Nitin Vagulde from VeloDB explains the reasons for AI agent failures after the demonstration phase

The Nasdaq index significantly declined after its record close on Thursday, losing more than 10%. This movement suggests a potential correction driven by concerns regarding massive artificial intelligence expenditures and increased scrutiny over data center expansion.

The composite Nasdaq index, which is heavily reliant on the technology sector, fell nearly 4% during trading in New York. This drop exceeded 10% from the record closing high reached on October 6th.

Thursday's decline erased previous daily gains as Democrats won both chambers of Congress. Investors began assessing the potential for closer attention to further infrastructure build-out for AI.

Samir Samana, Senior Global Market Strategist at Wells Fargo Investment Institute, noted that the transition from complete unified control by one party to complete unified control by another can cause significant uncertainty, disruption, and sharp policy shifts.

He added that when it seemed Democrats would gain a significant majority in both the House and the Senate, analysts anticipated at least a temporary pause in the markets.

The 10% drop from the last record high is widely viewed as a correction, marking the second Nasdaq correction this year. Sentiment toward AI-related stocks has worsened in recent weeks, and Nasdaq has lost over $2.5 trillion in market value in October.

Several large technology companies are tapping debt markets to finance multi-billion dollar investments planned for data centers, chips, and other necessary infrastructure for AI development.

Investors are increasingly questioning the speed at which these investments will translate into profits, especially against the backdrop of persistent geopolitical tensions in the Middle East and fears of prolonged inflation, which sustain high borrowing rates.

Reporting published last month by Alphabet and Meta, two of the largest hyperscalers building and operating the world's largest data centers, failed to reassure investors about the return on AI spending, contributing to a broader sell-off of AI-related assets.

Despite the growing availability of artificial intelligence models, turning them into reliable commercial products remains a complex task, especially when enterprises face fragmented data infrastructure.

At the TechSparks 2026 event, Michael Cronin, Vice President of Sales for Asia Pacific, Japan and India at Couchbase, participated in a discussion with Shivani Mutanna, Senior Director of Strategic Content and Video at YourStory. They discussed the transition of AI from the experimentation stage to full production, management challenges, customer experience, and the increasing role of India in AI innovation.

For decades, large corporations have implemented various databases and highly specialized solutions to solve specific problems. While each solved an immediate issue, the result was often a fragmented data layer composed of multiple systems that interact poorly with each other.

Cronin considers this fragmentation a serious obstacle to AI adoption. He cites industry estimates suggesting that 70–80% of AI development projects never reach the production stage, and MIT research showed that 95% of such projects do not become profitable.

Couchbase's operational data platform is designed to address this problem by helping companies consolidate their data infrastructure. Instead of trying to fix everything at once, Cronin advised companies to start small, recommending that founders and enterprises struggling with data layer fragmentation first choose the most critical use cases.

He cited the example of UK retailer Tesco, which has a large Global Capability Center (GCC) in Bengaluru. Many of this retailer's systems negatively impacted the customer experience, especially during peak shopping periods like Black Friday and weekends. Tesco began consolidating its data layer, starting with one instance and one use case before expanding efforts across its entire technology stack. This consolidation allowed the retailer to process millions of transactions per second with less than a millisecond latency. In Cronin's view, this example demonstrates how companies can begin solving the fragmentation problem through one use case before scaling changes across their entire operation.

The pressure associated with developing AI products has prompted many enterprises to prioritize models and applications. However, pilot projects are often conducted on small datasets in controlled environments, which does not prepare organizations for real-world production demands. Cronin pointed to the growing scale of machine traffic. He stated: 'In eighteen months, machine traffic surpassed human traffic on the internet. That is the scale of AI. That is the scale of machine speed at which businesses must operate.'

When AI applications encounter real-world workloads, fragmented data systems can prevent them from functioning reliably at scale. The increasing use of AI agents exacerbates this issue. As agents become more autonomous and take on decision-making functions, enterprises must ensure they can access the necessary data within appropriate governance structures.

Fragmented systems can hinder communication and management, limiting agents' ability to act autonomously. Cronin also urged enterprises to reconsider their approach to allocating AI investment. While models evolve rapidly, corporate data remains a critical asset. He emphasized: 'Models can be rented. The best model today is yesterday's newspaper. The true asset for an enterprise is data.'

As enterprises began to realize that fragmented data was slowing down their AI initiatives, many responded by adding APIs and connectors to link existing systems. However, Cronin warned that this method could create additional complexity instead of solving the core problem. Every extra connection increases latency, risk, and creates another potential point of failure.

Instead, he recommended consolidating the data layer so that AI applications and agents could interact with it more directly, bringing data closer to the customers. According to Cronin, many enterprises, including large GCCs in India, are currently re-architecting to eliminate data fragmentation. Without these changes, further investments in AI development may lead to the same problems.

Data consolidation can also improve customer experience by helping systems handle growing demand, remain available, and respond quickly. As customers expect more from AI-based services, having data at the edge is becoming increasingly important.

Cronin noted that business cultures vary across the Asia Pacific region, but fragmented data infrastructure remains a common challenge for AI adoption. At the same time, India is emerging as a vital hub for AI innovation and decision-making. He noted that 80% of Couchbase's top 50 global customers have a significant presence in India, with much of their AI innovation and leadership based in the country. Couchbase is also expanding its presence in India, moving its Asia Pacific headquarters to Bengaluru. Among Couchbase's corporate clients in India are Air India, Airtel, BigBasket, and Mahindra Finance. In Cronin's opinion, India's role in global business extends beyond cost savings; innovation, leadership, and important business decisions are increasingly happening there. He mentioned Air India's next-generation entertainment platform, which is built on Couchbase. This platform aims to provide passengers with a personalized shopping and entertainment experience even without network connectivity on board the aircraft. For Cronin, this project illustrates how Indian enterprises are using technology to enhance customer experience in real-world operations.

Although artificial intelligence is capable of generating code at unprecedented speeds, the process of testing, verifying, and deploying it into a working environment remains a separate complex task. Engineering teams are increasingly using AI agents to write code, fix bugs, and create new features, but faster coding does not guarantee faster software delivery.

As code generation gains momentum, the volume of checks increases, testing slows down, and teams find it difficult to coordinate the growing amount of changes.

At the TechSparks 2026 conference, Tushar Matur, CEO of Forge Code, highlighted this issue in his session 'Anatomy of a Software Factory.' He explained how AI-focused engineering organizations can move beyond automating individual coding tasks and start creating systems that manage the entire Software Development Life Cycle (SDLC).

Matur noted that just because code generation has become easier and faster does not mean the overall SDLC will accelerate. He reminded that the traditional development approach focused on writing code efficiently, requiring teams to define requirements, prepare specifications, and design architecture to reduce the load on the implementation phase.

However, since AI takes on much of the coding work, bottlenecks shift to other areas. Matur described a situation where organizations face a huge number of pull requests awaiting review, longer testing cycles, and more errors than existing regression suites can detect.

He emphasized that solving the problem does not necessarily involve hiring more engineers or launching more AI agents. Additional staff can increase coordination costs, and adding agents without effective feedback mechanisms can lead to code that overwhelms the verification and testing systems.

'The reason for this is the lack of signal,' Matur explained, clarifying why simply connecting automated steps does not create an effective software factory.

Matur presented Forge Code as a software factory that coordinates AI agents throughout the development cycle. The process begins with sorting requests, creating specifications, and implementing features. Then, agents check the changes before they are deployed, with human engineers intervening when decisions or feedback are required.

The system also monitors in the working environment, using the gathered data to inform future tasks. This forms a feedback loop, allowing the factory to identify issues, improve its workflows, and generate new work based on the experience gained. During the presentation, Matur reported that Forge Code used about 100 agents in the background to work on features, compatibility, and bug detection.

He outlined six principles for building an effective software factory: autonomy, scalability, cost-effectiveness, measurability, self-improvement, and alignment with business goals. The last principle is particularly important.

The factory should not operate as a task of writing code for every request, but rather be aimed at achieving specific outcomes, such as increasing customer retention or meeting regulatory compliance. Teams can then prioritize features based on the value they bring. In this model, engineers spend less time managing each task and more time observing results, analyzing signals, and improving the system itself. Business leaders can also provide requests and priorities directly, allowing the factory to translate organizational goals into a development backlog.

Code writing speed is good, but creating better software is the real victory. For engineering teams, the key is transforming AI-driven speed into reliable products, measurable progress, and real business results.