AI in the Enterprise: Accelerating Performance Through Intelligent Systems

Leading enterprises are leveraging AI to boost profitability, enhance customer engagement, and streamline operations. Klarna, BBVA, Lowe’s, and more overcome adoption barriers and build long-term innovation capacity.

Key Points

• As enterprise AI delivers strong returns—like Klarna’s $40M profit gain and BBVA’s 2,900 GPTs—smaller firms face cost and access barriers, prompting wider adoption of fine-tuning, model compression, and open-source tools to scale AI affordably and effectively.
  
  
• As consumers demand more personalized and explainable AI—evidenced by Indeed’s 13% hiring boost—firms are overcoming data inconsistencies and analytics gaps by fine-tuning models and using transparent communication to scale trust and human-centric engagement.
  
  
• As AI agents like Codex CLI and Gemini 2.5 Flash redefine competitive advantage, firms are addressing developer and integration bottlenecks by adopting automated, no-code platforms—such as Mercado Libre’s Verdi with 99% fraud detection—to drive precision, speed, and intelligent workflow efficiency.
  ​
  
• Decentralized AI tools like ChatGPT Enterprise are accelerating cross-functional innovation, but centralized control and skill gaps remain barriers—prompting firms to upskill teams and democratize access, as seen in BBVA’s rollout of 2,900 GPTs to foster scalable, employee-driven solutions.​

Download the Full Report

Integrate key insights into a holistic strategy for your company.

te2504284_-_ai_in_the_enterprise_accelerating_performance_through_intelligent_systems.pdf
File Size:	761 kb
File Type:	pdf

Download File

---

AI delivers measurable improvements in productivity, profitability, and operational efficiency across industries

Enterprise AI adoption is driving profit gains, faster service, and enhanced workforce performance

AI adoption is delivering measurable results across workforce productivity, automation, and customer engagement. Klarna, for instance, has projected $40 million in profit gains through the rapid deployment of an AI assistant that reduced customer service resolution times from 11 minutes to just 2. BBVA has empowered employees to create over 2,900 custom GPTs in just five months, fundamentally transforming internal workflows. These results underscore the broader enterprise benefits of adopting AI, with OpenAI reporting measurable improvements across workforce performance, automation of routine operations, and enhanced product capabilities.
​
Developer bottlenecks and premium AI tool pricing strain innovation and widen adoption gaps

Developer resources are frequently cited as the main bottleneck and innovation inhibitor. Iterative model refinement—such as that conducted by Klarna—is costly and time-consuming, while premium access tiers from vendors like Microsoft and Anthropic exacerbate disparities in access. Copilot’s upcoming Computer Use feature and Claude’s advanced reasoning tools are currently reserved for top-tier users, compelling organizations to invest heavily just to maintain a competitive edge.

Fine-tuning, model compression, and open-source training lower costs and build internal resilience

To navigate these financial and operational pressures, companies are leveraging fine-tuning and model compression techniques to boost efficiency and performance. Lowe’s achieved a 20% increase in product tagging accuracy and a 60% improvement in error detection through GPT customization. Indeed collaborated with OpenAI to deliver the same job-matching performance using 60% fewer tokens by training a smaller, task-specific model. Open-source tools and training resources, such as DeepLearning AI’s autonomous browser agents course, enable organizations to reduce reliance on costly vendors and build internal capacity.
​

​Evolving consumer expectations demand personalized, transparent, and responsive AI engagement

Customers increasingly expect AI-driven experiences that offer real-time, human-like relevance

Customer expectations are evolving rapidly. The demand for contextual, transparent, and personalized interactions is reshaping how firms approach AI deployment. Indeed’s use of GPT-4o mini to generate tailored 'why' statements in job recommendations led to a 13% uplift in successful hiring outcomes. As digital agents mediate more customer experiences, users increasingly expect intuitive, human-like interactions backed by real-time reasoning.

Data inconsistencies and rising AI expectations challenge organizations’ ability to meet customer needs

Organizations face major hurdles in responding to changing behaviors. Incomplete and inconsistent product data—such as that faced by Lowe’s—make it difficult to align AI output with user intent. Moreover, advanced AI platforms like Claude set new benchmarks for contextual relevance, requiring firms to build adaptive, continuously trained systems. Those lacking behavioral analytics or real-time integration risk diminished loyalty and engagement.
​
Fine-tuned models and transparent practices improve satisfaction, trust, and brand loyalty

Addressing these issues requires technical refinement and trust-building. Fine-tuned models trained on proprietary datasets improve accuracy, maintain brand consistency, and reduce human intervention. Transparency through storytelling—as illustrated in the McKinsey AI Agents report—increases accessibility and trust. These efforts ensure that AI enhances engagement while reinforcing brand reliability.
​

​AI platforms streamline development, improve accuracy, and unlock scalable automation potential

​New agentic platforms and UX redesigns are reshaping how firms compete and integrate AI

The release of agentic platforms like OpenAI’s Codex CLI and Google’s Gemini 2.5 Flash reflects a wider market shift toward autonomous, decision-capable systems. Microsoft’s Copilot and Semantic Kernel further support AI agent navigation and collaboration across digital systems, setting a new standard for intelligent automation.
​
Integration issues and talent shortages hinder AI scalability and operational consistency

Integrating rapidly evolving models into legacy systems remains a core challenge. Developer bottlenecks, integration latency, and lack of agility stymie innovation. Mercado Libre’s solution—a unified platform called Verdi powered by GPT-4o—allows developers to build secure apps without touching source code. This streamlined infrastructure has yielded measurable efficiency gains, including 99% fraud detection accuracy and scalable localization of product descriptions.

Embedded AI workflows and structured prompting enhance speed, accuracy, and value creation

Operational excellence is driven by strategic use of AI in workflows. OpenAI’s automation platform processes hundreds of thousands of tasks monthly, reducing repetitive work. Structured prompting strategies from GPT-4.1 enable long-context, agentic workflows that reduce error and increase speed. These innovations support scalable value creation through automation and intelligent task delegation.
​

​Structural changes and skill development are essential to sustaining enterprise-wide innovation

​Decentralized AI development and agent-first design drive speed and collaboration

In response to evolving needs, companies are undergoing structural transformation. BBVA’s global deployment of ChatGPT Enterprise decentralizes AI development, empowering teams across departments to create use-specific GPTs and cut project timelines from weeks to hours. Simultaneously, firms are rethinking UX to align with agent-first workflows, supporting seamless cross-team collaboration and faster execution.
​
Centralized control and lack of access prevent widespread AI adoption and experimentation

Transformation is still hindered by cultural resistance, technical skill gaps, and access inequities. Many organizations rely heavily on centralized IT teams for AI development, limiting experimentation. Premium access models exclude smaller teams and non-technical staff from innovation opportunities. This creates a divide between strategy and implementation, stalling widespread adoption.

Upskilling and inclusive tools empower teams to lead innovation from within

To foster innovation, organizations must promote inclusive skill development and experimentation. Equipping employees with tools to create GPTs—as done by BBVA—enables localized problem-solving and faster execution. Structured training in agentic workflows, paired with tools that save hours on complex tasks, cultivates a culture of continuous learning. With the right foundation, innovation becomes systemic rather than situational.
​