Much like Hamlet’s dilemma, society often frames artificial intelligence as a binary choice, but that’s a false dichotomy. AI is already embedded in our daily lives, from filtering spam to powering voice assistants (cue frustrated groan when they misinterpret us). It’s been quietly supporting daily writing tasks for decades: spell checkers, grammar tools, and vocabulary suggestions. Today, AI can generate entire documents based on user prompts and hopefully without “hallucinations.”
Despite AI’s current ubiquity, people still resist its further expansion. Common concerns include:
- Potential job loss: due to layoffs or shifting job requirements
- Decision-making concerns: trusting AI and assigning accountability
- Privacy fears: surveillance and data misuse
- Emotional resistance: rigidity and impersonal nature
- Risk of misuse: unintended consequences and ethical concerns
Reframe the Question
The real question isn’t whether to use AI, it’s how. Too often, leaders rush to implement AI under pressure to cut costs or appear innovative. In doing so, they choose AI projects that are not well thought-out and unlikely to succeed. The first question should be: Where can AI meaningfully support our mission?
Answering that requires three layers of understanding:
Know Your Organization
How does your organization make money and what are the operating costs? Who are the organizational stakeholders and what areas have they identified as inefficient? Considering the organization’s industry and competitors, where are the opportunities for innovation to achieve a competitive advantage? Tools like the Business Model Canvas from Strategyzer can help organize these elements efficiently. Crucially, gather input from those doing the work because documented processes rarely reflect reality.
Assess Your Technology Landscape
Do your systems support AI integration? Legacy systems may limit feasibility. Identify what needs to be updated for the technology landscape to meet the prerequisites. In “From Vision to Velocity”, I outline how to document a roadmap for readiness. Use that approach to identify gaps and prepare your tech stack.
Evaluate Your Data
Do you know where your data lives, and whether it’s complete, accurate, and unbiased? AI amplifies patterns without distinguishing between helpful and harmful ones. Missing or inaccurate data increases the risk of AI hallucinations, as AI fills in blanks or rationalizes inconsistencies. If your hiring data reflects unconscious bias, AI will enforce it with ruthless consistency. Humans can see nuance; AI cannot. If you made a decision one way 9 out of 10 times, AI will make it 10 out of 10.
Define the Path
Use the information collected to identify potential AI projects. I recommend using the 5 Ds Framework to help identify where AI is most likely to deliver value. This framework is discussed by David A. Wood in his book “Rewiring Your Mind for AI”.
“If a task is dirty, dull, dangerous, distant, or data-driven, it’s a candidate for AI—not because AI is magic, but because it’s built for scale, repetition, and pattern recognition.”
Here are some examples for each of the Ds:
- Dirty: Data cleansing, spam filtering
- Dull: Routine scheduling, invoice processing
- Dangerous: Fraud detection, cybersecurity monitoring
- Distant: Remote diagnostics, global customer support
- Data-driven: Forecasting, personalization, sentiment analysis
The next step is to create a simple prioritization matrix and rate as follows:
- Business Benefit (1 = high, 5 = low)
- Time to meet technical prerequisites (0 = no time, 5 = a lot of time)
- Time to meet data prerequisites (0 = no time, 5 = a lot of time)
- Estimated time to implement (1 = minimal, 5 = a lot of time)
- Overall complexity (1 = minimal, 5 = very complex)
- Sum the ratings to get an overall score, lower numbers indicate higher priority.
This helps prioritize responsibly. Don’t manipulate ratings to get the priorities you or your boss wants. Starting with a lower-stakes, simpler project builds familiarity and resilience. Mistakes are part of the learning curve; make them on a project that is not mission-critical. Adjust the roadmap mentioned above to reflect the priorities.
Avoid the Shortcut Trap
When a vendor pitches AI solutions, proceed with caution. Great products exist, but there are no magic solutions. If vendors skip the foundational work, they lack the insight to guide your organization effectively. Even when vendors do the research and analysis, it may be too high-level. Consider separating the research and analysis work from the implementation work across vendors to avoid conflicts of interest.
The research and analysis work doesn’t have to be slow. With the right strategies, you can gather accurate information quickly. Those strategies rely largely on a collaborative approach with your team members most familiar with the work.
“To AI or not to AI” is the wrong question. The right ones are: Where will AI serve us best? What do we need to prepare? How do we ensure we are working on the right things at the right time? Thoughtful answers lead to meaningful, sustainable AI implementation.
At Braided Strategy, we help leaders ask and answer those questions. Our advisory approach starts with understanding your organization’s unique context, then guides you through the strategic, technical, and ethical considerations that make AI work for you. If you’re ready to move from reactive adoption to intentional integration, let’s schedule a time to talk here.
