How do you decide which business processes are best suited for AI optimization?

When I think about this question, it's not a single decision. It's a matrix. You're weighing multiple factors at the same time, and where a process lands across those factors tells you whether to prioritize it, defer it, or skip it entirely.

Start With the Foundation

If you're just getting started with AI, choose the easiest items first. Any introduction of technology has a foundation that needs to be in place before it works. For AI, that means software setup, plugins, API keys, account creation with an LLM provider. If you're planning to integrate that software into existing systems, you need to understand the connection points: APIs, integration platforms, data formats. It also means establishing basics around security, file organization, backups, and access controls.

Get this foundation right once. Everything you optimize afterward builds on it.

Can AI Actually Reach the Process?

How much of the optimization lives in a system that AI has access to? If the process runs inside software with APIs or data exports, AI can work with it. If it requires capabilities you don't have yet (visual processing, audio transcription, live data feeds), you need to factor in the cost of adding those capabilities before the optimization is worthwhile.

Do You Have the Data?

AI optimization needs data to work with. If you have historical data, you can evaluate quickly whether optimization will help. If you don't, set the expectation that data collection happens first. AI can help you collect and structure that data over time. Then you re-evaluate whether the optimization is worth continuing.

Personalization and Context

AI tools do a lot, but they need guidance on what success looks like for your specific situation. They need to understand how you want something done, not just what. Maybe the output is factually correct but the delivery format doesn't work for your team or your customers. This layer is where you teach the system your preferences, your constraints, and your standards.

Cost vs. Impact

The easiest win isn't always the highest value win. Some processes cost you significant time or money but are straightforward to automate. Others are easy to set up but save you ten minutes a week. Map both axes, effort to implement and business impact if it works. Prioritize the quadrant where effort is low and impact is high.

The Human Review Question

Some processes need you to review the output before it goes anywhere. Client communication, proposals, anything a customer sees. Others can run without you, internal reporting, data cleanup, scheduling reminders. That distinction changes your actual ROI. A process that still requires you to approve every output saves less time than one that handles the full cycle. Factor in where human review is necessary and where it isn't.

Maintenance and Drift

Once AI is optimizing a process, how often does that process change? Monthly invoicing is stable. It stays optimized for a long time with minimal upkeep. Responding to shifting market conditions or seasonal changes requires constant re-tuning. Invest your setup effort in processes that will stay consistent long enough to justify the work.

The Matrix

Plot your processes across these factors: foundation readiness, AI accessibility, data availability, personalization needs, cost vs. impact, human review requirements, and stability over time. The processes that score well across the board are where you start. The ones that need work in multiple areas are where you plan for later.