Skip to content
MXFACTORY

Too Many AX Ideas? A Scoring Method for Deciding What to Do First

Rank AX project candidates by frequency, time spent, cost of errors, data readiness, and owner commitment, and learn which projects to avoid early on.

BY
MXFACTORY
PUBLISHED
Reading time
3 min read

Why Order Matters When You Have Many Candidates

Once a company starts exploring AX (automation and AI transformation), ideas arrive from every direction: sales, finance, production, customer support. The shortage is rarely ideas. It is people, budget, and the time of the staff who must carry the work. Start everything at once and nothing finishes.

What you need is a shared yardstick, not gut feeling or the loudest department's request. Scoring each candidate on five criteria keeps the discussion calm and comparable.

Scoring on Five Criteria

Rate each candidate low, medium, or high (a simple 1 to 3 scale is enough) on every criterion. This is not precise math. It is a tool that lets your team look at the list the same way.

  • Frequency: How often does the task recur? Daily or weekly work compounds its benefit. A task done once a year can wait.
  • Time spent: How long does one run of the task tie up a person? Combined with frequency, it gives a rough sense of total burden.
  • Cost of errors: How much does a mistake hurt? This one runs in the opposite direction: the lower the cost, the better the first project. Start with work where a person can easily catch and fix a wrong result.
  • Data readiness: Is the needed information digital, consistently formatted, and accessible? If it lives in people's heads, on paper, or scattered across files, half the project is data cleanup.
  • Owner commitment: Does the person who actually does this work want it automated? This is the most underrated criterion. A tool the front line does not use will sit idle, however well built.

The best first candidates are frequent and time-consuming, cheap to get wrong, backed by ready data, and championed by the people doing the work. If one or two criteria score very low, flag them even when the total looks high.

Practical Tips for Using the Scorecard

  • Have several people score independently. Let front-line staff, team leads, and executives score separately, then discuss the items where they disagree. Hidden circumstances tend to surface there.
  • Look at blockers, not just totals. If data readiness or owner commitment is the lowest rating, park the project as "revisit after prerequisites are solved," whatever the total.
  • Pick a small, visible first project. An early success earns the organization's trust and cooperation for the next one.
  • Revisit the scores every quarter. Data gets cleaned up, owners change, and the ranking shifts with them.

Projects to Avoid

Regardless of score, some project types are best avoided early on.

  • High-stakes work with no way to verify results: contract terms, payments, or legally sensitive judgments handed over without human review.
  • Work whose process is not yet defined: if everyone does it differently and no written standard exists, clarify the process before automating it. Automating a messy process only makes the mess faster.
  • Projects the owners oppose or ignore: pushing from the top alone leads the front line to find workarounds.
  • Projects with no way to measure the effect: if you cannot describe the starting point, you cannot describe the improvement.
  • Sweeping projects that try to change every department at once: a wide scope stretches the schedule and drains momentum. Split it and begin with the smallest piece.

Summary and Next Steps

Prioritizing is not about picking the most impressive project. It is about picking the one that will actually be finished and used. Here is a simple way to begin:

  1. List every candidate in one table.
  2. Have several people score each one on the five criteria.
  3. Among the top two or three, choose the one with no major blocker as your first project.
  4. Move candidates matching the "avoid" types to a hold list, with the reason written down.

If building the candidate list or scorecard feels daunting, MXFACTORY can help you work through it.

NEXT STEP

Let's start with a diagnosis.

One team handles AI development, branding websites and marketing. Tell us where you are, and we will decide together what to do first.

Send an inquiry

This post is general guidance. Every company is different, so review it against your own situation before applying it.