Publishing: when the bottleneck is the first reply
Some companies have a traffic problem and some have a response problem. This one was the second kind: enquiries arrived, but too much time passed between someone raising their hand and someone actually answering them — and in that gap they were lost. Names and identifying details are withheld: some of these projects were commissioned by third parties.
01 Where it started
A platform in the book world, with authors arriving from different channels, at different moments, at completely different stages: some with a finished manuscript, some with an idea, some already published, some with no clue where to begin. They all fell into the same funnel and received the same message.
The predictable result: those who were ready got bored, those who were not got scared, and whoever handled the enquiries spent their days working out who they were talking to instead of talking to them.
02 What was built
A contact-handling system where AI does the understanding: it reads what the person wrote and works out where they are in their journey, what they need now and what the sensible next step is. From there the message, the timing and the call to action all change.
The CTAs are no longer one-size-fits-all: they change depending on who is reading. An author with a finished manuscript is offered one thing; someone still writing is offered another. It sounds obvious put like that, but doing it by hand at real volumes is impossible — and it is exactly the kind of work where AI earns its keep.
03 The number
40% of the readers who enter the journey book an appointment. Four in ten: in a funnel where everyone got the same message and conversion was counted in low single digits, that is a different order of magnitude, not an incremental gain.
It is the case that best demonstrates a principle we repeat often: AI is not there to write more messages, it is there to send the right one to the right person at the moment they are willing to hear it. The work continues — this is the one project of ours still in active development on this front — and the data it gathers sharpens the system month after month.
04 If this sounds like you
This is not a publishing case: it is a case for anyone receiving more enquiries than they can follow up properly. Agencies, firms, training, services. The symptom is having a single message for audiences sitting at different points of the journey.
The question to ask is: how many of the enquiries you received in the last three months got an answer written for them? If the number makes you uncomfortable, there is room — and it usually shows quickly.
Does this mean AI writes to my customers instead of me?
No, and that would be a recipe for ruining a relationship. AI understands and prepares: who replies and what you offer stay yours. In projects that need a personal touch the system suggests, it does not send.
Do I need a lot of contacts for this to make sense?
You need enough that you cannot follow them all properly by hand. Below that threshold the honest advice is to handle them yourself: you convert more and you learn more, and what you learn will feed the system when you build it.