CloudFactory recently hosted AI IRL, a live event in London at the Menier Penthouse near London Bridge. The gathering brought together AI and machine learning experts focused on applying AI in real-world scenarios, helping to bridge the gap between theoretical models and practical deployment. Attendees engaged with speakers actively moving AI out of the lab and into real-world settings, sharing insights on tackling the implementation challenges.
CloudFactory’s Matt Beale discussed how human processes drive inference-centric AI models, which deliver business value. Alberto Romero, Director of Gen AI Platform Engineering at Citi, discussed the often-overlooked vulnerabilities where data can be corrupted and the guardrails that keep them in check. Dan Saunders, Data Science Team Lead at Ripjar, explored the evolving role of training for decision-making within data-sensitive industries.
Here are three perspectives that stood out from the talk:
1. “What businesses care about is how an inference is correct. What we’re talking about is not necessarily building a perfect model to solve a problem but making sure the result that gets to the end-user or system is correct because that’s what a decision is made on.”
Mr. Beal explained that CloudFactory practices using inference assessments essentially do the opposite of confidence scores, measuring where a dataset’s inferences are most likely to be incorrect. With that information in place, CloudFactory relies on HITL (Human-In-The-Loop) as a manual review to improve data pipelines and deliver more sound inferences.
2. "In large banks, you can expect money to be thrown at compliance, but smaller companies will have trouble keeping up with those costs.”
Mr. Romero noted he hasn’t yet seen businesses unable to deploy due to the high costs associated with security. Still, he did highlight a surprising set of entry points where data can be manipulated or stolen, an especially dangerous challenge for finance organizations.
3. “Many bleeding-edge models (that aren’t pre-trained) require extensive expertise to optimize, and we want to be users, so we don’t want to rely on having that expertise in-house.”
Mr. Saunders spoke about the value of using pre-trained models as a practical approach for organizations that want to leverage AI without heavily investing in specialized skills. He highlighted how pre-trained models allow teams to implement robust AI solutions quickly, focusing resources on application rather than development, thus accelerating deployment while minimizing the need for deep technical expertise.
Meanwhile, other experts have expressed their own opinions about getting AI out of the lab and into the real world. Here are four interesting statements from the same week as our event.
4. “The novelty phase is over.”
Stefano Puntoni, Sebastian S. Kresge professor of marketing at the Wharton School and co-director of AI at Wharton believes “companies are no longer just exploring AI’s potential—they are embedding it into their strategies to scale growth, streamline operations, and enhance decision-making… We're now starting to see the integration of AI into various business processes as companies look to unlock its long-term value across the enterprise."
5. “After all, it is a balancing act – capturing opportunities and mitigating risks.”
“...We will work hand in hand with the financial regulators and industry players to foster a healthy and sustainable market environment, thereby facilitating the financial institutions to seize the opportunities and adopt AI in a responsible manner.” The Hong Kong Financial Services and Treasury Bureau (FSTB) recently introduced new financial services policies to enhance efficiency, security, and customer service, promoting the responsible application of artificial intelligence across the industry.
6. "We have been wowed at just the pace of growth, particularly on the consumer side.”
“...Even our enterprise businesses. They are young, but they are already doing an incredible amount of annualized revenue. We’re really excited by the potential there,” Said OpenAI CFO Sarah Friar to BloombergTV.
7. “Organizations are seeing the business returns of focusing time and investment in data intelligence programs when a clearly articulated data intelligence strategy is in place.”
“…The challenge for organizations today is to balance their attention between getting more business value from reliable data right now while at the same time laying the groundwork to reduce the risk from and accelerate value from future AI use,” said Stephen Catanzano, senior analyst at Enterprise Strategy Group, on data governance.
The collective insights from industry experts suggest that AI is transitioning from experimental phases to practical, real-world applications. Organizations are moving beyond exploring AI’s potential to actively embedding it into their strategies to drive growth, improve operations, and enhance decision-making. The shift involves a balancing act of seizing opportunities while mitigating risks related to compliance, data security, and scalability.
Watch the full presentation recording and get an exclusive look at how industry leaders are shaping the future of AI.
