How to become AI-ready: defining “Winning”

Practical methods from real-world examples

In my last article, I explained how important it is to define what winning looks like.

Because before you dive into the technology, you first have to look in the right direction: toward the customer / user / citizen. You can read up again on why this is so important here and here.

That is the North Star. Now the task is to define the game. What is winning and what is losing?

That is easy to understand and hard to implement. There are a multitude of methods, and none of them fits 100%.

The danger is always that you are either too generic or get lost in the weeds.

Hence this article. A practical guide to how you can get something off the ground here with the help of AI.

In this article you’ll get:

• practical tips
• hands-on methods, and why they matter
• real-world examples of where we applied them

Use AI to help you learn

I find the following approach helpful:

1. Know a few methods, that means read about them – or have an AI explain them to you and think about them
2. Write down notes on your own situation
3. Identify the right questions with the help of AI
4. Answer the questions
5. Give this input to the AI and ask the AI
   1. Which of this is absolutely unrealistic to implement in my situation?
   2. Which of this is definitely realistic to implement in my situation?
6. Then listen to your gut feeling:
   1. which of the AI’s answers feel right to me,
   2. which ones bother me
   3. or do I have no feeling about it at all? That happens when there is still too little experience to draw on. In that case, take the steps that are easiest to implement
7. Put first small steps into practice, specifically for the points that feel right or the ones that are easiest.

In the following, I present methods that I have applied myself together with teams and leaders in large companies across various industries, but never exactly as in the template – always modified and adapted to the situation!

There are many, many more methods, and I can’t cover them all. What matters more is that you start thinking about it and that you start putting something into practice that feels right. Which method you choose in the process is secondary.

Real-world examples from my career

Example: Insurance company, 7 teams, approx. 80 people

• Winning: Each team and all teams together develop solutions (new software + new workflows + new automations + new standards in other departments) that handle customer requests to the customer’s complete satisfaction and it works fully automatically – without manual work.
• Losing: Customer requests are not resolved, customer requests are resolved manually, customers are dissatisfied, teams develop solutions that nobody needs and that contribute nothing to winning.
• Connection to Data and AI: The automation rate was calculated from data and measured every sprint. One team built this data calculation and measurement.

All of this was still before AI. If the teams had to do it today, it would be child’s play for them, because this organization has now practiced for years the way of working of aligning with the customer, delivering every sprint, and being measured against KPIs. Perfect starting conditions.

Example: Car manufacturer, 10 teams, approx. 150 people

• Winning: All teams together successfully develop a new head unit (this is the thing in the car where your radio, your phone app and your navigation lives) with new functionalities, new apps, a new chip, and new hardware in order to be able to offer customers additional services for the vehicle.
• Losing: The new head unit delivers no added value compared to the old one.
• Connection to Data and AI: None, really, because the example is already a few years old.

But I included it anyway because winning and losing were lived out in exemplary fashion here and everyone was measured against numbers, data, and facts, which I thought was cool.

Every sprint, the new head unit was tested while driving the car and using the head unit, and all delivered features, all bugs, all costs, etc. were captured in data and measured, and decisions were made solely on the basis of numbers, data, and facts.

Example: Pharmaceutical company, 5 teams, approx. 50 people

• Winning: In this initiative, winning meant: We understand which doctor treats the diseases for which we have medications, we understand the diseases and how we can help, and we understand which doctors the sales reps should visit next and in what order (recommendation model).

(In Germany, there are strict data protection laws in health care, that means, you are not allowed to know who has which disease and who goes to which doctor.)

• Losing: Building dashboards that nobody needs and that don’t answer these questions.
• Connection to Data and AI: It was a huge data platform team, which we divided into smaller teams. In addition, a machine learning team was founded. The recommendation model was a complete success.

Example: Fashion shop (online and stores), 2 teams, approx. 20 people

• Winning: The business knows which items of clothing from the warehouse should be offered next to which target group online and in the stores.

(This company had a huge warehouse because its business model was based on selling past-season items from well-known brands.)

• Losing: A great data warehouse and zero actionable insights for the business.
• Connection to Data and AI: A data platform team that provided data for the business departments. AI was still in its infancy here.

Example: Energy supplier, 3 teams, approx. 30 people

• Winning: Everything new! New data platform (in parallel, new CRM, new ERP), new cloud infrastructure, new machine learning – and all of it AT THE SAME TIME!
• Losing: Millions were burned. The customer was miles away. The data platform program was scrapped. Machine learning went quite well, though. Cloud made progress. I didn’t follow CRM and ERP any further, but the new data platform never went into production, and they continue to work with the old one.
• Connection to Data and AI: 3 teams that built the cloud infrastructure, data platform, and a machine learning initiative.

A great textbook example of how NOT TO DO IT in the data space. I’ll come back to it in later articles.

Not a single method was used. The only way of working was (micro-)managing people and painting colorful slides.

Some useful methods

Now for the methods. Please research the materials on them yourself. For one thing, the article would otherwise get too long, and for another, the point of the exercise is for you to think about them.

Or ask an AI of your choice about each method. The methods are ancient and world-famous, that means any AI is trained on it.

Value Proposition Canvas

There are nice templates for this. What matters is not that you fill out exactly these perfectly.

• You should understand how to put the customer at the center,
• and you should understand whether what you are doing solves the customer’s pain points
• or whether what you are doing brings the customer a gain.

You should think this through and map it systematically. That’s the whole exercise.

If what you are planning is neither a pain reliever nor a gain creator, then what is its reason for existing?

That is what you should think about.

Lean Startup Method

In the Lean Startup method, you formulate hypotheses, think about how you can test them, and run experiments. Based on what you measured in the experiments, you decide how to proceed.

Here, too, the point is not to impose some template from the internet exactly as is. Think first.

• What assumptions have you made?
• How can you validate or refute these assumptions?
• Which experiments help here?
• What can be measured?
• What kill criteria do you want to set?

You can use this way of thinking for all initiatives. Instead of assuming that you know everything, assume that you are merely making assumptions.

These can be assumptions about what winning really is for the customer, but they can also be assumptions about whether the path to get there is the right one.

This way of thinking forces you to put winning and losing into words and to translate them into data and measurement points.

This way of thinking is the discipline of actually following through on it.

Practice this mindset and look for an example where it fits in your context.

Impact Mapping

There are nice templates here, too. And here again: Don’t turn it into dogma.

But: Think about how you want to implement “winning.”

Who has to do what and how, so that what comes out in the end?

Again, it’s about building the bridge from strategy to execution in a disciplined way. What good is a great vision if you don’t know how to implement it?

That’s what the method is for.

It forces you to actually write down how the idea can become reality.

This is not about details. But it is about getting grounded.

If nothing comes to mind, if you don’t know who would have to do what and how so that winning becomes real, then you realize: Your wording for winning is too generic. Back to the value proposition and do your homework again.

Only when you are able to formulate winning in such a way that you can then say who has to do what and how, and which deliverable ultimately contributes to winning, does strategy actually become execution.

Balanced Scorecard

Here I’m thinking of a simplified version of it.

It is SUPER IMPORTANT not to focus only on financial goals. Focusing purely on ROI or money-related goals won’t get you where you want to go.

The workflow that brings you profits in the short term – is it stable, repeatable, independent of individual people, and can it be sustained this way over a long period without everyone burning out? That is the Process dimension.

A workflow that is repeatable and profitable but doesn’t force you to learn something new every time – that is the moment where it becomes a “legacy system”. The world keeps turning and you stand still because you failed to factor in from the start that the organization has to keep learning permanently.

And finally, the customer is put back in focus.

The Balanced Scorecard

• takes the customer into account,
• takes the way of working into account,
• and takes learning into account.
• And, of course, profit as well.

Assuming you know what winning means in concrete terms and you have a plan for implementing the strategy, then make sure that all leaders have to improve KPIs from all areas of a Balanced Scorecard. That’s how you prevent dysfunctional incentives.

The customer as the North Star prevents a one-sided focus on short-term financial benefits. Improving the way of working brings everyone on board. Improving the organization’s learning forces a leader to cooperate with other leaders.

Such a simple tool, and so incredibly helpful.

KPIs

KPIs follow on from this. Define KPIs that cover all areas of a Balanced Scorecard. Both teams and leaders are measured against the KPIs.

• Better fewer than more.
• Better pragmatic than scientifically correct.
• The measurement has to be quick and uncomplicated to do.
• Better to measure a 5-star rating based on your stakeholders’ gut feeling than to measure nothing at all.
• Better to start right away.

It’s about improvement, and it’s about establishing facts. You can optimize it bit by bit but don’t turn overcomplicate it - in my German dialect we say: don’t turn it into a doctoral thesis!

Circle of Leaders

There is no official method called like this. I saw it in some organizations and adjusted it.

At some point I realized that I achieve much more when I work with groups.

From the very beginning, I bring all the leaders who are responsible for an initiative into one group and call it the xyz Circle. Xyz can be anything. Pick a fitting name.

• Everyone has to be in (everyone who is a leader and is responsible for the initiative)
• you meet weekly or sometimes even daily,
• and anyone who isn’t there supports the group’s decisions.

Everything that needs to be clarified gets clarified. Everything that needs to be measured gets measured. Everything that needs to be decided gets decided.

Keep the agenda simple. But follow through consistently.

It always works.

• Responsibility is spread across several shoulders,
• people understand each other better,
• you move forward much faster, everyone feels relief and progress.

The method is so good that every time I ask myself why it hasn’t long since become a standard in management literature.

• You don’t have to change an org chart,
• you don’t have to run a change project,
• you just have to give the whole thing a sexy name,
• make the meeting cool, effective, and efficient,
• and work with the dynamics of the group.

You have to be a suitable host for a meeting like this. Focused, friendly, not a chatterbox, not putting yourself at the center, but making everyone’s life easier. Then it works.

Reviews are Pitches to the Board

This method is bulletproof. It’s awesome. And it’s simple.

Even in a huge global corporation, this is doable. We did this.

Once you know which teams are involved in the initiative, what winning looks like, how you want to measure it, and once you have formed your leadership circle, then take the sponsor and find the highest level of the company that has an interest in its success. Ideally the board.

All teams and the leadership circle pitch regularly to the board, like a startup.

The agenda MUST HAVE at least this:

• a live demo from something running in production (uh, yes, we did this in large enterprises, the stuff was live in production)
• an explanation how this demo connects to the value proposition for customer/user/citizen
• data that proves how you validated the hypothesis with experiments
• data that proves what you learned and what you change because you learned
• and a short overview how you will operate and govern this

The board decides whether or not there will be further funding. That means this is not an annual funding process. That means all vendors can not rely on an annual funding. This changes the whole vendor structure. We will discuss this in later articles.

Of course, the whole thing only works if you actually follow through with it.

Stop: This is NOT only leadership talking to leadership.

It is an EVENT. All team members are present. The teams show their stuff themselves. It’s a pitch! Don’t waste the board’s time with boring technical details. And make sure everyone who isn't attending feels serious FOMO.

Prerequisite: It must be possible to kill initiatives because the assumptions did not prove true, without anyone losing face or damaging their career in the process.

Sure, not every organization will manage that right away. There is a lot to learn! That will then be the topic of further articles.

That’s it for today.

Your Exercise

1. Take this article, use an AI of your choice, and additionally give the AI the following information about you and your initiative:
   1. Industry
   2. Company size
   3. Your role
   4. What the initiative is about
   5. Which teams are working on it
   6. Which stakeholders you have
   7. Which problems you have already identified
2. Tell the AI to give you a list of questions that are still open and that you need to answer before you decide which methods you can apply
3. Answer the questions
4. Give this input to the AI and ask the AI:
   1. Which of this is absolutely unrealistic to implement in my situation?
   2. Which of this is definitely realistic to implement in my situation?
5. Then listen to your gut feeling:
   1. which of the AI’s answers feel right to me,
   2. which ones bother me
   3. or do I have no feeling about it at all? That is because the experience is lacking.
6. Implement first small steps, specifically for the points that feel right or seem least difficult.

Try it out and get in touch with me. Did it help?

Best,

Bianca

P.S.: Many, many other articles and methods will follow. We just started.

I know that everyone talks about redesigning workflows for AI. But you must get the fundamentals right first.

If you mess up at the start, things will head in the wrong direction and political dynamics will throw a wrench in your plans. The whole thing is a complete setup, it takes a lot of building blocks. Join me on this learning journey.