AI Strategy for Business Impact
AI is evolving so quickly that daily news and “news bites” can make it difficult for executives to keep up.
In this environment, an effective AI strategy requires disciplined choices: focusing on real business needs, setting realistic expectations, and understanding where generative AI fits versus other AI approaches.
This article is adapted from a video interview conversation in the engatica interview series featuring Andreas Welsch, Chief AI Strategist at Intelligence Briefing, discussing AI in business, generative AI, and the metaverse.
Executive Summary
- AI momentum is shifting from automation and insights to content creation via generative AI.
- Leaders should avoid “technology looking for a problem” and start from business needs.
- Generative AI improves accessibility, cost, and speed for producing first drafts of content.
- Risks include intellectual property ambiguity and convincing but inaccurate outputs.
- Hands-on trials (often free) help leaders make informed decisions and guide adoption.
Key Takeaways
- Andreas Welsch explains that earlier AI waves sometimes underdelivered due to missing processes, unrealistic expectations, insufficient data, and weak business–technology collaboration.
- AI has become pervasive in modern software—often working “in the background” to improve efficiency and effectiveness.
- Generative AI creates new momentum by enabling content generation beyond traditional automation and analytics.
- Business value today is largely productivity-driven (e.g., long-form text, images, and faster content production).
- Leadership must monitor IP and ownership debates as models are trained on public web data and lawsuits emerge.
- Large language models can produce plausible-sounding inaccuracies, requiring fact checking and human oversight.
- Metaverse experimentation is underway, but broad adoption will take time; generative AI could enable more dynamic personalization within immersive experiences.
What is AI strategy?
AI strategy is the leadership discipline of deciding where AI should be applied to create business impact, and how to do it responsibly. In this conversation, Andreas Welsch emphasizes starting with a real business need, confirming whether AI is the right approach, and then selecting the appropriate type of AI—such as generative AI for content creation versus other AI for automation or insights based on historical data. AI strategy also includes setting realistic expectations, ensuring data and processes are in place, and keeping humans involved to validate outputs when models can produce inaccuracies.
Why this conversation matters
The discussion is aimed at business leaders navigating rapid change across AI and adjacent trends. It highlights a practical leadership challenge: separating daily hype from durable value, while still moving fast enough to stay competitive.
For AI leadership and workforce transformation, the central message is pragmatic: when tools are available for free, leaders and teams can try them directly to build informed judgment, participate credibly in internal discussions, and make better decisions about adoption.
From AI hype to business reality: lessons leaders should retain
Welsch observes that AI surged as a “hype topic” around 2016–2017, promising large value. Many organizations invested, ran proof-of-concepts, and built platforms.
However, results were uneven. Welsch points to common blockers: processes not in place, expectations set too high, missing or inadequate data, and the ongoing challenge of collaboration between technology experts and business stakeholders.
Key Insight: AI programs often stumble not because AI is inherently ineffective, but because organizations treat it as a technology project instead of a business change effort—requiring clear problems to solve, realistic expectations, strong data foundations, and sustained business–technology collaboration.
AI strategy is becoming more pervasive in modern software
Welsch describes AI as increasingly embedded in everyday applications—so much so that users may not even notice it is present. In his view, AI is becoming part of the “fabric” of modern software and helps people complete tasks more efficiently and effectively.
This shift matters to executives because it changes how AI value shows up: not always as a standalone AI initiative, but as measurable improvements to workflows, customer experiences, and employee productivity.
Key Insight: As AI becomes embedded in common tools, leadership focus shifts from “building AI” to governing and guiding how AI-enabled capabilities are used—ensuring business relevance, appropriate oversight, and informed decision-making across functions.
Generative AI: the next major push for AI strategy
Welsch identifies generative AI as the next wave that goes beyond automating tasks or generating insights from historical data. Generative AI can create new information and content, which is why it has renewed momentum for AI discussions in business.
He also cautions against repeating past mistakes. The risk is treating generative AI as “a technology looking for a problem,” rather than first identifying where the business needs help and whether AI is the right tool.
Key Insight: Generative AI can reignite hype cycles unless leaders anchor decisions in business needs—deciding where content generation truly creates value and where other approaches (or no AI at all) are more appropriate.
What generative AI is (in simple terms)
Welsch explains generative AI as technology that creates content based on historical information used in training. Two highly visible forms include:
- Text-to-image generation: a user prompts the system to create an image (e.g., “a photorealistic image of a robot in a flower field” in a particular style).
- Large language models: prompt-based systems that generate long-form text (e.g., blog posts) by predicting the next word in a way that can sound human-like.
Where generative AI helps today: productivity and content velocity
Welsch frames near-term business value as productivity improvement. Generative AI can reduce cost and effort to produce initial ideas and drafts, potentially at a higher volume and useful quality compared to traditional approaches.
Examples he references include marketing use cases such as creating long-form blog content, using video avatars to produce videos more quickly, and generating images from text prompts.
Key Insight: The immediate executive opportunity is not “fully automated creativity,” but faster and cheaper first drafts—text, images, and video-like assets—that teams can refine, validate, and align to brand standards and business goals.
Risks leaders must address: IP ownership and factual accuracy
Welsch highlights two leadership risks that come with rapid generative AI adoption.
1) Intellectual property and ownership debates
Because many models are trained on public web information, Welsch notes it is not always clear what sources were used and who owns the rights to content that influenced outputs.
He points to emerging legal pressure, citing examples where stock image companies such as Getty and Shutterstock filed lawsuits, arguing AI outputs may be based on (or resemble) content from their databases.
2) Convincing but inaccurate outputs
Welsch warns that large language models can generate inaccuracies while sounding highly plausible. He shares a simple example: when asked whether cow eggs or chicken eggs are bigger, the model confidently answered that cow eggs are bigger—despite cows not laying eggs.
His implication for leaders is straightforward: at least for now, a human should remain “in the loop” to fact-check and validate outputs used in business contexts.
Key Insight: Responsible AI strategy must account for both legal ambiguity (ownership, training data, and resemblance) and operational risk (hallucinations). Governance is not optional when outputs can be persuasive yet wrong.
The metaverse: game-like origins, serious long-term potential
On the metaverse, Welsch describes it as a digital environment where people can immerse themselves and have experiences. He notes that the space emerged largely from gaming, even though the longer-term promise could be much larger—potentially comparable in significance to the internet over decades.
He also emphasizes timing: it will take time for the metaverse to become pervasive. Even so, some consumer brands and retailers have experimented, especially after Facebook rebranded to Meta and signaled its strategic focus.
Key Insight: The metaverse should be treated as an experimental frontier rather than an immediate mainstream channel; leaders can learn by piloting, while recognizing broad adoption will likely be gradual.
How generative AI could shape personalized experiences in the metaverse
Welsch connects generative AI to a future of more personalized customer experiences. He notes that today’s customer experience is fragmented across vendors and brands. If brands had more context about an individual, experiences could shift from one-to-many to one-to-one personalization.
In Welsch’s view, generative AI can help create more dynamic and personalized experiences by generating imagery, sound, music, text, and narratives on demand—potentially extending into immersive environments when needed “at the second or at the minute.”
Key Insight: The strategic link between generative AI and the metaverse is personalization at speed—content and experiences created dynamically, rather than pre-produced for broad segments. That prospect makes experimentation today a workforce readiness advantage.
Staying current: the simplest executive habit that scales
Welsch underscores that AI news volume is now overwhelming. His practical recommendation is to use the availability of free tools: sign up, try them, and build firsthand experience.
He argues that firsthand experimentation helps leaders and teams participate in conversations with peers—on both technology and business fronts—and make more informed decisions about whether, where, and how to use these tools.
Key Insight: In a fast-moving field, direct tool experience becomes a leadership capability: it improves judgment, reduces overreliance on headlines, and enables better governance decisions about adoption, appropriate use cases, and human oversight.
Leadership Implications
- Start with business problems, not tools: avoid “technology looking for a problem” by defining needs before selecting AI methods.
- Set realistic expectations and ready the foundations: ensure processes, data availability, and business–technology collaboration are addressed early.
- Keep humans in the loop: require fact-checking and validation for generative AI outputs that may be plausible but incorrect.
- Monitor IP and ownership exposure: track evolving debates and lawsuits tied to training data and output resemblance.
- Promote hands-on experimentation for workforce enablement: encourage leaders and teams to try tools (often free) to build shared, informed understanding.
Conclusion
Welsch’s message is that AI is now pervasive, and generative AI has accelerated both opportunity and risk. The differentiator for executives is not simply adopting new tools, but building a grounded AI strategy that starts from business needs, incorporates oversight for accuracy, and anticipates IP uncertainty.
For leadership teams, the fastest path to better decisions may also be the simplest: try the tools directly, learn what they can and cannot do, and guide responsible adoption with informed governance and workforce readiness.
FAQ
What should an executive AI strategy prioritize right now?
An executive AI strategy should prioritize real business needs, realistic expectations, and selecting the right AI approach for the problem. Generative AI adds new options for content creation, but it still requires human oversight for accuracy and careful attention to IP risk.
How is generative AI different from traditional AI in business?
Generative AI differs by creating new content rather than only automating tasks or generating insights from historical data. In the interview, Andreas Welsch describes prompt-based systems that can generate images or long-form text, expanding AI’s role in productivity and content creation.
Why did some AI investments fail to deliver in earlier waves?
Some AI investments underdelivered because processes were not in place, expectations were too high, data was missing, and collaboration between technical experts and business stakeholders needed more nurturing. Those organizational gaps can limit business impact even when AI technology is strong.
What business use cases are most practical for generative AI today?
The most practical generative AI use cases today are productivity-driven, such as drafting long-form blog content, generating images from text prompts, and accelerating video production with AI avatars. These outputs can reduce effort and cost, but should be reviewed before publishing.
What are the biggest governance risks with generative AI?
The biggest governance risks are unclear intellectual property ownership and the possibility of convincing inaccuracies in outputs. Andreas Welsch notes lawsuits involving stock image providers and highlights that large language models can generate plausible but wrong answers, requiring human validation.
Why does “human-in-the-loop” matter for AI adoption?
Human-in-the-loop matters because generative AI outputs can sound credible while being incorrect. Welsch’s example shows a model confidently describing “cow eggs,” which do not exist. For responsible AI adoption, teams should fact-check and validate outputs before business use.
Is the metaverse mainly a game, or a serious business trend?
The metaverse began largely from gaming experiences, but it has longer-term potential to become significant over time. Welsch suggests it will take time to become pervasive, though some brands and retailers are already experimenting following broader industry moves like Meta’s rebranding.
How could generative AI and the metaverse connect in the future?
Generative AI could help create highly personalized metaverse experiences by generating imagery, sound, music, text, and narratives on demand. Welsch links this to a shift from fragmented, one-to-many experiences toward more one-to-one personalization if brands have richer user context.
How can leaders keep up with fast-moving AI developments?
Leaders can keep up by trying tools directly, especially when they are available for free. Welsch recommends signing up and building firsthand experience to participate in informed discussions with peers and to decide whether, where, and how AI should be used responsibly.