Rethinking publishing, from script to screen. Inkstair is an AI-powered CRM for modern publishing borned as hackaton project. While filtering manuscripts, crafting assets, and reducing reader friction Inkstair empowers publishers through AI.

Concept ideation & development, Surveys & User Interviews, Brand Ideation & Visual Identity, Product Strategy & Flow, User Interface, Prototyping , Video Prompt Ideation.

The adoption of AI is transforming how we manage, process, and retain information.

During a design hackathon, we were challenged to explore how AI could be meaningfully integrated into the publishing industry — not just to improve workflows, but to engage both creators and and readers.

With an open-ended brief and many possible directions, one of the biggest hurdles was defining a concept that was not only innovative, but also feasible and relevant in a real-world editorial context.

It hasn’t been easy to delve into a sector in which any creative solution seemed like a minefield of ethical implications and possible violations of authors' rights (we initially thought of a chatbot to chat with your favorite character from a book). And above all, by keeping high the desire not to degrade the reading experience by making it summarized or simplistic, but to enhance it.

That’s why after 3 inconclusive days of brainstorming, we decided to stop questioning and start asking, letting an extremely data-driven process to guide us.

For inkstair I worked side by side with a service designer. With the rest of the team we took care of aspects such as technical feasibility of the AI models, creating the case study and defining the name and tone of voice of the product. As designers, on the other hand, we both took care of all the brand identity, information architecture and market flows (with me putting focus on user flows and the service designer putting more focus on the service blueprint). We then took care of the whole research and ideation phase and the UX, UI and prototyping part. As well as the defining of the case study and presentation video, which was entirely done with AI plus minimal video editing.

To fully understand the needs and daily challenges faced by publishing houses, we conducted semi-structured interviews with professionals who work in the publishing industry on a daily basis. The interviews, conducted remotely, revealed numerous insights useful for guiding our planning.

The insights gathered were indispensable in defining the final product and gave us enormous motivation to continue.

We were fortunate to have an editorial translator on the team who was able to find the right hooks for us in less than 48 hours.

The primary research for inkstair involved:

3x publishing house owners

1x professional writer

36x readers

We used as Research Methods:

Semi structured interviews for editorial professionals

1 survey for readers

From semi-structured interviews with professionals who work daily in publishing and deal with its problems, it immediately became clear to us that there were a huge elephant in the room that no one seemed to have considered (perhaps because it is a very niche sector? We can't know).

Here the main insights:

• Publishing houses are overworked and saturated with manuscripts to read and review. Many of these are automatically rejected due to lack of time.

• Publishers are forced, when deciding which manuscripts to review, to opt for authors they know personally from events, making the system for extremely unbalanced, unfair, and disadvantageous for the publishing house itself.

• Secondary frustrations include managing social media and creating marketing campaigns to promote the publishing house's offerings. The obstacles for small and medium-sized publishing houses are lack of time and limited financial resources.

• Writers know how broken this selection process is and feel discouraged, frustrated, and ignored. They often convince themselves that having connections is the only way to get ahead of others and that talent doesn't matter.

We could have stopped at the first problems that emerged, but I didn't want to cut readers out of our project. So we created a quick survey that opened our eyes on several fronts.

Out of 39 respondents, 38 said they often didn't have the energy or time to read. What surprised me, however, was not the alarming data itself, but the reasons that emerged to support it. 60% of the readers in the sample said they didn't read due to lack of time (and we couldn't do anything about that). However, 77% said they did not read due to mental fatigue, and 36% due to lack of concentration.

Reducing the user's cognitive load is one of the main areas of focus for UX. I did not want to ignore such a relevant piece of data with such a wide range of possible solutions.

Furthermore, our secondary research process shows us that:

• Several studies confirm a global decline in the ability to sustain attention on long-form content.

• Cognitive load has increased significantly over the last 10 years due to:

• The amount of information we are exposed to (email, news, social media, messages, apps, dashboards, etc.)

• The growing complexity of digital tasks (e.g., managing multiple tools, continuous onboarding, dark patterns)

• The almost complete lack of so-called "decompression time."

• According to the World Economic Forum, the average person today consumes approximately 34 gigabytes of information per day, up from 5 GB in the 2000s.

• In 2023, a study by Oracle and Workplace Intelligence showed that 75% of people feel overwhelmed by the amount of data they process every day.

Furthermore:

• In Italy, according to MIUR statistics, diagnoses of learning disabilities (LDs) have increased from 0.9% to 6% in just four years. At the same time, diagnoses of ADHD (including adult ADHD) and other types of neurodivergence have increased. Consequently, awareness of the issue and related market needs have grown.

• Integrate artificial intelligence into work processes with the ultimate goal of optimizing publishing houses' workflows.

• Be able to select the 5-10% of manuscripts* worthy of human intervention, without any prior human involvement in the process.

• Help publishing houses manage their marketing, advertising, and promotion flows, both on social media and offline, ideally at a low cost.

The technical feasibility of the ideas was the main obstacle. The AI ​​models used were carefully chosen to ensure optimal results.

Generating images and marketing campaigns is a demanding task, especially due to computing power. Everything was integrated by deciding to manage the process via API.

To avoid one-size-fits-all solutions, a system of settings and preferences was developed to help the publishing house define its specific target audience and desires in terms of tone of voice and values. In short: personalization.

• Reducing critical reading issues (very long descriptions, socio-cultural references unknown to the reader, vocabulary too complex for the target audience).

• A seamless reading experience that integrates without "subtracting," and adds value to digital books.

• We wanted AI to act as a bridge between traditional reading and innovation, simplifying, integrating, and adding rather than summarizing, devaluing, and "sapping the soul" of works. A discreet, sensible, and non-invasive aid.

• The proposed product was already a service for publishing; we couldn't shift the offering to the readers and propose a second service for which it was the primary driver.

• Make reading clearer and more comprehensive for those who want it (even on a contextual/occasional basis) without degrading the reading experience itself.

• Allow authors to collaborate directly, adding value to their content and making it more relevant and inclusive.

By connecting to the publisher's email inbox, inkstair offers a three-step manuscript filtering system.

They evaluate, in order:

• Grammatical accuracy of the text and its affinity with the publisher's published genres, discarding manuscripts submitted "in bulk" and those not carefully curated

• The content's sentiment and matches it with the publisher's targeted sentiment, discarding volumes that don't align with the editorial tone of voice.

• The match between the keywords in the manuscript under consideration with those trending in the current world, particularly useful for medium-large publishing houses.

Leveraging the most powerful AI models currently on the market through API integration, inkstair enables the creation of customized graphics, book covers, promotional posts, event posters, marketing campaigns, and editorial plans.

Content is stored in a structured archive that can be filtered by category, making it reworkable even after a long period of time.

Publishers and their teams can organize content by project (displayed as folders) to further streamline workflows for urgent deliveries or medium- to long-term assignments.

It targets publishers and readers together, analyzing the works (now edited, revised, formatted, and ready for publication) and comparing them to the target audience.

This generates a clear and comprehensive list of potential points of friction for the reader (excessively long descriptions, socio-cultural references, overly complex vocabulary).

The generated suggestions can be further processed by AI, which will use them to generate alternative texts/explanations, or they can be sent in XML format to the author, who can edit them personally, adding further value to their works and embracing the needs of readers who are passionate about reading but tired, stressed, or less trained in text comprehension, whatever the reasons.

The generated alternatives don't replace the original text in any way. They integrate seamlessly into the reading experience, in the form of disappearing tooltips and dotted underlines that are perfectly distinguishable from current underlining systems. They are also visible on black-and-white e-ink screens.

The goal of all this is to give digital books that added value that the staunch supporters of paper and ink have not yet been able to identify.

The reading experience isn't diminished, but enriched: authors' collaboration can be as passive as active, if they choose to participate in the process. Contributions provided directly by the author are always highlighted in the tooltips themselves, as are those generated by AI, thus addressing one of the major ethical concerns surrounding AI adoption: when it's used to generate content, users want to be aware of it.

The brand identity was developed through collaborative brainstorming sessions. The team used a shared board to finalize the name, define the service's tone of voice, and the tagline, through generative workshops conducted by me and the service designer in the team.

Branding was subsequently developed by me and the other designer. Emphasis was placed on linear, legible fonts and colors that were serious and professional, yet also conveyed an innovative and modern concept.

The logo, a ladder composed of books, references the service's values: helping publishers scale their offerings to ever-higher quality; and helping readers reduce the friction that keeps them away from paper and digital media. Inkstair, these media become clickable, interactive, and more valuable.

The name was chosen with the goal of being easy to pronounce and memorize, easily positioned in terms of SEO, and, above all, in terms of deliverability (the .com and .AI domains have been verified and are currently free).

It's incredible how much this project taught me, from start to finish. Although fictitious and created for fun, it was an incredible professional training ground.

A team assembled on the fly, interviews with the right people by pure luck, and a lot of stress due to the tight deadline (while we were all working full-time jobs) nonetheless led me to the best creative period of my life.

I understood even more how research can open up new horizons and help you focus on real problems closely. How invaluable (especially in niche sectors like this) talking to those directly involved rather than assuming needs is. How dedicating 50% of the process to research can make the ideation phase easy, spontaneous, and interesting. I will carry this baggage with me forever.