Emerj Artificial Intelligence Research

Barclays’ AI strategy shows where banking AI creates value first: Workflow fragmentation. Workload resilience. Operational control. In our latest analysis, we look at two AI use cases at Barclays, one of the largest universal banks, employing more than 100,000 people worldwide and reporting £7.7 billion in total income for Q1 2025. Barclays’ article confirms these figures and frames the two use cases around Microsoft 365 Copilot and IBM workload automation. The takeaway for financial services leaders is clear: AI value in banking does not begin with a broad innovation mandate. It begins where operational complexity already creates cost, risk, and service pressure. Barclays is applying AI in two areas: 1. Solving fragmentation with generative AI Barclays deploys Microsoft 365 Copilot to unify systems, reduce attrition, and streamline workflows across its workforce. The business problem is not just productivity. It is the friction created when employees navigate disconnected tools, policies, workflows, and internal systems at enterprise scale. 2. Improving workload efficiency with predictive analytics Barclays implements IBM workload automation tools to centralize IT operations, predict disruptions, and enhance service reliability. For a bank managing high-volume payments, online banking, card processing, regulatory reporting, branch accounting, and data warehousing, workload reliability is not an IT metric. It is a customer trust and compliance issue. The pattern is worth noting: Barclays is not treating AI as a standalone capability. It is embedding AI into the systems that already govern work, service delivery, and operational resilience. That is where financial institutions are likely to find measurable AI impact first. For banking leaders evaluating AI use cases, the Barclays examples offer a useful filter: start where workflow fragmentation or operational risk already has a measurable cost. Full Emerj analysis here: Artificial Intelligence at Barclays – Two Use Cases. https://zurl.co/1ZtKA

AI in healthcare and life sciences is not adopted because it is impressive. It is adopted when leaders can trust how it fits into care delivery, operations, compliance, and the operating model. Across recent Emerj conversations, five leaders pointed to the same adoption reality: Kathy Azeez Narain Chief Digital and Customer Innovation Officer, Hoag Brad Kennedy, MHA Senior Director, Business Solution Strategy, Orlando Health Linda VerPlanck Senior Director, Pivot to Growth Strategy Office and Transformation, Teva Raman K. Vice President of Technology, Elsevier Andrew Cuppia Senior Vice President, Service Line Management, Savista The shared lesson: AI adoption in regulated environments depends on domain expertise, organizational alignment, and trust. Not speed alone. For healthcare and life sciences teams, the question is no longer whether AI can create value. It is whether the business can adopt AI in a way that is measurable, governed, and useful inside the workflows where decisions get made. That is where AI strategy becomes operational reality.

The reason enterprise AI investments fail has less to do with the model or infrastructure than leaders think. The breakdown starts earlier. Teams rush to ship AI capabilities before they’ve clarified the human decision problem underneath them. In this HTEC-sponsored episode of the AI in Business podcast, Carsten Wierwille, Global VP of Product and Design at HTEC, explains why treating design as a late-stage layer in AI initiatives creates downstream dysfunction at scale. One point stood out: AI has reduced the cost of generating concepts. It has increased the cognitive burden of evaluating them. Senior leaders are now flooded with AI-generated ideas, prototypes, and feature proposals. Without design clarity upfront, enterprises end up reviewing noise instead of solving operational problems. The conversation also explores: • Why enterprise teams still build around technical possibility instead of business clarity • Why MVP frameworks break down for novel AI products • How AI should amplify human judgment instead of replace it • Why design is shifting toward cognitive design, focused on trust, perception, and decision-making The financial advisor example in the discussion captures the distinction well: The goal is not replacing expertise. The goal is helping professionals make better decisions under pressure. For leaders inheriting or scaling AI initiatives, this episode offers a useful framework for where design belongs in enterprise AI strategy. Listen here: https://lnkd.in/dUysAyEC 

An AI answer delivered with confidence but no awareness of your business context is not acceleration. It is an operational risk at scale. In this UpperEdge, LLC-sponsored episode of the AI in Business podcast, John Belden, Chief of Strategy and Research at UpperEdge, LLC, joins Emerj Artificial Intelligence Research’s Yolandi de Weerdt to outline two questions CIOs should ask before approving an AI-enabled delivery model. Where is the human in the middle? Does the system understand the context of now? The conversation focuses on a growing procurement challenge inside enterprise AI programs: separating real delivery capability from polished demos and automation theater. Procurement and program leaders evaluating SI proposals will find a practical framework for assessing governance, accountability, and contextual decision quality before deployment. Full breakdown: https://lnkd.in/d38dc-bB  

AI vision programs in manufacturing fail at deployment for organizational reasons, not model performance. In this Roboflow-sponsored episode of the AI in Business podcast, Jeff Witt, Digital Transformation Leader at a Fortune 500 global leader in building materials and fiberglass composites, joins Emerj Artificial Intelligence Research Editor Yolandi de Weerdt to examine why computer vision initiatives struggle to move beyond pilot environments. The challenge is rarely the vision model itself. The breakdown happens across infrastructure, governance, and operational ownership: • Fragmented camera environments across plants • Disconnected data pipelines between systems • No feedback loop between IT and operations teams • Vision deployments treated as isolated projects instead of scalable platforms The conversation also explores a critical shift in enterprise manufacturing AI: Vision programs scale differently when business units own deployment strategy instead of centralized IT teams. For manufacturing leaders responsible for AI outcomes, the episode offers practical insight into the architectural and operational decisions that determine whether a vision initiative stalls or expands across sites. Listen here: https://lnkd.in/dZX5qD7m

A voice deepfake does not need to defeat your authentication stack. It only needs the one agent trained to help the customer. In this Modulate-sponsored episode of the AI in Business podcast, Jon G Shende Global CTO for Data and AI at Thales Group, joins Yolandi de Weerdt to explain why fraud accountability cannot sit with security teams alone. Attackers exploit three things simultaneously: operational gaps customer friction workflow risk The vulnerability is rarely the model itself. It is the process that allows a caller to trigger a high-cost action without stronger verification. For CISOs and CX leaders reviewing voice authentication vendors, start by mapping which actions can still move forward after a successful social engineering attempt: password resets account recovery payment approvals escalation requests Shende’s joint-control model clarifies: who sets verification thresholds who owns the workflow who accepts residual risk That governance layer determines whether voice AI becomes a fraud surface or a controlled enterprise channel. Full episode here: https://lnkd.in/d7iEEe48  

Third-party risk did not get harder because companies stopped paying attention. It got harder because vendor ecosystems scaled faster than the systems built to manage them. Periodic reviews and static checklists were designed for a different operating reality. Enterprise teams are now dealing with: – Thousands of vendors – Shifting regulations – Real-time dependencies across the business – Risk signals that cannot wait for the next review cycle That is the shift Aravo is focused on addressing. Organizations are moving toward centralized, scalable approaches to third-party risk management, where risk teams can see more of the vendor ecosystem without adding more manual work. Across enterprise teams, the pattern is clear: Continuous monitoring is replacing periodic assessments Risk data is becoming more centralized and visible Systems are expected to surface signals without adding manual overhead Third-party risk is also expanding beyond compliance. It is becoming part of how enterprises think about governance, resilience, and performance. Read the full article: https://lnkd.in/gE-YgGpg At Emerj Artificial Intelligence Research, this is our focus: Connecting vendor perspectives like Aravo with how enterprise teams manage risk in practice, and bringing in operator insight that reflects what it takes to scale these systems. For teams looking to engage that audience: https://lnkd.in/dXsnJy8C

Retail teams do not need another broad AI overview. They need a clearer answer to one question: Where should we start? Across retail, AI is being applied to use cases like chatbots, personalization, logistics, and other workflow-level priorities. The challenge is not awareness. It is translating AI terminology and vendor claims into use cases a retail team can evaluate, discuss, and apply. Emerj’s AI Cheat Sheets are built for that gap. This brief 8-page PDF helps retail professionals quickly understand key AI concepts and apply them in their roles. Inside, you’ll find: • Applicable AI use cases across the retail landscape • Key AI terminology explained in simple, nontechnical terms • Direct links to deeper research on retail AI use cases, including chatbots, personalization, logistics, and more A focused guide for understanding how AI is being applied in retail today. Download link in the comments section.

Trying to reach VP+ enterprise AI buyers in the Fortune 500? Do not lead with your own story. Lead with the perspective your prospects already trust. Emerj’s Thought Leadership Series programs place sponsors side by side with Fortune 500 leaders to shape narratives that drive business. The result is not another vendor message. It is a market conversation built around the priorities, pressures, and language of the enterprise buyers you want to reach. Learn more: https://lnkd.in/dXsnJy8C 

AI‑driven modeling is exposing how much strategic risk enterprises carry when dose decisions still rely on legacy, phase‑by‑phase clinical assumptions. In this episode of the AI in Business podcast, Shefali Kakar, Global Head of PK Sciences and Oncology at Novartis, joins Emerj Artificial Intelligence Research’s Matthew DeMello to examine how deeper data integration is reshaping dose selection, safety interpretation, and patient variability assessment. Shefali shows how longitudinal analysis and exposure–response modeling cut down on unnecessary sub‑studies and give leaders clearer evidence for adjusting doses across diverse patient groups. She also explains how these methods bring earlier clarity to high‑stakes decisions and strengthen alignment across clinical functions. Executives can listen to the full episode to understand exactly what they gain when dose strategy becomes data‑driven — and why it’s emerging as a competitive advantage: https://lnkd.in/dwKjmudz