The Folder You Own

The hardest thing on a blank application form can be the question that sounds easiest: what are you good at?

You have done useful work for years. People have relied on you. Yet when someone asks you to explain your value, you reach for the title on your last payslip. Remove that title and the answer can seem to disappear with it.

I want Songbird's first encounter with that person to produce something worth having before we discuss anything they might buy: a written account of what they bring, made with them, corrected by them, and handed over as a file they can keep.

This is the service we are designing, not a report of completed personal-retelling results. We have experience clarifying people's business ideas. A truthful account of the person is a different craft, and its first complete exemplars still need to be made by hand. The worked session below is an illustration.

The work you stop noticing

Polanyi's account of tacit knowledge starts with the gap between knowing something and being able to explain how you know it. You may recognise a face while being unable to specify the features that made recognition possible. Skilled judgement has similar difficulties of articulation.¹ Asking for a skill list can therefore miss the judgement that actually makes someone useful.

There is another difficulty. You know the world you have worked in; you may not know where your ordinary competence would be unusual or helpful elsewhere. The nurse knows nursing. A family trying to make sense of several services has a different starting point. Connecting a capability with that need requires looking at both sides. It does not license telling every nurse that she possesses a unique commercial asset.

And then there is the story you have become used to telling about yourself. A job title can become the boundary of what you notice. An outside perspective can reveal something you missed. The Johari window calls attention to differences between what a person and others perceive; Luft described it as a heuristic, not a machine for finding somebody's true identity.²

The outside eye has work to do. It must look at the evidence, propose a connection and survive correction. Confidence alone is cheap.

Put the work where both of you can see it

Imagine a former customer-service manager who says she enjoyed customers but has no desire to go into sales. She brings a short presentation she made about a recurring handover problem. It contains no customer details or former employer's confidential material; she has recreated the relevant example in her own words.

An interview could turn that into a list of attributes. I would rather spread the material out as a case file. The presentation sits beside a note about one difficult customer conversation. A possible role sits beside both. The person can point to the connection and say it is wrong.

The first sketch says she might lead customer relationships and manage a small support team. She objects to the team. It is tempting to record that as rejection of the whole possibility, especially if the system only has a like button and a dislike button. But her objection is much more useful than either.

She explains that she enjoyed finding out what had gone wrong, showing the customer the next sensible step, and arranging a clear handover. She disliked chasing sales quotas and being responsible for everyone's roster. The corrected page becomes:

That is already a better account than “good with people”. It tells the next reader something they could act on. It also leaves room for a mistake: perhaps she likes explaining only when she has deep knowledge of the subject. That question belongs on the page before somebody turns her into a general customer adviser.

Now make a possible morning. There is preparation, a customer conversation, a written handover and a clear finish. She can say which part she would welcome and where the imagined day still feels wrong. Recognition supplies cues that an abstract request does not; Nielsen Norman Group's guidance explains that distinction without claiming recall never works.³

Keep the first role sketch. Mark it superseded and explain that the business interested her while staff management did not. The earlier page now protects the corrected understanding against the same mistake next time.

A page made for a partner

Then she says she needs to discuss it with her partner. The next useful thing is a page for that conversation.

With a business partner, the page should make responsibilities discussable: which customer conversations she might hold, who would find suitable enquiries, who would own follow-through, and which gaps remain unresolved. It must not quietly allocate unpaid work to the other person.

With a life partner, the questions change. What time would this occupy? What would happen to evenings? Is any household income at risk? What has actually been agreed, and what is still imagination? A useful page gives the partner room to disagree. A personalised sales argument tries to remove that room.

The person chooses what to share. An expert can later receive the relevant material and begin with the difficult question rather than collecting the same biography again. Being helpful does not give Songbird blanket permission to invite people into the folder.

The description must be able to fail

Consider this account of our illustrative manager: “You are a resilient, people-centred problem solver whose experience can create meaningful impact.” It is pleasant. It does almost no work.

Forer's classroom demonstration showed the danger of treating acceptance as validation: students received the same supposed personal sketch and still rated it highly. A description that fits broadly can feel individual.⁴ AI makes it easy to produce a polished version of that mistake.

A useful retelling is narrower:

In the handover example, you noticed that the customer was being passed between people without anyone owning the next step. You made the responsibility explicit and explained it in language the customer could use. That suggests a strength in clarifying service handovers. It does not yet establish that you want to sell, manage staff or advise outside the field you know.

Every part can be challenged. Did she identify the problem herself? Was it a repeated pattern or one lucky intervention? Did she enjoy the work, or was she rescuing a situation she would rather never repeat? Is that capability useful to a specific group that lacks it? The retelling becomes more valuable when those answers change its words.

A name might eventually help. Customer guide, service interpreter, handover specialist: each catches something and misses something. Offer them as candidates. A person should be able to reject every one without the service interpreting refusal as resistance to insight.

A file, with no hook

The folder must leave with its owner. In practical terms I want a readable offline copy and an editable text copy, with the agreed evidence, current account, earlier corrections and outstanding questions included. A downloaded file that opens without a Songbird account is a much stronger demonstration than an ownership slogan above an export button that does not work.

The full folder and a shareable selection should be different things. Someone may want an adviser to see the role question without reading everything said about family commitments. A copy already given to another person cannot be made unseen; the choice should be clear before the handover.

There are three legitimate ways to leave the first part: explore piloting a company, offer know-how to someone else's, or go away with a better account of yourself. The third needs the same complete file and the same goodwill as the first two.

Why would Songbird pay for this?

Because a good folder can create several kinds of value together. The person receives the immediate result. Songbird learns which questions and forms of explanation help, with permission to reuse only what has been agreed. People may choose to become future founders or experts. A generous experience may earn a referral.

Those are reasons to test the offer, not to declare it profitable. The cost is the whole delivery: preparation, model use, human review, correction, support and any promised follow-through. A few dollars of inference would not make an hour of founder time disappear from the accounts.

Paul Graham's advice to do early work manually offers a useful precedent: serve people closely enough to learn what should later be automated.⁵ For Songbird, that means making the founders' folders and a friendly outsider's before admitting a large public queue.

Set a cost allowance for a completed folder and a weekly admission limit supported by both money and review capacity. Count the distribution of all three exits, useful corrections, voluntary referrals and lessons that actually improve later work. If the cost cannot be supported, or quality requires more human attention than is available, reduce or stop new admissions. Finish the gifts already promised. Do not turn the download into an upsell to repair the budget.

A clinically trained adviser should help set the care floor: truthful, non-manipulative, consented, clear about the limits of this work, and careful about documents that belong to somebody else. Songbird still owns the warmth. A folder can be technically correct and leave a person feeling processed.

I would be proud to build a business that looked closely at someone, helped them say something true they could not quite say alone, and let them carry it away. That is a first encounter worth making. The next step is small enough to do with paper: bring one story, write one specific observation, and leave room for the person beside you to change it.