Don’t have an account? Sign up with Email

4 steps to getting practical about scaling AI for success

Organizations frequently ask how they can become great at scaling AI. Here’s are the key elements to consider if you want to scale AI well, repeatedly.


Lately, everyone’s ‘doing AI.’

Leaders now understand the urgency, but the ‘how’ remains elusive: according to our research, more than 75 percent of executives know how to pilot but are struggling to scale AI across their businesses. But scaling well matters. Our data indicates Strategic Scalers (those who successfully scale their AI) are seeing as much as three times the return on investment, compared with their counterparts stuck in the Proof of Concept (POC) stage.

My experience on the ground confirms the stats. My clients frequently ask me how to become great at scaling AI. Here’s what I’ve learned about the key elements to consider if you want to scale AI well, repeatedly.

Focus on value: think of your business strategy and AI strategy as one and the same
If you can define what really matters to your business, you can align your AI agenda to the company’s highest-level strategic goals, e.g., organic growth, or development of new products. The trick is to focus intently on value as it stands today, while at the same time having a vision for tomorrow. Where is your organization and your industry headed, and what does value look like three to five years from now?

While you’re thinking about value, get your data strategy in place. It’s the foundation and drives value as much as the AI itself. Strategic Scalers understand this and have a clear design and intent around what data is being captured, in what way, and for what purpose.

Reimagine how your business and your people work.
Deconstruct job roles and look at which tasks will be automated, which will require human-machine collaboration, and how this might change how people and teams intersect and interact—and their skilling needs. That boundary may shift as the organization’s AI maturity increases.

For example, applying AI to real-time data can help support decision making—but that doesn’t mean the decision itself is being offloaded to AI. Establish the right talent mix and be mindful of thinking data scientists are the only ones who matter when it comes to creating a route scaling AI. A diverse set of voices and talents – be they HR, data integration experts, business analysts, marketers, or software engineers, among others.
server wiring.jpg

When it comes to the build/buy decision, take advantage of what’s out there for success at speed and scale. Reuse, partner or buy before you even consider building new proprietary technologies in-house.

Get responsible AI on the board’s agenda
We’ve seen cybersecurity and data privacy moved from departmental to board-level issues, and AI should be no different. Responsible governance of AI must be elevated to the boardroom. If a machine’s decision turns out to be erroneous or unlawful, the potential fines and sanctions could threaten the business’s commercial sustainability. Bias and unintended consequences must be ‘designed out’ as far as practical using robust ethical frameworks (which will have to be defined before legislative change catches up with technology).

Obsess about getting to production to realize value
This is the crux of it and the most confused aspect. Much of the confusion comes from the fact that the AI roadmap looks and feels different from the time-honored software implementation approach.

Here’s a scenario I’ve seen more than once: a company spends six months on a proof of concept without considering what is needed to put it into production, the risks of the concept, data bias, data privacy, ethical considerations, etc. And when the time comes to move to production, the team realizes the POC isn’t built for scale and they’ve reinvented the wheel for no clear benefit.

To mitigate this risk, we use an AI start-to-end roadmap with our clients to define an AI use case's route to live. It lays the path for how to multiply value from the use case through continuous engineering, optimization and the extension of the feature to new use cases.

There’s even an argument that the POC will fall away entirely for companies who are well versed in AI already. We may almost be entering a ‘post-POC’ world, where we now know enough about AI’s potential to skip the pilot altogether and move straight to production in AI-mature businesses.

The prize is for the taking.

More for you

Loading data for hdm_tax_topic #better-outcomes...