The ROI of Ambiguity: How Great AI Can Emerge from Vague Questions
Companies love clarity. Define the problem, find clean data, choose the model, deploy, measure – that’s been the standard playbook for developing enterprise AI. It keeps projects on time and risks low, but it also keeps innovation boxed in.
When you use GPS to get to work, you often follow the same suggested route every day. It’s efficient and familiar, but it also means you rarely explore alternative paths unless prompted by traffic alerts or detours. You may never discover a faster, safer, or more scenic route. The same thing happens inside organizations pursuing AI projects. Teams often stick to the familiar, chasing well-defined use cases and predictable results. But this comfort with the known can blind them to more innovative or disruptive possibilities, ones that may only emerge by questioning assumptions, experimenting with uncertainty, and venturing beyond the beaten path.
The most transformative advances in AI didn’t come from precise business cases or ROI-driven roadmaps. Breakthroughs like AI copilots and LLM search assistants emerged because someone asked a question that didn’t have a clear answer and was willing to explore the gray space between business need and technical potential. Ambiguity can feel inefficient or even risky, but it is the raw material of innovation. While structured analytics optimize what’s already known, ambiguity opens the door to entirely new ways of creating value.
(Taris Tonsa/Shutterstock)
Defining Ambiguity in the Context of Enterprise AI
Ambiguity is frequently misunderstood. While most associate the term with confusion or lack of focus, it’s actually a deliberate exploration of questions that aren’t fully formed yet, datasets that are unstructured or difficult to model, and domains where value is real but not yet measurable. A better term might be “strategic ambiguity.” Strategic ambiguity enables organizations to uncover unmet needs and unknown unknowns. It allows them to experiment in spaces where customers, markets, and even competitors have not yet defined demand. In that sense, ambiguity isn’t a gap to close, but a frontier to explore.
Despite its potential, ambiguity runs counter to most corporate instincts. Large organizations are structurally and culturally wired for precision. Clear timelines, defined KPIs, and measurable ROI make project planning easier and executive approvals smoother. Plus, most tests are run by data scientists, who are predisposed to using a scientific method that values repeatability.
However, this precision bias has a cost. It pushes organizations toward incremental use cases (churn prediction, lead scoring, process optimization, etc.) – projects that are safe, measurable, and fundamentally limited in impact. When every AI initiative must justify its existence through a near-term ROI lens, innovation narrows. AI becomes another reporting mechanism rather than a transformation engine.
(Collagery/Shutterstock)
The GenAI Factor
If you’re a manager on the hook to prove ROI, all this talk of unpredictability might make you uncomfortable. But fear not, because the large language models and multimodal systems that power GenAI are built to handle ambiguity, and don’t require perfectly structured data or tightly-framed questions. Rather, they thrive on open-ended prompts and fuzzy context. Here are a few ways of positioning the concept to appeal to any skittish stakeholders:
- Reframe ambiguity as a feature. This fundamentally shifts how enterprises interact with information. For example, in sectors like law and compliance, where data is text-heavy and nuanced, teams can now explore questions that defy clear articulation. Ambiguity, far from being a flaw in the process, becomes the mechanism that unlocks cross-disciplinary understanding and previously inaccessible insights.
- Treat ambiguity like a shortcut to insights. Traditional analytics demand exhaustive data preparation and clearly defined objectives before any value can emerge. GenAI reverses that order. It allows teams to engage in rapid hypothesis generation that can test ideas with data before investing in full-scale model development. With this framing, ambiguity becomes a creative accelerant that can spark new ideas, uncover creative solutions, and expose unique variables that structured models overlook.
- In marketing, ambiguity unlocks creative potential. Creative teams have traditionally relied on intuition and precedent to guide messaging and design. GenAI changes that. It enables marketers to explore dozens of visual and messaging directions, many of which outperform human-created versions in real world tests. AI-generated content allows for experimentation at scale, revealing ideas that resonate with micro-audiences, reflect emerging trends, or challenge conventional assumptions. Ambiguity in this context is not noise, it is signal.
The ROI of Exploration
Traditional ROI frameworks struggle to capture the value of early-stage AI exploration. When outcomes are measured purely in revenue or cost savings, the learnings that precede those outcomes disappear.
(Source: aniqpixel/Shutterstock)
Organizations need alternative metrics that acknowledge the strategic value of exploration. The real question becomes: How many viable pathways did the project open up? How many new questions or hypotheses did it generate? How quickly did the team learn what truly matters, even if the immediate deliverable wasn’t production-ready? The most meaningful progress often comes from what is uncovered along the way. Did the effort help surface new opportunities, steer the team away from unproductive directions, or deepen their understanding of customer needs? And most importantly, what did the team learn that they didn’t know before? In GenAI-driven organizations, this kind of insight and iteration becomes the true signal of momentum.
By reframing ROI around exploration, discovery, and learning velocity, leaders can justify investments in ambiguous work as foundational data. The returns may not appear on the next quarterly report, but they shape the organization’s long-term capacity to innovate.
How to Make Ambiguity Operational
Embracing ambiguity requires building a different kind of structure that’s designed for exploration. To operationalize ambiguity, consider the following:
- Create safe spaces for ambiguity. Set up dedicated sandboxes or internal labs where teams can explore ideas within clear timeframes. Encourage them to design with questions, not fixed requirements, so they can test early signals, follow curiosity, and surface unexpected insights.
- Assemble mixed teams. Create mini-pods that bring together domain experts, data scientists, and GenAI specialists to drive fast experimentation. This structure encourages collaborative inquiry over linear thinking, helping teams uncover better questions and faster paths to insight.
- Track learning, not just delivery. When learning is defined as a core output, knowledge creation can be treated as its own measurable KPI.
- Evolve governance to support uncertainty. Introduce flexible processes like early-stage funding or innovation councils that allow experimentation while maintaining compliance. Use narrative as a governance tool. Instead of heavy review decks, ask teams to share short, one-page stories: What did we try? What surprised us? What did we learn? What should we do next? These narratives surface real insight and keep exploration aligned to business goals without slowing progress.
It’s also helpful to keep in mind that ambiguity can be especially valuable in environments that lack clarity, like: (a) new markets, where consumer behavior, regulations, and demand signals are still in flux; (b) voice-of-customer (VOC) analysis, where the richest insights often hide in unstructured, informal feedback; and (c) early-stage GenAI exploration, where use cases aren’t fully defined yet. In each of these settings, demanding clarity too early can close off the very discoveries that drive differentiation.
(SuPatMaN/Shutterstock)
Creating a Culture of Experimentation
When Harvard Business School compared the stock performance of companies with a strong experimentation infrastructure (Amazon, Facebook, Google, etc.) against the performance of the S&P 500 over the past 10 years, they discovered that these “experimentation organizations” consistently and significantly outperformed the S&P. So, while precision might satisfy senior leaders looking for predictability, it’s clear that having an experimentation culture is increasingly necessary to staying competitive.
Google is a classic example of a company that has institutionalized ambiguity. Through its famous “20% time” policy, employees spend most of their week shipping core products, but have a protected slice to explore open-ended ideas. That culture has produced new business lines and capabilities, from AdSense and Google News to features like wheelchair-accessibility in Google Maps, with many of these starting as side projects before becoming part of the core. In other words, Google didn’t just tolerate ambiguity, but gave it a formal slot on the calendar.
To realize the ROI of ambiguity, organizations must cultivate this new cultural mindset. Teams need permission to stay in problem-finding mode longer and prioritize questions over answers, and executives must learn to value experiments that reveal possibility even when they don’t lead directly to production.
Remember, not all ambiguity is productive. Exploration must still be guided by intent and aligned with strategic objectives. But rejecting ambiguity entirely is far more dangerous, and can close off pathways through which the next major breakthroughs may emerge. In the age of GenAI, asking vague questions might just be the only way to find the right ones.
About the Author
Boobesh Ramadurai is Vice President of LatentView Analytics, where he leads customer-facing analytics teams focusing on revenue growth and client success. He has delivered data-driven solutions across industries, including Technology, Financial Services, Travel, Retail, and Media & Entertainment. With hands-on expertise in analytics, digital marketing, and visualization, Boobesh manages global teams across data engineering, analytics, and data science. He has built and scaled onsite-offshore models and advises clients on topics like GenAI in marketing, B2B sales, the Creator Economy, and subscription models.
Related
Nvidia Releases Nemotron 3, Expanding Its Open Models for Agentic AI
Nvidia has released Nemotron 3, the latest of the company’s open reasoning models designed to…
Untamed Data Is Undermining the AI Revolution
Across industries, organizations are drowning in unstructured data: files, videos, images, chat logs, design documents,…
As AI Scales for Science, the DOE Turns to Nuclear and Federal Land
Data centers running advanced AI systems now require steady and high-density energy at a scale…
A (Mostly) Non-Technical AI Primer
I have studied and watched artificial intelligence grow over the last forty years. Like many,…
This Week in AI: Nvidia in China, OpenAI’s Disney Deal, and Federal AI Oversight
This week’s AI news cycle captured where the AI sector stands as 2025 draws to…
Unconventional AI Wants to Solve AI Scaling Crunch with Analog Chips. Will It Work?
Serial entrepreneur Naveen Rao lifted the curtain a little bit this week on his latest…
The ROI of Ambiguity: How Great AI Can Emerge from Vague Questions
Companies love clarity. Define the problem, find clean data, choose the model, deploy, measure –…
Google Cloud Backs Genesis Mission with Gemini for Government and Secure AI Infrastructure
Dec. 23, 2025 — The White House’s Genesis Mission has set a bold ambition for the…
AWS Backs the Genesis Mission with Cloud and AI Infrastructure for National Labs
Dec. 22, 2025 — The Department of Energy recognized Amazon Web Services (AWS) as a supporter…
Anthropic to Support DOE Genesis Mission with AI Tools for Energy, Biology, and Research
Dec. 22, 2025 — Anthropic and the US Department of Energy (DOE) are announcing a…
DOE Announces Genesis Mission Collaboration Agreements with 24 Organizations
WASHINGTON, Dec. 19, 2025 — The U.S. Department of Energy (DOE) has announced agreements with 24…
Accenture Partners with DOE to Support Genesis Mission Program
ARLINGTON, Va., Dec. 19, 2025 — Accenture Federal Services, a wholly owned subsidiary of Accenture,…
