An agency sends a content request. Twenty four questions in a document, with a friendly note asking the client to fill it in when they get a chance.
The client opens it, sees twenty four blank boxes, closes it and intends to do it at the weekend. Six weeks later four boxes are complete and the project has not moved.
The agency chases politely, absorbs the delay, and eventually writes the answers itself for nothing. This happens on almost every project and it is planned for on almost none.
The first failure: a blank document is an impossible ask
Writing from nothing is the hardest form of the task. A client who could easily correct a draft cannot face producing one, and the request is what causes the delay.
The second failure: the client is not the bottleneck by choice
They have a job. This is somebody else's project with no deadline that costs them anything. Politeness without a mechanism produces six weeks of silence.
The third failure: the agency asks the wrong person
The person who signed the contract is rarely the person who knows the pricing, the process or the policy. Every question routed through them adds a hop and a week.
The fourth failure: nobody harvests what exists
Website pages, brochures, email templates, past enquiries and proposals. Most of the answers already exist somewhere and a first draft can be assembled from them in an afternoon.
The fifth failure: the request is one large task
Twenty four questions at once is a project. Five at a time, with a short deadline each, gets completed. The shape of the ask determines whether it happens.
The sixth failure: there is no consequence for a slip
Without a stated date that moves the launch, the content stage has no clock. This is the single change that most reliably fixes the problem, and it belongs in the proposal.
The seventh failure: approvals are not scoped
Content that must pass a compliance, legal or clinical review adds weeks nobody planned for, which is exactly the ground in scoping a client assistant project properly.
The eighth failure: the agency does not know what to ask for
A generic list produces generic answers. The right list comes from the client's own enquiries, which is what the questions every small business website should answer sets out.
The ninth failure: the interview is never offered
Forty five minutes on a call, recorded and transcribed, usually produces more usable content than six weeks of a document. Clients will talk and will not write.
The tenth failure: the drafts are not written as answers
Marketing copy repurposed as an answer reads wrongly and gets rejected. Writing them as plain answers from the start avoids a rewrite.
The eleventh failure: the cost is absorbed silently
The agency ends up writing the content, unpaid, and calls it goodwill. Pricing it as a variable changes the conversation entirely, as set out in pricing assistant builds for clients.
The twelfth failure: the same problem recurs at every update
A client who could not supply content for the build will not supply it for the refresh either, which is why the maintenance arrangement has to assume extraction rather than delivery.
The fix: draft first, ask for corrections, and collect the real questions
An assistant deployed early collects the client's genuine enquiries while the content is being written, which both proves the topic list and gives the client something concrete to react to instead of a blank page. The same approach of drafting and refining client answers is described in how financial advisors use chatbots to answer client questions. Start at try, see features, and the sector view sits under agencies.