Data management 2025 predictions: GenAI changes data

1. Data readiness will remain a priority

Data readiness has been a consistent theme over the last few years, with companies pouring money into data governance, privacy regulations and clean, streamlined data sets. Preparing a solid data foundation is essential for generative AI because it needs accurate, diverse and governed data to produce effective results, especially when using enterprise data with a large or small language model.

For the first half of 2025, most organizations will focus on achieving data readiness for their generative AI projects. Research from Informa TechTarget's Enterprise Strategy Group showed that, on average, organizations have 15 or more generative AI tools they plan to build and bring to market. Each has their own unique set of data to support their initiatives. Data readiness will continue to be an ongoing effort for each initiative, along with updating data and adding new data sources, including third-party data sets, to drive new insights. Third-party data will become more prevalent once generative AI applications launch and organizations want to add more diverse data.

The focus will not just be on making data "ready" but making it "AI-ready." AI models will need contextualized, annotated and accessible data in real time, which will accelerate the adoption of tools that can automatically tag, catalog and clean data for AI consumption. By mid-2025, we should see some tangible market penetration.

2. Internal generative AI applications lead the way

In 2025, the early adoption of generative AI will revolve around internal productivity and efficiency. Businesses will have AI copilots and virtual assistants that cater to the needs of employees rather than external users. AI can create a safer, less open environment to experiment, eliminate the possibility of public-facing failures and let enterprises optimize generative AI services before they are available to customers. AI will help automate mundane and repetitive tasks, some of which we are already seeing today:

• Document automation. AI models will analyze applications (mortgage, credit, loans and insurance) and generate first drafts of reports, proposals and internal communications, freeing employees to focus on higher-level tasks.
• Meeting summaries and note taking. AI copilots will automatically summarize meetings, flag action items and track decisions, significantly reducing administrative burdens.
• IT and development support. Developers will increasingly rely on AI assistants to write code, debug issues and generate technical documentation.
• HR and talent management. Generative AI will generate personalized learning pathways to assist in talent acquisition, onboarding and employee training.

An internal-first approach will be especially appealing to industries with strong regulatory requirements because it enables controlled testing and refinement of AI systems without risking reputational harm. By mid-2025, internal-facing copilots will be essential productivity boosters.

3. From internal to customer oriented

Once companies are confident in their internal deployment of generative AI, they'll start deploying more customer-facing applications in the second half of 2025. By now, most companies have tackled the issues of model reliability, transparency and bias -- the three challenges that customers need to face when using AI applications.

For customer-focused tools, we should see more advancements with three types of applications:

• Personalized customer service. AI-powered chatbots and virtual agents will shift from simple Q&A guides to conversational agents that provide human-like service, contextual responses and real-time sentiment.
• Product recommendations. Online retail and online e-commerce brands will use generative AI to provide highly personalized recommendations for each user's actions, leading to higher customer satisfaction and conversions.
• Content and marketing personalization. From ad copy to email campaigns and social media posts, generative AI will empower marketers to rapidly create, experiment and optimize messaging for specific customer groups to ensure better customer experiences.

It will not be an easy transition from internal to external use. Businesses will have to impose greater limits on ethical AI use, explainability and data privacy. The "crawl, walk, run" strategy will prevail as enterprises deploy controlled pilots and phased introductions of end-user AI applications driven by enterprise data.

If you want to build a customer service application, you will need the purchase, search and support history of the customers. For example, healthcare companies might need a long history of medical records. Data readiness is especially crucial in the phase of making generative AI public-facing.

4. Ethical AI and trust will drive adoption

When generative AI enters enterprise workflows, questions of AI ethics, explainability and accountability will come to the forefront. In 2025, regulators and industry groups will develop more rigid regulations around the ethical application of generative AI. The new U.S. Administration is already talking about this, and we have the AI Acts in the EU and California, with more to come, focused on the ethical use of AI and consumer protections. Regulations will focus on three main areas:

• AI model transparency. Companies must be transparent about how their generative AI models make decisions, especially in the financial, medical and legal markets.
• Bias and fairness. Organizations will implement tools to flag and reduce bias in AI-generated outputs, including hiring, lending and customer services.
• Data privacy and consent. As generative AI applications harvest increasing volumes of customer data, organizations will need to gain permission from users to process data and comply with more demanding privacy laws such as GDPR, CCPA and new AI-specific regulations.

Organizations need strong data management practices to create trusted data sources that meet regulatory standards, protect privacy and achieve the results organizations want to see from using enterprise data in generative AI tools internally and externally.

5. Business data will be the heart of AI architecture

Organizations will realize that their most valuable asset in 2025 and beyond aren't their AI models -- it's their data. Everyone will have access to AI tools, large and small language models, which affects the quality of generative AI tools. The big differentiator is the contextual, proprietary data that each enterprise uses to build its unique generative AI tool. Quality data is a competitive advantage, enabling the ability to create unique generative AI experiences and advance innovation.

2025 will be a measured, but transformative year for data and AI. Although ethical AI, data privacy and bias issues continue to loom, companies committed to responsible AI development will be in the best position to take advantage of increasing opportunities. Enterprise data will be the catalyst for AI-based value creation to change from a "back-end value stream" to a "front-line enabler."

We'll remember the year 2024 as the start of a new era for generative AI -- one in which data, trust and ethical AI made the market winners in this ever-evolving space. If companies take a little time to prepare for data today, the second half of 2025 could be unprecedented.

Stephen Catanzano is a senior analyst at Informa TechTarget's Enterprise Strategy Group, where he covers data management and analytics.

Enterprise Strategy Group is a division of Informa TechTarget. Its analysts have business relationships with technology vendors.