Hello. This is Alex Bakker with what’s important in the IT and business services industry this week.
If someone forwarded you this briefing, consider subscribing here.
GCCs and AI Share Outcomes
Over the last few years, the two biggest trends in the IT and business services industry have been the adoption of generative AI and the increasing interest in building or growing global capability centers (GCCs).
While those two trends might seem disconnected – one is a technology, and one is a delivery model – they both share the same objectives: saving money and driving output.
Enterprises are adopting AI in the hopes that it will drive new ways to use data and automate more complicated tasks and types of work. And they will be able to do so cheaper.
Enterprises are building or growing GCCs to get value through labor arbitrage. And, in addition to the cost-reduction business case, they can construct a talent pool to drive innovation while retaining a low-cost profile, making GCCs more compatible with the kind of high-risk-high-return projects that are likely to actually be innovative.
So, of course, organizations are going to try to put these two concepts – AI and GCCs – together. We wrote last week in detail about how much organizations are spending on GenAI and the types of skills that are most in demand. This week we’ll look at what might prevent that from working as intended. (See Data Watch)
Data Watch
Background
From one perspective, AI capabilities are software applications that have proven to be highly flexible across a variety of business domains. From another, AI is just user interface wrapped around some very complicated math. And the words, numbers and information on which you perform that math will have a significant impact on the results you get. As they say, garbage in, garbage out.
In a lot of the use cases for AI, this has been ignored. The tools available have come with their own data embedded by the training data, and end users have been primarily leveraging generic information.
As organizations get more mature in their adoption of AI, they are recognizing that AI tools built around specific business, domain or process information work better than off-the-shelf tools. In other words, generic AI is table-stakes, but working with custom data is the only way to build differentiated value.
What's Next
That’s where AI and GCC challenges meet up. To drive differentiated cost or differentiated innovation with a GCC, organizations must build it using the data they have. And the data needs to be good.
Looking back at the chart, we see poor data drives poor outcomes or delays the ability to deploy AI. This highlights the costs of AI that are disconnected from business outcomes. And when users struggle to integrate AI into their processes or workflows, they remain unconvinced that AI is actually any better than other methods they have at their disposal. This builds resistance that prevents people from wanting to learn the skills needed to leverage AI to its fullest potential.
For GCCs to successfully drive value from AI – and for businesses to see ROI on the kinds of high-risk-high-return innovations they want to drive through their GCCs – they need to improve the quality of their data and the processes that maintain it.
About the author
