You Have AI Options—Use Them Wisely

Artificial intelligence isn’t a monolith. It’s more like a toolbox—and the tools you pick or ignore can determine whether your content strategy and operations thrive or falter. In our firm’s long-running research into content operations, we’ve found that the biggest gains don’t come simply from “doing AI.” They come from choosing the right AI, in the right place, within the right operational framework.

We see too many organizations take a “copycat” approach where they jump into generative AI because a competitor or an organization they admire is doing it. If you’re a leader or executive in product, marketing, customer experience, or communications, it’s time to move your organization beyond AI imitation into AI strategy, where you make thoughtful choices. 

To jump start your thinking, I’m sharing five sets of AI options to consider.

AI Option 1: Content Creation or Other Use Cases?

Because it’s easy for all of us as individuals to use gen AI like ChatGPT and Claude to create content, it’s tempting to think that’s what an enterprise should do. But our latest study of content operations and AI finds that enterprises operating at the highest levels of content ops maturity are much less likely to focus first on the content creation use case. And they’re having much more success with scaling AI.

For instance, in our recent panel discussion about content operations and AI, Senior Director Laura Barnes of Red Hat shared that her team tackled the use case of enhancing metadata first. AI freed her team from the tedium of updating and maintaining metadata so they could focus on more strategic and rewarding work.

Yes, generative AI (for drafting, rewriting, image-creation, summarization) has its place. But it’s also risky when your processes, governance, and measurement aren’t ready. Consider AI use cases where the risk is lower and value more immediate such as

• Personalization and distribution (e.g., delivering tailored messages or flows)
• Content intelligence and optimization (e.g., tagging, metadata enrichment, audit automation)
• Governance (ensuring brand voice, compliance, quality)
• Operational efficiency (workflow automation, routing, version control)

AI Option 2: Build or Buy?

Once you’re committed to using AI, you face a classic question: Do you build your own models and systems internally, or do you buy existing tools/platforms? Each path has trade-offs.

Sample build (proprietary model) trade-offs:

• Potential for competitive advantage and customization (you tailor models to your domain, voice, data)
• Higher initial cost: infrastructure, talent, integration
• Longer time to value

Sample buy (existing tools/platforms) trade-offs:

• Much faster to deploy
• Lower up-front investment
• Possibly less tailored to your content situation or brand voice

And it’s very possible to use both in a smart way. Buy commercially available platforms to get moving, then build components that differentiate (e.g., domain-specific fine-tuning, proprietary datasets, custom workflows).

AI Option 3: Deep Learning (Generative) AI vs. Machine Learning AI

This choice is often overlooked, but it’s foundational, as I discuss in The Content Advantage. Let’s take a close look at these options.

Deep Learning / Generative AI

Generative AI (e.g., LLMs, diffusion models) is powerful, creative, and flexible. But it also carries inherent unpredictability. For instance, deep learning models

• Can produce variations each time they run (stochastic outputs).
• May hallucinate facts or misinterpret context.
• Struggle with certain structured tasks like precise numerical reasoning and math.
• Are harder to audit and govern.
• Introduce additional risk in regulated or high-stakes environments.

Machine Learning (ML)

Machine learning approaches—such as classification, clustering, regression, recommendation systems, or rules-based hybrids—offer:

• Higher predictability and reproducibility.
• Lower risk, especially where accuracy must be consistent.
• Stronger data traceability and interpretability.
• Excellent performance in structured tasks (e.g., scoring, routing, detecting anomalies).

Forward-thinking organizations treat generative AI as one powerful tool among many. And I think the best systems emerging combine them. Machine learning governs decisions and structure, while deep learning enhances creativity, translation, or natural-language interpretation.

AI Option 4: Large Language Models or Small Language Models?

If you choose deep learning AI, then you’ll face another choice. Do you go big with large, general-purpose LLMs (large language models) like GPT, or do you opt for smaller, more specialized SLMs (small language models)? Again, factor in the trade-offs.

Sample LLM trade-offs:

• Powerful and flexible
• Handle wide variety of tasks out of the box
• Often expensive (compute, licensing) and harder to control in terms of data security, bias, domain specificity

Sample SLM trade-offs:

• Less flexible across diverse tasks
• Easier to fine-tune for niche or domain-specific needs
• Lower resource cost
• Easier to align with your brand voice, data privacy, workflow constraints

Organizations that understand when they need broad flexibility and when they need focused or specialized accuracy will make solid choices here.

AI Option 5: Productivity or Transformation ROI?

Finally, consider your approach to ROI from AI carefully because many organizations are struggling to see returns. Is your goal short-term productivity (faster drafting, editing, publishing), long-term transformation (new offerings, deeply personalized experience, business model change), or a combination?

If you focus on productivity, the benefits might show up quickly, such as less time spent, fewer bottlenecks, and faster time to market. But those benefits might not add up to a return that exceeds your investment, and you’ll quickly hit a ceiling.

If you aim higher, using AI to fuel business transformation such as reimagining customer experiences, enabling personalization at scale, or creating entirely new offerings and capabilities, then the ROI can be exponential. 

Productivity is like tuning your engine. Transformation is like building a whole new vehicle equipped to go where you haven’t been before. You might want the engine tune-up now, but if you don’t eventually plan for a new vehicle, you’ll be stuck in yesterday’s race.

And regardless of the option you choose, assessing ROI will be nearly impossible if you don’t have a system to measure content effectiveness and impact before and after deploying AI.

Closing Thoughts

Your enterprise relies on content, the substance of digital products, services, communications, and customer experiences.  So when you make choices about AI, you’re not just selecting a tool—you’re making a strategic statement about how you view content. Is your choice ad hoc, tactical, or transformative?

If your organization tries to “just use AI” without regard for why or how, you risk deepening chaos or missing out on ROI. But if you deliberately choose your path, AI stops being a risky experiment and becomes a force multiplier.