How AI impacts the build versus buy decision for data marketplaces

Artificial intelligence has already had an enormous impact on how enterprise software is created, deployed, and maintained. Generative AI can be used to write software code from scratch, without requiring any programming skills. Agentic AI can create end-to-end processes that replace the need for specific software programs.

All of this is causing organizations to take a fresh look at how they meet their software needs, changing the balance between building in-house and buying from an external vendor. In this blog, we look at how AI impacts build versus buy calculations when it comes to data product marketplaces.

AI makes it easier for anyone to code and create software solutions without requiring technical training, especially through generative AI and agentic AI.

Generative AI

In the same way that users can prompt generative AI tools such as ChatGPT and Claude to provide information, they can also prompt them to write code to meet a specific need. While outputs clearly vary depending on the quality of the prompt engineering used to generate them, this dramatically reduces the time (and resources) required to create code.

Agentic AI

Agentic AI links together multiple AI agents to automate processes and decision-making in specific areas. By tapping into data in underlying systems this can remove the need to create an application at all, replacing it with an agentic workflow.

Data product marketplaces provide a single, centralized space for sharing data with the entire business. An intuitive, e-commerce style experience builds trust and confidence with business users, driving greater usage, while data product owners can easily publish and track consumption of their data products by humans and AI. 

When it comes to deployment, organizations have to make a choice about whether to build their data product marketplace in-house or partner with an external, specialist vendor. 

We have covered common misconceptions about the build versus buy debate in a recent blog. Building on this, how does the rise of AI impact this vital decision? For example, does the ability to create code more easily in-house mean that organizations should now look to build their own data product marketplaces, rather than buy a SaaS solution?

Based on our market experience and analysis, there are four reasons why for the majority of organizations, buying not building is the best option for long-term data product marketplace success

1 Value comes from experience, not code

The process of coding any solution can now be automated. However, coding is only around 20% of the work involved in creating software. The bulk of development involves testing, reviewing, and quality assurance. Companies need to understand technical and functional specifications, measure adoption, write documentation, and plan workflows. Optimizing the 20% of time involved in coding doesn’t remove the fact that most organizations have never built a complex tool such as a data product marketplace. This means they don’t have the skills or experience to understand user needs and design an experience that meets them.

2 Solutions still need to be maintained and developed

Once a solution has been created it then needs to be supported, maintained, and updated, over the long-term. This means adapting the solution and adding new features in line with user requests and requirements. While AI can help with coding, future development needs humans in the loop to work with users, understand their pain-points, and deepen functionality. This adds to resource requirements and costs. By contrast, buying a solution from a specialist vendor provides access to a roadmap of new features that are developed to meet the needs of all customers, reducing time to market and removing technical support and maintenance overheads. 

3 Scalability and security remain vital

Whether built in-house or externally, data product marketplaces have to be able to scale to cope with enormous volumes of data and potentially thousands of concurrent users. Building a solution with AI doesn’t remove this scalability challenge – or the need to securely protect data and meet compliance requirements. Cloud-based external vendors offer the ability to scale seamlessly to deliver for millions of users without impacting performance, while protecting data through the powerful security capabilities provided by hyperscaler partners. Trying to replicate this in-house through on-premise solutions will strain existing IT resources and further add to costs.

4 The need to deploy internal AI where it delivers greatest value

Organizations have a long list of potential AI use cases. Choosing which to pursue and where to focus internal resources is therefore critical. While data product marketplaces clearly deliver great value, as discussed, building them in-house does not increase this value. Instead, working with a trusted external partner enables limited internal AI resources to be better deployed on complex AI use cases that have to be completely tailored to the organization itself. 

For most organizations, the rise of AI doesn’t change the decision on buying, rather than building, a data product marketplace, in order to maximize ROI and accelerate deployment and adoption. 

Instead, artificial intelligence introduces new criteria that must be taken into account when choosing a vendor. You need a partner that is embracing AI, has a strong roadmap of future innovations, and above all offers value that goes beyond the software code they provide. Look for a vendor that offers these five capabilities:

Incorporates AI to increase engagement and usage

Features such as AI-based search, integrated AI exploration agents, and a built-in MCP server are all crucial to connecting your data with human and AI users. These maximize the impact of your data, and your data product marketplace. For example, Huwise’s Huwy AI exploration agent lets users ask questions in natural language and receive reliable, contextualized answers in seconds—with no technical expertise needed. As it is connected to the Huwise MCP server, Huwy explores your actual data—not just the metadata—to accelerate access to insights.

Focuses on AI to improve efficiency, administration and data product development 

Through automation AI can dramatically reduce workloads for marketplace administrators and data product owners. AI can help through capabilities such as assisting in completing metadata to very complex agentic workflows. For example, the marketplace’s AI should be able to take a selected dataset, create visualizations and ensure that it is discoverable, or even automate the creation of a data product. This frees up time to focus on talking to users and selecting the best data sources to meet their specific business needs.

Harnesses AI to accelerate its own development

AI is fundamentally changing the programming process. It accelerates development and enables vendors to create optimized features that add value more quickly. Therefore ensure that you work with a marketplace vendor that is clearly embracing AI internally, both for development and to aid marketplace deployment. They need a comprehensive vision and plan to standardize AI across the delivery chain to maximize the value that they deliver to your business. 

Offers access to experience across sectors and countries

As coding is commoditized, added value is delivered by the service and experience provided by the vendor. Focus on vendors with a strong track record and a large customer base with low levels of churn – this enables you to learn from your peers and accelerate the adoption of best practice. Demonstrating this, Huwise has 15 years of experience, over 350 customers in 25 countries, and has worked on more than 3,000 data marketplaces, providing skills and knowledge to support the success of all of its clients.

Understands and is close to the data

Being able to access and share your data across the business is central to organizational success. That requires a partner that truly understands data, how it flows, and how it can be harnessed by you. Your partner should be close to the data with a platform that supports access to all data sources, provides data in relevant formats, and underpins complex and business critical workflows. The platform must be an enabler to build use cases for data sharing within the company, allowing data and metadata to flow smoothly around the organization. This is something that cannot be quickly replicated with internal AI coding.

The successful deployment of AI relies on access to reliable, trustworthy and understandable data. This accelerates the need for businesses to deploy comprehensive, self-service data product marketplaces to deliver this data for LLMs, generative, and agentic AI. Despite the growth of AI for coding, choosing an external vendor still remains the best option for most organizations in terms of features offered, access to best practices and minimizing required resources. What has changed are the capabilities to look for when choosing your supplier – you need an AI-enabled partner that has the experience and understanding to meet your requirements, now and in the future.