The AI Balanced Scorecard Coach skill for Claude

This copy holds 1,324 words of instructions that tell Claude how to translate a strategy into a Balanced Scorecard: which questions get asked first, in which order the four perspectives are worked out, what a KPI should look like and what may never be invented under any circumstance. Download the zip at the top of this page, put it in your Claude environment, and from that moment on Claude builds scorecards following this logic.

The problem the skill solves is named literally in its own activation list, as the sighs it responds to: we only steer on revenue. I have no idea whether we are on track. Everyone is running in a different direction. Our strategy is a nice story but nobody does anything with it. We measure everything but nothing tells us anything. Anyone who says something like that to Claude while the skill is loaded gets it automatically. Those are not marketing lines, those are the triggers from the frontmatter of the file.

It is written for anyone who wants to make a strategy measurable: SMEs, corporates, healthcare, government, non-profits and education are named explicitly in the file. It is part of the skill library we make available for free through the AI and automation service, with no account and no sales email afterwards.

One thing you should know beforehand, because the skill says so itself: the Balanced Scorecard translates an existing strategy into measurement and action, it does not invent a direction. If you do not yet have a strategic ambition, the skill first helps you formulate that one sentence before it writes down a single KPI. And it is explicitly not a replacement for your financial administration or annual accounts.

The word balanced refers to three kinds of balance at once: between financial and non financial, between short and long term, and between the internal and the external view. A management team that discusses only last month's revenue every month has none of the three. It only sees that something is wrong once it is already in the figures, and by then the cause, somewhere in the customer experience or a stalling process, is already months old.

What that means in practice is clearest in how the skill rewrites a vague intention into a KPI. Every KPI in the scorecard is SMART: specific, measurable, with an owner, a measurement frequency and a target value. If that target value is missing, there is no estimate but [FILL IN], with a suggestion for how you can get the figure. In diff form, with lines towards the example from the file:

The difference is always the same: nobody can check the left line, everybody can check the right line. It states who looks at it, how often, and where the data comes from. That is the whole scorecard in miniature.

Then follows the theory in short: the origin with Kaplan and Norton, the four perspectives and the thinking behind the Strategy Map. Then comes the heart of the file, a method of nine fixed steps. First the organisation is explored: sector, size, ownership structure, growth phase, mission and time horizon, twelve months by default.

Then it asks for the strategic ambition in one sentence, and if the user cannot give that, that sentence is formulated first before a single KPI reaches paper. After that it works out the four perspectives one by one, formulates two to three SMART KPIs per perspective, builds the Strategy Map and closes with an implementation note: who discusses the scorecard monthly, where the data comes from and what the first measurement point is.

What stands out is how strict the file is about what may not go in. For the internal processes the skill must explicitly name which processes are NOT critical, so the scorecard does not become an inventory list of the whole organisation. Target values the user has not supplied become [FILL IN], never an estimate. And the implementation note must end with the three biggest risks for this specific scorecard, so the skill names its own weak spots before the management team does.

Before the writing starts, the skill runs through an input checklist of six points: sector and size, the ambition or the honest answer that it does not exist yet, the time horizon, the figures that already exist, the existing measurement systems such as bookkeeping, CRM, survey tool or HR system, and who is going to use the scorecard. That last point determines the language and the level of detail: a scorecard for the board sounds different from a scorecard for the whole team. Only what is missing gets asked about, so you get no form but at most a few targeted questions.

The file closes with a list of seven things the skill never does, worked out below in Section 04, and a source list of five titles: four by Kaplan and Norton and one by Doran, on the origin of the SMART format. More on that below.

The framework comes from Robert S. Kaplan and David P. Norton. Kaplan, professor at Harvard Business School, and Norton introduced the scorecard in 1992 with the article The Balanced Scorecard: Measures That Drive Performance in the Harvard Business Review, based on research at twelve companies. In 1996 the book The Balanced Scorecard: Translating Strategy into Action followed, in which the framework grew from a measurement tool into a management system.

The Strategy Map got its own book in 2004. In Strategy Maps: Converting Intangible Assets into Tangible Outcomes, Kaplan and Norton worked out further the causal chain you saw in FIG.02: learning and growth feeds processes, processes feed the customer promise, the customer promise feeds the financial result. In 2008 The Execution Premium followed, about linking strategy to operations, and that is exactly where the monthly discussion rhythm comes from that the skill enforces in every implementation note.

The SMART format comes from George T. Doran. His article There's a S.M.A.R.T. Way to Write Management's Goals and Objectives from Management Review, 1981, is the fifth source in the file. Whatever comes out of it follows that format. Anyone who wants to learn to formulate goals SMART outside the scorecard can turn to the SMART Goals Coach skill: the two have deliberately been kept separate.

Even without installing the skill you can improve your own steering with these principles: do not measure only the final stop but the whole chain, give every measurement an owner and a frequency, and treat every causal link as a hypothesis you test instead of a fact you assume.

After that, using it is simple: give your organisation context and your ambition, and answer the few questions the skill asks back. The more honest your input, the more usable the scorecard, and that goes especially for the answer that a figure does not exist yet. If you want to learn to write a file like this yourself, read writing a SKILL.md. And if you are new to this kind of file, the knowledge base has the broader explanation of working with AI.

There are also situations where you are better off leaving it alone, and the skill names them itself. If you do not yet have a strategy, the scorecard is too early: the model translates a direction, it does not invent one. If your business moves so fast that the plans tip over every month, the model is too slow, and that is stated literally in the list of limits the skill never conceals. Work that cannot be quantified is not captured by this scorecard either. And a scorecard with no monthly discussion rhythm is a document, not a steering instrument, so reserve that meeting before you start, not afterwards.

One more honest limit: the scorecard makes your steering better, not your strategy. If the direction was chosen wrongly, the scorecard only measures more precisely how you are driving the wrong way. What it does do is make that conclusion visible months earlier than a revenue figure ever would, and that is exactly why the causal chain in FIG.02 is treated as a hypothesis you test.

A scorecard makes visible whether you are on track, but in the customer perspective everything starts with inflow: are you actually found by the people looking for you? The free SEO scan shows in seconds where your site stands, and that is immediately a first baseline measurement for your own customer KPI. And if you want to talk further about what AI can do for your business beyond this, from separate skills to full automation, we simply do that in a conversation.