Like Their Only One

A customer returns with a request that changes the plan. That is a revealing moment for a company. Someone has to understand what changed, remember what still matters, and decide what to do about it.

Consider an illustration. A business has engaged an organiser for a corporate golf day. The purpose is to spend unhurried time with existing clients. Later, the host asks whether a valued guest who cannot play could still join the occasion. The request is perfectly reasonable. It also tests whether the organiser understands why the occasion exists.

The host should be able to raise that need without becoming an event coordinator. The organiser should investigate a suitable arrangement, explain any limits and carry the agreed change through. The history of the engagement should help.

That is the kind of difference the Songbird name must stand for. We want every company we build to give customers the care its founder would give a single, deeply understood account. Their circumstances should matter in the decisions the company makes, even when the founder is occupied elsewhere.

This is a build standard we are committing to. It is also a standard against which customers can find us wanting. The scenarios here are illustrations; I am not presenting a completed Songbird customer case.

The promise reaches beyond the founder

It is easy to explain Songbird entirely from the founder's side: their knowledge gains a business, their preferred work gets more room, and the surrounding administration takes less of their day. But their customers need a reason to choose that business. The founder's easier working life cannot be the entire proposition.

The missing part is what that freedom and knowledge do for someone else. A customer should receive careful preparation, an appropriate response to a changed situation and reliable follow-through. The founder's standards should influence ordinary service, rather than appear only when the founder personally intervenes.

Attentive service has a long history. Paul Graham's advice to early startups placed the customer's whole experience above the product alone, with individual attention central to that experience.¹ Our burden is to make such care a dependable property of the company as its work grows.

The foundations are set out in Experience Is the Unit and Software in the Middle. The software filter narrows an offer when the application decides which customer needs can be considered. Demand-pulled code supplies a tool when an understood engagement requires it. Testimonial-back design starts with the truthful account each participant should be able to give afterwards, then asks whether those accounts can coexist. Experience is the unit of value: success belongs to what people receive across the engagement.

Those foundations lead to eight commitments. Each has to survive contact with a customer.

Eight commitments we must keep

1. We read your situation before acting

   Every customer has a continuing folder: their purpose, relevant preferences, recent changes, decisions and anything we still owe them. Whoever picks up the work must consult it. A returning customer should spend their attention on what has changed, with the earlier context already available.

   Remembering includes accepting correction. A preference recorded last year may be wrong for this year. We must distinguish what the customer told us from what we inferred, and allow current information to change the approach. A folder earns its place when it improves the next action.

   What you notice: a new person can continue the work, and a correction you made last time stays corrected.

2. We agree on an experience that works for everyone involved

   You should know what worthwhile result we are working towards. So should other people whose participation makes it possible. An organiser cannot call a day successful because the host was delighted if guests were pressured or venue staff had to absorb undisclosed work.

   Starting from customer benefit has a precedent in Bryar and Carr's account of Amazon's Working Backwards approach.² Our commitment is to make the benefits compatible and the promises explicit. Participants remain free to tell us that the result disappointed them.

   What you notice: we can explain what a good result means for you, and achieving it does not quietly transfer the burden to somebody else.

3. We give the problem serious attention within your real constraints

   We should consider what excellent preparation could accomplish before settling for the usual arrangement. That means investigating options, anticipating awkward moments and making thoughtful choices while respecting the customer's ordinary budget.

   For the illustrated golf day, better attention might produce a more welcoming introduction, a less rushed programme or a host who knows when to leave a conversation alone. It need not purchase a more extravagant afternoon. The venue still has limited capacity. Guests still have their own wishes. A company must confirm what it can deliver before promising it.

   What you notice: the preparation addresses something important about your occasion, without creating extra work or unwanted fuss for you.

4. We hold the responsibility until it is resolved

   Sending a message to a supplier is an action within an obligation. If a guest's arrival arrangement remains uncertain, the company must keep pursuing it, tell the customer what remains unresolved and seek a decision when one is needed.

   The obligation survives a handover, a delayed reply or an agent stopping. Closure requires evidence that the agreed matter is settled, or an explicit conversation about a changed commitment. Otherwise the customer becomes the person remembering which of our completed tasks failed to produce a result.

   What you notice: you can ask where matters stand without first reminding us that they exist.

5. We adapt the delivery to the engagement

   A worthwhile request deserves consideration on its merits. The customer's first answer should explain feasibility, purpose and limits. The absence of a button in our current application should not settle the question.

   Tools may be temporary; the service commitment persists. The customer need not learn another application for every variation we accommodate. Nor does flexibility mean agreeing to everything: if the requested arrangement would compromise another guest, exceed the agreed scope or rely on an unavailable supplier, we should explain that and explore a feasible alternative.

   What you notice: an unfamiliar request receives a considered response, with any resulting work handled by the company.

6. We turn experience into usable, written know-how

   Each engagement should leave the company better prepared. We must record what mattered, what happened and what we would reconsider. Relevant learning should inform later work, including work for another customer, without exporting anyone's private circumstances.

   This has to be tested through use. Anthropic's guidance on reusable agent expertise recommends evaluating representative tasks and improving the material in response to observed difficulties.³ Our customer standard asks for a similarly concrete result: identify something we learnt, then show the later decision it improved. A growing library by itself proves very little.

   What you notice: we avoid a previously identified problem or make a better choice because we have encountered something relevant before.

7. We begin with people who know what good looks like

   A new company needs sound judgement before it can repeat sound judgement. We should draw on excellent practitioners early, capturing the distinctions they make and the reasons for them. Their contribution must reach the work, rather than remain a name attached to the business.

   In an event company, an experienced organiser may notice that a schedule leaves nobody able to welcome late arrivals, or that an introduction would embarrass both parties. Customers should benefit from such judgement even when the practitioner is elsewhere. We have to demonstrate that transfer; recording an expert conversation does not establish it.

   What you notice: the service shows practical understanding of situations that a generic plan would miss.

8. A human champion remains answerable

   Every company needs an identifiable person who stands behind its promises, can make consequential decisions and can arrange a remedy. Customers should know how to reach that responsibility when ordinary handling fails.

   This person does not have to approve every routine action. They do have to ensure that delegated work is competently done and that failures receive attention. A reply explaining which automated component went wrong is insufficient if the customer still needs the company to put matters right.

   What you notice: someone can decide, explain and act when your situation requires judgement.

The commitments depend on one another. Remembering a customer without adapting to them makes the memory ornamental. Adapting without follow-through creates promises that drift. Follow-through without learning makes customers endure avoidable mistakes again. And delegating all of this without an answerable person leaves the customer carrying the risk.

What three visits should reveal

Return to the anonymous event organiser. The sequence below is an illustration of the standard, including its successful outcomes. It is not an account of a delivered Songbird engagement.

At the first visit, the company earns a useful brief. The host wants time with established clients and wants colleagues to enjoy the occasion too. Several invitees are inexperienced players. These details should affect the programme: enough support for beginners, room for conversation and clear arrangements so the host can concentrate on guests.

The organiser checks that the venue can support that plan. If a sponsor participates, its involvement must suit the occasion. Nobody should discover on arrival that a relaxed afternoon has become a sales presentation. Before the day, the company has an agreed purpose, confirmed arrangements and a person answerable for delivery.

Now imagine a failure within the same illustration. Every invitation arrives, bookings reconcile and all scheduled messages go out. The dashboard is entirely green. But beginners feel stranded, the host spends the afternoon sorting logistics and a sponsor repeatedly interrupts the conversations the host wanted to have. The operational record cannot excuse that result. Those failures belong in the company's account of the day.

At the second visit, the prior engagement must do useful work. The company begins with what it learnt, checks what has changed and proposes a different arrangement where appropriate. The host should not have to repeat the problem with beginners or remember to ask whether it has been addressed.

Suppose the host now wants a shorter occasion and less competitive play. The organiser can explain why its proposal has changed. It can also show what still requires confirmation. Continuity means the company's understanding develops; it cannot simply repeat the previous event with a warmer greeting.

At the third visit comes the unusual request: the valued guest who cannot play would still like to attend. The organiser already understands that time together is the point. It investigates whether the guest could join a suitable part of the programme, checks access and timing with the venue, and agrees the arrangement with the host.

If that is feasible, the host receives a settled plan. If it is not, the organiser explains the constraint early enough to choose another approach. Either way, the request stays with the company until a decision and its consequences have been handled.

The folder now contains more than a booking history. It shows why decisions changed and where responsibility remains. Its value would become visible again when another worker takes over, or when a different client has a similar participation problem.

A customer's thank-you note would be the clearest account of whether that attention mattered. It might describe being able to spend the afternoon with guests, or relief that an awkward request was handled. We must wait for the customer to supply that account. Writing an attractive version ourselves would only demonstrate that we understand the story we hope to earn.

The company has to improve between engagements

That distinction changes what early work is for. A subsidised engagement can be a deliberate investment in learning: the customer receives an agreed service, while the company gains evidence about delivery and develops know-how it can use again. The commitment to serve that customer still applies in full.

The investment needs a boundary and a return. We should be able to identify the useful learning, account for the effort it required and see whether it helped later. If every new customer requires the founder to invent the service again, calling the work a learning investment has not made the company more capable.

A release labelled alpha, beta or version two tells us where software stands. It says little about whether the company understands its customers, holds its promises or learns. These terms are clues to inspect, especially when every conversation about progress ends at the next release. They are not a verdict on a business.

Steve Blank's June 2026 account of his Stanford class describes teams mistaking polished deliverables for understanding stakeholders and validating demand.⁴ Songbird needs to resist that temptation in its own work. An impressive demonstration can clarify the promise. A customer engagement has to test it.

Apply the remove-the-AI test honestly

Good service existed before AI. A capable organiser could perform every action in the illustration. That is why the example cannot, by itself, establish that the company is AI-native.

The test concerns the complete offer. Could we sustain the same preparation, specificity, response time and continuing attention for the same customers on the same commercial terms if AI were removed? If the answer is yes and only some administrative effort increases, we have described a conventional service using useful automation. We should say so.

If the offer depends on AI carrying substantial preparation and coordination, the dependence must appear in the evidence. Count the founder's interventions, the review, the corrections and the supplier work. A carefully concealed pile of human effort can produce a lovely customer experience while disproving the intended economics.

Bessemer's AI-native services framework makes a relevant distinction: access to agents does not establish an attractive business; delivery economics, defensibility and execution still matter.⁵ Our standard adds a customer obligation to that discipline. Attention must remain useful and dependable while the company learns how to provide it sustainably.

For Songbird's selection of companies, this means looking for valuable customer differences a business can respond to, expertise it can retain, and a promise people have reason to trust. A familiar subscription model or a fast software launch is insufficient evidence. The competitive opening may be an incumbent's difficulty adapting an established package, but we must assess it against a competent competitor with access to the same AI. Relevant learning and trusted delivery have to explain why customers would keep choosing the Songbird company.

What the name should come to mean

My co-founder has pressed the distinction between producing a product and building a company, and the deeper importance of meaning and trust. His view of brand as intellectual property belongs here: reputation contains accumulated expectations about how a business will behave.

For the illustrated host, the meaning of the service is time with people who matter. Trust grows when the organiser understands that purpose and can be relied upon when circumstances change. A new tool may improve the work. The customer should continue to recognise the care.

That is the durability we want. Methods can evolve while the promise stays recognisable. The name should remind a customer of sound decisions, responsibilities carried and problems put right. We cannot create that association by announcing it.

It also gives the founder something worth building. Their judgement can influence more engagements than they could attend personally, while they remain responsible for the company expressing it well. This is how the benefit to the founder and the benefit to the customer should meet.

Songbird must supply the first evidence

We cannot credibly ask founders to build this way while treating our own enquirers as interchangeable entries in a sales sequence. A person returning to Songbird should find that their circumstances, corrections and outstanding questions have carried forward. Work we promised should have an owner and a result.

I cannot yet point, on the evidence behind this article, to a completed Songbird engagement demonstrating all eight commitments. That limits what we can claim today. The standard is explicit; its complete demonstration remains work we owe.

The first evidence can come from a real continuing enquiry to Songbird. We should be able to show, with the person's permission, what they asked, what we understood, what we undertook and what actually happened. The customer must be able to correct our account.

For that engagement to count, the evidence needs to answer each part of the standard:

• Understanding: which earlier facts and corrections informed the next response?
• Shared purpose: what result was agreed, and were others affected by the work treated fairly?
• Attention: what useful preparation changed what the person received?
• Responsibility: which promises reached an evidenced resolution, and which remain open?
• Adaptation: how did we handle a meaningful change without making the person coordinate us?
• Learning: what written lesson improved subsequent work?
• Judgement: where did experienced human practice improve the delivery?
• Accountability: who answered for the result and acted when something needed correction?

Alongside those answers, we need an honest account of the work it took. One satisfied person is a valuable beginning. Repeating the service under changed conditions, with useful knowledge carrying forward and manageable intervention, is stronger evidence of a company.

This is why the standard matters before the success story exists. It tells us what to build towards, gives founders a clear obligation and gives customers grounds to challenge us. It makes the Songbird name answerable to conduct.

Judge us by what happens when you return, when your circumstances change and when a promise needs attention. If you can leave the matter with us, confident that we understand why it matters and will see it through, we will have begun to earn the feeling behind Like Their Only One.