The Unreasonable Family Business

Artificial intelligence presents family businesses with an unusual temptation.

For the first time, a relatively small firm can possess capabilities once available only to large organisations. It can research markets, translate correspondence, analyse accounts, produce advertising, answer enquiries, generate photographs and videos, compare competitors and preserve the knowledge of senior family members.

A small company can suddenly acquire something resembling an additional research department, advertising agency and administrative team.

That is potentially transformative.

It is also dangerous.

The danger does not lie simply in AI getting things wrong. Human beings have always managed that perfectly well on their own.

The deeper danger comes when we start asking machines to make decisions that properly belong to us.

There is an important difference between using AI to help run the business and allowing AI to tell us what the business should become.

For family firms, that distinction may prove particularly important.

Use AI to shorten the journey. Do not automatically let it choose the destination.

When Popularity Becomes Truth

Modern computer systems are remarkably good at recognising patterns.

If thousands of successful advertisements use a particular headline structure, AI can recognise it. If certain videos receive more clicks, an algorithm can recommend similar videos. If particular products sell faster, software can identify them.

All of this can be useful.

The trouble begins when measurement quietly becomes judgment.

Social-media algorithms do not normally ask whether something is wise, original or especially useful. They can measure whether people clicked, watched, shared or bought it.

Those are different questions.

A piece of business advice that can be compressed into thirty seconds may therefore spread further than a more complicated idea requiring twenty minutes.

“Work less and make £1 million” makes a better thumbnail than:

“Build a modestly profitable company that supports two generations, allows everybody to have Wednesday afternoons off, employs six good people locally and does not destroy the relationship between two brothers.”

Yet the second may describe a remarkably successful family business.

It is simply harder to systematise.

And harder to turn into a viral video.

Gradually, the most transmissible idea begins to look like the best idea. The algorithm rewards the message that travels easily, and creators naturally produce more of what the algorithm rewards. Business advice consequently converges around the same subjects: rapid growth, scale, funnels, productivity, personal wealth and dramatic financial outcomes.

The process does more than make business advice repetitive.

It narrows our imagination about what a business is for.

Popularity begins to masquerade as truth.

Growth begins to masquerade as progress.

And large amounts of money begin to masquerade as success.

The Algorithm Likes Simple Success

Algorithms have another limitation.

They prefer things that can be counted.

• Revenue is easy to count.

• Followers are easy to count.

• Profit is easy to count.

• Employees are easy to count.

The value of having dinner with your children every evening is harder.

So is having a business that employs your sister without ruining your relationship with her.

Or being able to live in the village where you grew up.

Or taking six weeks away each summer.

Or keeping a craft alive.

Or protecting the employment of somebody who has given thirty years to the company and has now become ill.

Or simply running a company that earns enough.

These things resist convenient metrics.

They are therefore less likely to dominate business media.

We repeatedly encounter stories about the entrepreneur who built a £100 million company.

We hear much less about the woman who deliberately kept hers at £2 million because she enjoyed knowing every employee and wanted enough time to paint.

The first story contains an obvious score.

The second requires explanation.

The algorithm prefers the score.

The algorithm wants one scoreboard. A family business may have several.

And this leads to a much more difficult question.

Success for Whom?

It is convenient to speak about what the family wants.

Families rarely want anything quite so neatly.

Different people within the same business may have profoundly different ideas about what constitutes success.

Suppose two brothers inherit a profitable engineering company.

One wants to maximise profitability.

He wants modern machinery, fewer employees, professional management and expansion into overseas markets.

His brother wants to preserve what their grandfather created.

He values the firm's reputation in the town, the long-serving workforce and the workshop in which three generations of the family have worked.

Neither position is obviously foolish.

But they contain different definitions of success.

To the first brother, retaining an employee who can no longer work at full capacity may look sentimental and financially irresponsible.

To the second, dismissing somebody who gave thirty years to the family may constitute a betrayal of everything the business supposedly stands for.

The argument may appear to concern payroll.

It actually concerns identity.

The same problem appears repeatedly in family companies.

• One member wants growth.

• Another wants security.

• One wants dividends.

• Another wants reinvestment.

• One wants the next generation brought into the company.

• Another believes jobs should go to the most capable candidate.

• One wants to preserve Grandfather's methods.

• Another believes Grandfather succeeded precisely because he was willing to change them.

• One wants to sell.

• Another cannot imagine life without the business.

These are not simply differences of opinion.

They represent competing versions of the future.

When One Person's Success Becomes Another's Failure

This helps explain why apparently sensible business decisions can create extraordinary family conflict.

Suppose profits rise after ten employees are made redundant.

Financially, the policy succeeded.

But the daughter responsible for the redundancies may now feel ashamed when she walks through the town.

Her father may regard the improved margin as evidence of good management.

Her mother may believe the family has broken an unwritten promise to people who supported them during harder years.

The accountants can calculate the saving.

Nobody can calculate quite so easily what has happened to trust.

Consider another case.

The business expands rapidly.

Revenue doubles.

The financial press would probably call this success.

But the managing director now works seventy hours a week.

His marriage begins to deteriorate.

His sister resents having inherited additional responsibilities she never wanted.

Their father, who expected retirement to bring family holidays, finds that his children are constantly unavailable.

Again, the numbers say one thing.

The system says something more complicated.

This is why family-business decisions cannot always be reduced to optimisation.

A decision affects money.

But it also affects status, loyalty, identity, relationships, time and the meaning family members attach to the enterprise.

Success in one part of the system can produce failure somewhere else.

The Problem of Loyalty

Long-serving employees provide a particularly difficult example.

Imagine a production manager who joined the company at twenty-two.

• He helped the founder through a near-bankruptcy.

• He worked weekends without being asked.

• He trained the founder's children when they eventually entered the company.

• Thirty-five years later, illness means he can perform only part of his former role.

An efficiency calculation may give a simple answer.

Replace him.

A family may arrive at another.

They may create a lighter role, accept some inefficiency or continue paying him for reasons that would look irrational on a management spreadsheet.

That decision carries a cost.

But so does the alternative.

• Employees watch what happens.

• Children watch what happens.

• Customers may watch what happens.

• Most importantly, the family watches itself.

The decision contributes to the answer to an important question:

What sort of people are we?

This does not mean that every employee must be retained forever.

A business that ignores economic reality eventually becomes incapable of helping anybody.

It means only that profitability forms one part of the decision.

There may be others.

Some costs appear in the accounts. Others appear in relationships.

Grandfather's Legacy Can Mean Opposite Things

Legacy presents another paradox.

Two relatives can sincerely claim to be protecting Grandfather's legacy while advocating opposite policies.

One says:

“Grandfather built this business around traditional craftsmanship. We must protect it.”

The other replies:

“Grandfather was an innovator. If he were here, he would have bought the new machinery.”

Both may be right.

The dead cannot settle the argument.

Legacy easily becomes a container into which present-day family members place their own hopes and fears.

The important question therefore changes from:

What would Grandfather have done?

to:

What principle in Grandfather's behaviour are we trying to preserve?

• Craftsmanship?

• Independence?

• Courage?

• Loyalty?

• Experimentation?

• Financial prudence?

• Family control?

Once the principle becomes clearer, the current generation can find a contemporary way of expressing it.

That is very different from turning tradition into imitation.

The Unreasonable Man

George Bernard Shaw captured something of this problem in Maxims for Revolutionists:

“The reasonable man adapts himself to the world: the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.”

It is easy to turn this into a celebration of stubbornness.

That misses the useful point.

Shaw's unreasonable person does not need to ignore evidence or insist that everybody else is wrong.

The more interesting interpretation begins by asking:

What matters sufficiently to me that I am willing to attempt something different?

Originality then becomes an expression of purpose.

• One family may want to build the largest company in its sector.

• Another may deliberately remain small.

• Another may preserve a craft.

• Another may create employment in one town.

• Another may want enough profit to support three households while allowing everybody to enjoy their lives.

• Another may value independence above maximum financial return.

Once motivation becomes clearer, an original strategy becomes possible.

Originality is not doing something different for the sake of being different. It is knowing what motivates you well enough to find your own way of expressing it.

This quality appears repeatedly in stories about the SAS.

What makes many of those stories interesting is not merely courage.

It is disciplined improvisation.

The objective remains clear.

The method changes.

Family businesses can learn something from this.

Hold the purpose firmly enough to experiment with the method.

Wanting Can Sometimes Make the Gap Larger

There is another psychological difficulty hiding inside goal setting.

Imagine a family repeatedly saying:

“We must become more profitable.”

On the surface this sounds sensible.

But every repetition also contains another message:

We are not profitable enough.

Likewise:

“We must become a happier family.”

“We must communicate better.”

“I must become a better leader.”

“We must finally sort succession out.”

Each desired future can draw attention to the fact that the desired future does not yet exist.

That does not mean goals are harmful.

It means that wanting has an interesting double edge.

A goal can organise action.

But excessive attention to the gap between present and desired circumstances can also increase tension.

The person begins watching themselves.

Am I leading properly yet?

The family begins monitoring every conversation.

Was that good communication?

The business owner watches every month's figures.

Are we successful yet?

The desired outcome becomes a permanent test.

And tests create anxiety.

This can produce a strange result.

The harder somebody tries to become spontaneous, confident, happy or relaxed, the more difficult the state becomes.

Family businesses can fall into something similar.

• A family determined to achieve harmony may become hypersensitive to disagreement.

• A founder determined to delegate may watch every delegated decision.

• A company determined to innovate may introduce so many innovation meetings that nobody has time to experiment.

The pursuit of the outcome begins interfering with the outcome.

A destination can guide us. Constantly measuring the distance from it can make the journey harder.

Enough Is Different From More

This is another reason why the idea of enough matters.

“More” has no natural end.

1. More revenue.

2. More profit.

3. More growth.

4. More market share.

5. More followers.

Every achievement simply resets the comparison.

1. Enough requires judgment.

2. Enough money for what?

3. Enough growth to achieve what?

4. Enough time for whom?

5. Enough security against which risks?

A family capable of answering these questions has done something more sophisticated than simply choosing a financial target.

It has begun describing a life.

Profit remains essential.

But profit becomes a means of sustaining that life rather than the definition of the life itself.

Profit keeps the business alive. It does not tell you what the business is alive for.

The Consultant Who Brings the Answer

These tensions create an interesting problem for consultants and mentors.

A family business may ask:

“What should we do?”

The temptation is to answer.

Sometimes that is exactly what the consultant has been employed to do.

But prescribing a solution too quickly can inadvertently intensify the problem.

Suppose the consultant tells the brothers that the company must grow.

The ambitious brother now has an expert on his side.

His cautious brother no longer faces only his sibling.

He faces his sibling plus an authority figure.

The disagreement has become a coalition.

The consultant may have improved the strategy while worsening the family system.

Human relationships therefore matter greatly in this work.

A useful mentor may sometimes contribute more by helping people describe what they are protecting than by immediately telling them what to change.

1. “What worries you about growth?”

2. “What would selling the company mean to you?”

3. “What would you lose if your brother's plan succeeded?”

4. “How much money would actually be enough?”

5. “What part of Grandfather's legacy matters?”

6. “What does retirement mean to you?”

7. “What are you afraid might happen if your daughter takes over?”

These questions bring motives into the room.

Once motives become discussable, strategy usually becomes easier.

The consultant's job then becomes less like programming a machine and more like helping a system hear itself.

Do Not Make the Client Want Your Outcome

There is a further danger.

Consultants naturally arrive with preferred models.

• Growth consultants favour growth.

• Marketing consultants see marketing problems.

• Accountants notice financial inefficiency.

• Family therapists notice relationships.

• AI consultants notice opportunities for AI.

None of these perspectives is useless.

The problem arises when the professional's preferred outcome quietly becomes the client's desired outcome.

A mentor who repeatedly tells an owner that they must delegate may unintentionally create another standard the owner is failing to meet.

Now the person has two problems.

They still cannot delegate.

And they feel inadequate because they cannot follow the advice.

The helping relationship works better when curiosity accompanies expertise.

Instead of:

“You need to delegate.”

Try:

“You keep asking me to help you delegate, yet whenever someone takes responsibility you become involved again. Something important seems to happen there.”

That observation opens a door.

It does not dictate what lies behind it.

When People Ask for Help and Then Ignore It

This brings us to one of the strangest problems in consulting.

1. Somebody asks for help.

2. They agree with the advice.

3. They may even pay considerable money for it.

4. Then they fail to do what was agreed.

5. Or they do something resembling it but subtly alter it until the useful part disappears.

6. They ask to delegate and then supervise every detail.

7. They ask for better financial control and then avoid looking at the accounts.

8. They ask for a succession plan and postpone every meeting.

9. They ask for help becoming less involved and then manufacture new reasons why only they can solve each problem.

From outside, this can look irrational.

Occasionally it is laziness.

Often something more interesting is happening.

The proposed change solves one problem while creating another.

The founder wants to retire.

But retirement may also mean losing status.

The mother wants her daughter to lead.

But accepting the daughter's authority may mean acknowledging that her own period of leadership is ending.

The son wants independence.

But overtaking his father may feel disloyal.

The siblings want professional management.

But appointing an outsider means surrendering the private language through which they have always controlled the company.

A person can therefore genuinely want an outcome while simultaneously protecting themselves from what that outcome means.

Both forces operate at once.

Incongruence

This is where apparent incongruity becomes important.

People often imagine themselves as having one coherent intention.

In reality, several intentions may coexist.

“I want to grow the business.”

And:

“I do not want to become like my workaholic father.”

“I want my daughter to take over.”

And:

“I do not want to become irrelevant.”

“I want the consultant to challenge me.”

And:

“I do not want to discover that my way of running the company has been wrong.”

“I want AI to remove work from my desk.”

And:

“If the machine can perform the work I have spent thirty years mastering, what does that say about me?”

The behaviour that follows may appear inconsistent only if we listen to the first sentence.

Once we hear the second, it becomes much more understandable.

The person is not necessarily sabotaging themselves.

They may be attempting to satisfy two incompatible requirements at once.

The resulting compromise often satisfies neither.

Success Itself Can Create Incongruence

Paradoxically, success can make this worse.

Suppose somebody spends twenty years saying:

“When the company reaches £10 million turnover, I will finally be able to relax.”

Eventually it does.

They do not relax.

They aim for £15 million.

Why?

Perhaps the original goal was never simply financial.

• Building the company provided status.

• Identity.

• Excitement.

• A reason to get out of bed.

• A way of receiving admiration from a parent.

• A defence against feeling ordinary.

• The financial objective could be achieved.

The psychological function continued.

This explains why instructions that seem perfectly sensible can repeatedly fail.

The instruction addresses the declared problem.

The behaviour may be serving another purpose.

Until that purpose becomes visible, more instructions may simply produce more sophisticated forms of resistance.

The Relationship Can Reveal the Pattern

This is one reason human mentoring remains important even in an age of extraordinarily capable AI.

An AI system can analyse the problem.

It can suggest alternatives.

It can remember every previous conversation.

It can identify contradictions in what somebody says.

All of this can be extremely useful.

But relationships generate information of their own.

The way somebody relates to a consultant can sometimes reproduce the very pattern they are struggling with elsewhere.

A client who says employees never take responsibility may also ask the consultant to make every decision.

Someone complaining that nobody challenges them may become irritated when the mentor disagrees.

A founder who says the family avoids difficult conversations may repeatedly change the subject whenever succession arises.

These moments contain information that no questionnaire may reveal.

The consultant does not need to diagnose the client.

It may be enough to notice what is happening:

“We seem to have reached this point several times. You ask me to challenge you, and when I do, we move quickly onto another subject. I wonder what happens for you at that moment?”

Now the relationship itself has become a place where the pattern can be examined safely.

That is something rather different from supplying another answer.

This Is Where AI Becomes Useful

None of this constitutes an argument against artificial intelligence.

Quite the reverse.

AI becomes extremely valuable once we understand which job we are asking it to perform.

• It can analyse accounts.

• Compare scenarios.

• Challenge assumptions.

• Summarise meetings.

• Preserve the knowledge of older family members.

• Produce alternative strategies.

• Model the financial consequences of retaining an employee.

• Ask questions nobody in the family thought to ask.

It can take large amounts of repetitive production away from human beings while leaving the originating observations and intentions with them.

But it should not quietly decide which family member's definition of success is correct.

Nor should popularity do so.

Nor should profitability.

Nor should the consultant.

These are sources of information.

Judgment belongs somewhere else.

Ask Better Questions of AI

This suggests a different way for family businesses to use artificial intelligence.

Instead of asking:

What should we do?

Try:

What assumptions underlie each person's position?

What might each family member lose if this plan succeeds?

Give me three strategies based on three different definitions of success.

How would this decision change if preserving employment mattered as much as maximising profit?

What would we do if growth were not the primary objective?

What would enough look like?

Argue against the most popular solution.

What important factors are difficult to measure here?

AI then stops acting like an oracle.

It becomes a means of widening thought.

That may be one of its most valuable roles.

Use AI More. Obey It Less.

This produces an apparently contradictory conclusion.

Family businesses should probably use more artificial intelligence.

And defer to it less.

• Use it to calculate.

• Use it to remember.

• Use it to search.

• Use it to model possibilities.

• Use it to remove repetitive work.

• Use it to challenge assumptions.

But retain human responsibility for deciding what matters.

The same principle applies to algorithms, consultants, management fashions and financial models.

• Popularity is evidence.

• Profitability is evidence.

• Efficiency is evidence.

• Tradition is evidence.

None constitutes truth on its own.

And none can define success for every member of a family.

• A family business may discover that growth matters.

• Or continuity.

• Or independence.

• Or relationships.

• Or protecting a craft.

• Or providing employment.

• Or having sufficient money and enough time to enjoy it.

Most probably it will want several of these things at once.

That makes the problem harder to systematise.

It also makes it more human.

The most interesting family businesses have rarely survived by mechanically following whatever everybody else was doing.

They adapted.

They borrowed.

They argued.

They experimented.

They protected some things and abandoned others.

And occasionally somebody unreasonable appeared and said:

There may be another way.

The purpose of good technology (and perhaps good consulting) is not to silence that voice.

It is to help the family discover what lies behind it.