How book publishers can automate royalty statements and rights work

Which publisher work an AI worker can take over

Most of the office time in a book publisher goes to the steps between a sales report and an author's statement, and between a signed contract and the next due date. An AI worker can handle much of that routine. In a trade, academic or educational list, it can gather sales files from distributors and retailers, match each line to a title and a contract, apply the royalty rules, and draft the statement. It can send reminders when a manuscript is due, when an advance instalment falls due, or when a rights licence is close to expiry. It can also log incoming manuscript submissions and sort permission enquiries into the right queue.

What it cannot do is decide. Whether a clause has been triggered, whether a reserve against returns is fair, whether an option has lapsed, whether a permission can be granted, and whether a manuscript should be published all need a person who knows the contract and the author. The rights team also checks the royalty figures before any statement goes out. The AI worker makes that check faster by showing the numbers and the rule behind each one. It does not replace the check.

For the office team, the change is in the type of work. The royalty administrator stops keying data and starts reviewing the exceptions the system flags. The rights assistant stops reading every enquiry from scratch and starts answering the ones that need a decision. Fewer hours go to copying figures between spreadsheets, and more go to the questions that carry risk: disputed sales, unusual contracts and authors waiting on a payment.

The publishing chain, stage by stage

The work runs in a fixed order. A contract is signed, a manuscript is delivered, a book is made and listed, sales happen, royalties are calculated, statements go out, and the money moves. Each stage creates the information the next one needs. The walkthrough below follows that order and says what staff do now and what an AI worker does instead.

Contract signing and advance schedule

Today, a contracts administrator or rights manager reads each signed agreement and keys the key terms into a royalty system or spreadsheet. That includes the royalty base, the rate or escalator, the advance and its instalments, the reserve rules, and the reversion or option dates. Errors at this stage travel into every later statement. An AI worker can read the contract summary a person has approved and build the milestone calendar from it. It can then flag each instalment and option date ahead of time. A person still confirms the terms against the signed document, because a summary can miss a rider or a side letter.

Manuscript submission intake

Submissions arrive by email, by agent, through a web form, or through a submission platform. An editorial assistant usually opens each file, checks it against the submission guidelines, records the title and author, and sends an acknowledgement. An AI worker can receive submissions from the set channels, check that required files are present, log the record, and send the standard acknowledgement. Scholarly publishers already use submission platforms for this. Aries Systems describes its Editorial Manager product, on its own page in 2026, as covering books as well as journals. The editor still reads the work and decides whether to take it further.

Delivery dates and editorial milestones

Production coordinators track when copy-edited files, illustrations, index and proofs are due, and they chase authors who are late. This is often done by email and a shared calendar, which means dates slip quietly. An AI worker can keep one schedule per title, send reminders to the author or editor before each deadline, and flag slippage to the production coordinator. It cannot make an author deliver on time. It can only make the gap visible early, so that the publisher can talk to the author while there is still room to move a publication date.

Metadata and listing

Before a book can sell, its metadata must reach distributors and retailers in the right format. The industry standard for this is ONIX. The Book Industry Study Group says that the U.S. and Canadian markets have moved to ONIX 3 and no longer support ONIX 2.1. Its 2025 guidance also points out that accessibility details can only be conveyed through ONIX 3 in the EU context. An AI worker can check that each title record has the required fields, flag missing subject codes or accessibility data, and compare the record with the last version sent. A metadata specialist still owns the content and signs off the file.

Rights and permission enquiries

Rights managers receive a steady flow of questions. Can a translator quote a passage? Can a film company option the title? Does a foreign publisher hold the territory? Often the answer is in the contract and the rights schedule, but finding it takes time. An AI worker can read each enquiry, match it to the title and the rights record, and draft a reply that states what the record shows. It can also route enquiries that touch a sold, reserved or expiring right to the rights manager. The decision to grant, refuse or price a right stays with a person.

Sales file intake from distributors

Each distributor and retailer sends its own sales report, in its own layout, on its own schedule. Someone opens each file, reformats it, and adds it to the royalty run. Many publishers still do this in spreadsheets. An AI worker can pick up the files as they arrive, detect the layout, map each line to an ISBN and a format, and report any line it cannot match. Many royalty platforms advertise this kind of automatic import. The reliability depends on the reader's own files, and the first few runs should be compared with the old method.

Royalty calculation and rights team check

This is the stage where the risk sits. A royalty can be based on list price or on net receipts, and the net figure depends on discounts, returns, distribution costs and sometimes a sub-licence share. Royalty rates can rise in steps once sales pass a set level. Joint accounting can recoup one book's advance from another book's royalties. An AI worker can apply those rules to the sales lines, show the working for each title and author, and list the lines that break a rule. The rights team then checks the totals and the exceptions before the run is released. The check is not optional, and the statement should not go out until a person has approved it.

Statements, payments and author queries

Once approved, statements are generated, sent and filed. Authors and agents then ask questions, and some disputes follow. Payment timing also matters. Distributors often pay after the sales period closes, so a publisher must wait until the slowest payment has arrived before it runs royalties. An AI worker can produce the PDF or Excel statements, send them to the agreed recipients, and store a copy against the contract. It can answer simple questions about where a figure came from. It should not reply to a dispute, offer a settlement or explain a clause. Those replies go to a person.

Accounts and reconciliation

Finance then moves the money, records the royalty expense and holds the liability until it is paid. The royalty system produces the bills and the accounting system takes the payment and bookkeeping steps. An accounts payable clerk currently matches the bills to bank payments by hand. An AI worker can prepare the bills in the format the accounting system expects, match payments to statements, and flag any payment that does not agree with the approved total. The accountant still signs off the reconciliation and the payment run.

Staff today vs with AI workers

The table below lists the office roles that a book publisher usually employs, under the titles the trade itself uses. It does not give headcounts or salaries, because these vary widely by list size and by country. The right-hand column is the one to read with care. It describes the checks that stay with a person, and those checks are what keep the royalty and rights work accurate.

The highest-value automations, ranked

Not every task is worth automating first. The table ranks the automations by the mix of effort to set up and value once running. Effort means the work needed to connect files, agree rules and test. Value means how much routine work leaves the office team and how much risk the change removes. The ranking is a starting point. The reader's own sales files and contract mix decide the real order.

Automations 1 and 2 carry the most effort, because they depend on clean data and agreed rules. They also carry the most value, because they remove the copying that causes most errors. Automations 3 to 5 are quicker to set up and help the team keep to dates. Starting with the low-effort items gives the team early practice without touching the royalty figures.

Compliance and regulatory points

Book publishers work across borders, so the rules that apply depend on where the author, the reader and the publisher sit. Three areas matter most for this work.

The first is personal data. Royalty systems hold author names, addresses, tax details and bank details. Under the EU's General Data Protection Regulation, Regulation (EU) 2016/679, personal data must be processed lawfully and fairly, collected for stated purposes, kept to what is needed, kept accurate, and not stored longer than necessary. The publisher must be able to show that it complies. An AI worker that reads contracts and sales files therefore needs access limited to the fields it uses, and a retention rule for its logs and drafts. Publishers outside the EU that handle data of people in the EU can fall under the same rules, so the reader should check the position for its own readers and authors.

The second is accessibility of ebooks in the EU. Directive (EU) 2019/882, the European Accessibility Act, has applied since 28 June 2025 and names e-books and dedicated reading software among the covered services. Member states had to transpose it into national law by 28 June 2022. The directive also sets transitional periods for some existing arrangements that run to 2030. For a publisher, this means the metadata that describes accessibility features needs to be correct and passed along the supply chain. The rights and metadata team should check that the automated listing checks include these fields.

The third is the metadata standard itself. ONIX 3 is the format that the book industry now uses for product data in the markets the Book Industry Study Group describes, and ONIX 2.1 support has ended in those markets. A listing that is sent in an outdated format can be rejected or misread by a retailer. The rule is set by the industry standard body rather than by a statute, so the publisher should confirm the current version with its distributors.

Tax on royalties is a further area. Withholding and reporting rules on payments to authors differ between countries and depend on the author's residence and the publisher's location. This report does not set those rules. The publisher should confirm them with its tax adviser for each country where it pays authors, and the AI worker should be set up to show the tax treatment as an input for a person to confirm, not as a decision.

This explains the rule in general terms; it is not legal or tax advice for your case.

How the transition happens without disruption

Start with the team that already does the work. For the first few royalty periods, the existing process keeps running, and the AI worker produces its own draft in parallel. The royalty administrator compares the two, lists every difference, and decides which one is right. This shows where the rules need tightening before the new system is trusted with a live run. It also gives staff the chance to learn the tool without carrying a deadline alone.

Next, move the exceptions to the people who already own them. The royalty administrator receives the exception list, the rights manager approves the run, and the accountant signs the payment. Nobody is asked to give up a check to make the system faster. The change is in what each person looks at first. Once the exception list is short and reliable, the routine lines can pass straight to the approval step, and the review becomes the main task.

Only then should the roles shrink, and only by moving work, not by cutting it at a stroke. As copying and chasing fall away, some roles spend more time on rights questions, author relations or metadata quality. Where a role really does become smaller, the change needs care. Notice, consultation and severance rules in the country concerned apply to any reduction in roles or hours, and the change should be agreed with the employee or checked with a lawyer before it is announced.

Common mistakes

The most frequent errors are predictable, and most of them come from moving too fast.

• Automating the royalty calculation before the contract rules are written down in a form the system can apply.
• Calculating royalties on the calendar date instead of on the date the distributor's payment has actually arrived.
• Treating an advance milestone as a fixed date when it depends on a delivery that has not happened.
• Keeping the old spreadsheet running in secret, so that two sets of figures exist and nobody knows which one was sent.
• Letting the AI worker reply to an author who is disputing a figure, instead of routing the dispute to a person.
• Holding author bank details and tax data in a shared inbox or an unlimited log.
• Skipping the rights team check because the figures look right on screen.

Each of these mistakes can be avoided with a short list of rules agreed before the first live run, a named owner for each exception, and a record of who approved each statement.