get past novelty and focus on the return on investment of their GenAI investments to close the gap between promise and reality.
- Organizations must establish clear business objectives and deploy GenAI solutions that solve business challenges end-to-end.
The economic potential of generative artificial intelligence (GenAI) is staggering. According to Goldman Sachs, GenAI could boost annual productivity by 1.5%, driving $7 trillion in added economic value over the next decade.
McKinsey’s projections are even more optimistic, predicting that GenAI could add up to $7.9 trillion annually to the global economy.
What’s more, empirical research and business practice results confirm that optimism is well-founded.
Professor Erik Brynjolfsson from the Stanford Institute for Human-Centered AI adds further colour to such findings in a recent interview with the Stanford Graduate School of Business: “These are huge numbers. I’ve done lots of work on the introduction of new information technology over the years, and often, companies are happy to get 1% or 2% productivity gains.”
Business practitioners are also reporting great improvements in productivity. Mercado Libre, for example, found that augmenting their 9,000 human developers with GitHub CoPilot resulted in a 50% reduction in time spent coding.
An interesting trend is emerging in the productivity gains driven by GenAI. The productivity gains increase as the complexity of the task increases.
One such example is the increase in productivity gains as we generate content formats that are more complex to produce by traditional means. Typing text requires less skill than writing code, while more visual formats such as video are even more complicated to produce. Business data suggests that this complexity leads to higher productivity uplift from GenAI.
For example, Zoom used Synthesia’s AI video platform to create interactive training for more than 1,000 sales representatives globally. Zoom accelerated video production by 90%, saving $1,500 per employee.
While such applications of generative text, code and video give reasons for optimism, many remain sceptical about large-scale GenAI deployment.
A muted response
Business operators on the ground however, don’t seem to share the full excitement.
Deloitte’s report echoes similar sentiments. Quoting multiple sources, Deloitte highlights both price sensitivity and high expectations for enterprise GenAI adoption. So, it notes that the enterprise is bullish on GenAI’s long-term prospects, one survey participant from the report poignantly states: “Good luck trying to get me to pay for it.”
Focusing on ROI
It speaks to the fact that novelty should not distract from what is fundamentally key: return on investment (ROI).
While GenAI’s low barrier to entry is a great feature, it should not distract from other measurable objectives and a hypothesis about the business outcomes we expect to observe. As Gartner found, the biggest challenge when adopting GenAI is estimating and proving the business value.
GenAI’s long-term potential depends on evidencing its short-term value. That means GenAI pilots should have clearly defined success criteria before they launch.
Like with any truly transformative technology, GenAI can drive measurable outcomes along two main dimensions:
- Improving customer experience.
- Lowering unit cost.
Past platform shifts can indicate how to unlock ROI along these two dimensions.
Over the past decade, user-generated content, combined with the internet, led to new forms of expression and some of the most lucrative business models ever invented.
Over the next decade, AI-generated content will unlock similar results. Those who leverage the features unique to GenAI to improve customer experience will be the ones to see real returns on their investment. An early example is PepsiCo’s Messi messages campaign, powered by Synthesia’s AI video platform, which leverages the unlimited scale of AI video.
Messi messages allowed for personalized videos to be generated at scale with the football star’s likeness. Fans generated over 7 million videos in eight languages, a feat impossible to achieve with traditional video production.
Such top-of-funnel marketing deployment is only one example of GenAI’s applications. Customer experience improvements can impact revenue across the entire customer lifecycle. Depending on business function, GenAI can upgrade mid-funnel product demos, improve outbound sales conversions through personalized prospecting journeys or increase retention by transforming customer support.
Similarly, GenAI can lower the unit cost of the customer experience, as demonstrated by the measurable productivity gains mentioned earlier. GenAI deployment will differ according to organizational priorities and the value can be measured by cost, employee productivity or any other efficiency metric; doing so will ensure leaders share in the long-term economic upside GenAI is meant to unlock.
GenAI: a tools to solutions
Most of us have yet to solve the challenge of holding ourselves accountable to measurable pilot outcomes. Luckily, some early adopters are already moving into the next phase of GenAI adoption.
Once we have set and proven initial value, the hard part begins: reinventing business processes and the enterprise operating model.
Economic research has shown a lag between the invention of general-purpose technologies and their subsequent impact on growth. While the gap becomes smaller with each new platform shift, the difference can be measured in years.
For example, manufacturing processes needed to be reinvented for the productivity boom from electricity to materialize. Only after the introduction of the assembly line did we experience the true benefits of the technology. This suggests that the enterprise needs to adjust how it operates so that GenAI can reach its full potential.
GenAI pilots are limited in scope and focus on a narrow task. This allows testing of the initial hypothesis about where the value may lie. Once validated, however, the scope needs to be broadened from singular tasks to solving business challenges end-to-end.
For example, with sales prospecting, a pilot might test the conversion uplift – and hence the downstream incremental revenue – of using video instead of text in the prospecting journey, most likely limiting the scope to a few AI videos embedded into existing prospecting channels, such as outbound email. This scope could usefully prove value but not unlock the full potential inherent in GenAI.
Sales prospecting is a complex, multi-step journey, however. GenAI’s true potential comes from re-engineering the prospecting journey end-to-end, by combining multiple GenAI tools to create a truly differentiated customer experience. It could entail prompt engineering the brand’s tone into script generation, creating brand-compliant video templates that service each journey step or integrating customer relationship management data for a personalized experience.
Combining tools into a solution also lowers the unit cost of the differentiated experience, allowing the sales team to auto-generate content at scale without eating into valuable employee time they could spend connecting with prospective customers.
Creating GenAI systems is only half the equation, however. As important is reinventing the enterprise’s operating model and changing human behaviour, just like the assembly line needed, from training to safety policies.
Taking content localization as an example, the current model invokes most companies to use a hub-and-spoke model to translate content into local market languages with traditional overheads. Additional challenges apply if the model involves work from external transcreation agencies, easily adding weeks to completion.
Augmented by AI translation, organizations must rethink their operating model from the ground up. Central roles might shift from coordinating stakeholders to more hands-on execution. Tasks usually delegated to agencies might be made in-house. These inherently change the success criteria of those roles, straining both enterprise training programmes and the employee psyche.
Some evidence suggests that adoption might be more seamless than previous transitions. As GenAI systems develop better reasoning capabilities, human-machine collaboration becomes more effective, with AI coaches supporting employees at scale unmatched.
As Brynjolfsson notes: “Often with new technologies, there’s a bit of a decline before it takes off because it’s difficult and costly to implement changes, retrain workers and change business processes. In this case, we did not see a lull and performance took off over just a few months.”
While there is much work ahead, I couldn’t be more excited for the value GenAI can drive in the enterprise – as long as we keep our eyes on the prize: the ROI from GenAI.
Frequently asked questions
