From Data to Decisions: How Synthetic Data, Custom Agents, and Human Co-Creation Will Shape the Future

At a time when technology is more powerful and more accessible than ever, we must remember that it’s a means to an end, not the end in itself, and focus on what it empowers people to do.

Marketing and insights teams are faced with the challenge of balancing efficiency and effectiveness by doing things right and doing the right things. That’s a tall order, as teams continue operating with tighter budgets and fewer resources. Agentic AI has been hailed as a game-changer, and it can be, but only when it increases efficiency and efficacy.

It’s telling that Gartner recently reported that they expect more than 40% of agentic AI projects to be canceled by the end of 2027, citing unclear value, high costs, and immature technology as a few of the key barriers. This echoes a point I made a few years ago, in an article on what I then called the paradox of insights tech: “insights tech is essential, but it’s not the point.”

The industry has come a long way since I first shared that perspective, but the core principle still holds. The risk of many AI initiatives isn’t that the AI won’t work, but that teams will lose precious time trying to retrofit AI into existing processes where it doesn’t belong, instead of focusing on where the technology adds clear value.

This is a challenge for teams under pressure to get sharper about where and how AI can effectively serve them.

It’s no surprise that the first half of 2025 has served as a testing ground for AI in the insights space. However, amid a flood of announcements and bold claims, distinguishing real progress from hype remains a challenge. So, how can marketing and insights teams make sure that their AI implementation works with them to drive smarter, faster business decisions?

The Next Evolution of AI-powered Insights 

Commentary from marketing and insights leaders and industry indicators all point to the fact that the next evolution in AI for the insights industry isn’t just about faster outputs and complete automation, but about smarter, more strategic collaboration between people and technology that will lead to better outcomes.

We can anticipate seeing increased momentum around three powerful shifts:

Human-AI co-creation

With deeper integrations, AI is evolving beyond a productivity tool and is becoming a genuine thinking partner. When connected to trusted data sources, AI can help teams frame the right questions, explore potential solutions, and uncover new strategic directions. For example, building an effective AI Assistant for innovation isn’t a one-off effort, it requires iteration and crucially, collaboration.

The way to create truly effective AI solutions is through close, ongoing collaboration between technology providers and end users. This partnership ensures that solutions are designed to fit into existing workflows rather than forcing customers to adapt to the AI. When AI is purpose-built to the team’s context - embedded to improve established ways of working, tuned to language and logic - it provides impact not hurdles.

Human feedback and collaboration can help identify a balance between which parts of the process require automation or a faster starting point for productivity.

Synthetic data

The promise of synthetic data is fast becoming a reality, with AI enabling scalable, research-backed consumer insights that can support decision making, particularly in marketing and innovation workflows where often early ideas and concepts are built on outdated or gut feeling, and not leveraging consumer insights. In fact, research recently published in Harvard Business Review found that 81% of respondents already use or plan to use GenAI to generate synthetic data

From simulating market responses, to scenario planning, to building dynamic consumer personas, synthetic data unlocks powerful new possibilities. Applied effectively, it can enable teams to stress-test ideas early, prioritize investments, and quickly assess whether further concept development is worth the time and resources.

However, as above, the value of synthetic data is only as strong as the data it is built upon. Generic tools may be able to generate surface-level ideas, but does it have access to all the data necessary to truly benefit? Without organizational context, such as connections to internal knowledge bases, these tools risk producing outputs that lack consistency, trustworthiness, and company alignment.

To deliver real impact, synthetic data must be grounded in robust, relevant sources and tuned to the specific needs of the organization. When powered by the right foundations, synthetic data presents a significant opportunity to reduce risk and waste but also to accelerate the delivery of sharper, more relevant concepts to the market.

Customizable agents

In addition to general-purpose Assistants, we’re seeing a growing focus on bespoke intelligent Assistants with agentic capabilities that deeply understand an organization’s unique context and are able to produce valuable outputs.

Each of these agents go beyond basic support and are trained to speak a brand’s language, align with nuanced business objectives, respect governance and privacy protocols, and adapt to the specific needs of different brands, categories, regions, and markets within complex organizations.

What Comes Next

We're in the midst of a complete redefinition of what insights work means, and it’s imperative that marketing and insights teams remain the central nervous system that helps organizations navigate complexity, uncertainty and opportunity in real time.

The future is about enabling faster, smarter decisions that promote strategic business outcomes. When AI is tailored to an organization’s specific context and tuned to a team’s language and logic, it becomes a multiplier, not a blocker.

The next wave of innovation will come not only from how GenAI evolves, but also from how teams put it to work.

One of the best parts of my job is getting to shape the future of insights alongside the people who actually use them. And while it has historically been about finding the right insights from the data, in this new era our accountability should extend to seeing those insights truly drive impact.