Depends. If you know exactly what you want to automate, start building. If you're asking "where should we use AI?" then yes, strategy first saves you six figures.
Skip strategy if
You have a specific, well-defined workflow to automate. The problem is clear: "We need to process these documents faster." You've seen the solution work elsewhere in your industry. Budget is under $50K and timeline is under 8 weeks.
In these cases, a strategy phase just delays the build. Start building, learn from shipping, then strategize about scaling.
Do strategy first if
You're asking "where should we use AI?" not "how do we build X?" Multiple teams are requesting AI capabilities. You're planning $200K+ in AI investment this year. You've had failed AI projects before.
A strategy sprint prevents the most expensive mistake: building the wrong thing well.
What a good strategy sprint delivers (2 weeks)
A prioritized list of 3–5 high-impact opportunities. A data readiness assessment for the top pick. Architecture recommendations. A 90-day execution plan with clear ownership. What to skip and why.
If the output is a 60-page deck with no action items, you hired consultants, not strategists.
Related reading:
Frequently asked questions
What does a strategy sprint cost?
Typically $10K–$30K for a 2-week diagnostic and prioritized roadmap. Anything more is over-scoped for a first pass.
Can we do strategy and build in parallel?
Yes, if you already have a clear first use case. Run the strategy sprint to prioritize what comes next while building the obvious one.