When the slide library becomes infrastructure

Every corporate team has a presentation queue. A proposal due tomorrow, a leadership alignment on Friday, an onboarding session next month. The queue is usually treated as a capacity problem: not enough people to make slides. Most of the time, it is a problem of origin. Every request starts from scratch.

The queue that never got shorter

In 2021, when Marja Amim joined the Business Design team at TOTVS, the backlog of visual requests was huge. Sales proposals, RFP responses, positioning materials and internal presentations all competed for the same calendar. Producing each piece faster would have fixed the week, not the month.

The decision was to change the question. Instead of "how do I deliver faster?", Marja started asking "what needs to exist so people do not depend on me to get started?". The answer became a living deck, updated every week, with templates for many different uses.

Enlarge

What was inside the deck

The library was not a folder of pretty slides. It was an internal product, with its own navigation architecture. To give the material a longer life, the placeholder text often doubled as the instructions for using the slide. For example: TITLE, Subtitle, three lines of body copy, item 1, item 2, item 3, footer, call to action, tag, and so on.

• A clickable index: templates organized by use, such as cases, proposals, services, solutions and agenda.
• Instructions for use: right at the opening, so nobody needed a meeting to learn how to use it. The material sold itself.
• Support libraries: small icons, GIFs and graphic elements ready to apply.
• Constant updates: every week or two we gathered what was new to keep the material aligned with the products and with the current messages and identity.

Who this resource is for

A visual resource library is not just a designer's tool. It is a tool for anyone who needs to explain something at work, which today means almost everyone:

• Sales and presales teams, who build proposals, RFP responses and value presentations on tight deadlines.
• Consulting and implementation, who need to explain scope, stages and next steps to the client.
• Leaders, in executive alignments, results presentations and strategy communication.
• Marketing and product, in positioning, launches and sales enablement materials.
• HR and corporate learning, in onboarding, training and internal announcements.
• Event organizers, from sales conventions to internal gatherings.
• Designers themselves. Whoever builds the library also gains from it: they stop redrawing the same slide for every request, start from ready-made pieces and save time for what is genuinely new in each project.

The last group is often forgotten, and may be the one that gains the most. At TOTVS, the library was created precisely to speed up the requests that reached Marja. It ended up speeding up everyone who depended on her, and became the showcase of the Business Design team inside the company.

Why does this speed up the everyday work?

The speed comes from separating two jobs that usually happen together: deciding what to say and deciding what it will look like. When the form is already solved by a reliable system, the team's time goes back to the content. This shows up in three recurring situations.

There is old and consistent evidence that visual support changes the outcome of a presentation. In a University of Minnesota study with 3M, presentations using visual aids were 43% more persuasive than the same presentations without them.

Proposing

In sales proposals and RFP responses, deadlines are short and the margin for brand mistakes is zero. A template with the commercial and technical proposal structure already solved keeps someone from spending the night aligning text boxes. The night goes to the argument.

And the proposal is not read only by the people who were in the meeting. A Gartner survey of 632 B2B buyers found that buying groups bring together 5 to 16 people from up to four different functions, and that 74% of them experience internal conflict during the decision. Groups that reach consensus are 2.5 times more likely to close a high-quality deal. In practice, the proposal becomes the material that circulates among these people when the seller is not in the room. It has to help the client convince itself from the inside.

Form also shapes how trustworthy a proposal feels. In a classic cognitive psychology experiment, the same statements were judged true more often when they were written in a way that was easy to read. Visual clarity does not replace an argument, but a confusing proposal charges the reader an effort they tend to discount from the content.

Explaining an understanding

Many people can only evaluate an idea once it takes shape. A ready-made visual resource makes it possible to materialize an understanding in the meeting itself, and the group starts discussing the same thing. Marja Amim calls this the Cognitive Mirror: form helps people see their own reasoning before it turns into implementation cost.

Teaching

Internal training ages quickly when every instructor builds their own material. With a common base, the content changes and the language stays. Newcomers learn from pieces that look like the ones they will find in their real work.

Learning research points in the same direction. When content combines words and pictures, instead of words alone, people apply what they learned better in new situations.

What the teams said

The thank-you notes came from different areas over two years, which says more about the system than any single compliment.

The templates you and your team built make Consulting's life so much easier!
Consulting team working alongside several regional units, May 2023 (translated from Portuguese)

The slide library material has made my communication easier, faster and much more effective.
Business professional, January 2022 (translated from Portuguese)

Thank you so much for releasing the templates so we could build the presentation for the Convention.
Organizers of the Sales Convention held the following January, December 2021 (translated from Portuguese)

The first and third testimonials are revealing. The library did not serve only the Large Enterprise sales team: it supported the presentation of a corporate event and spread to parallel teams that used to ask us for help, freeing up time for both the design team and the teams that needed support. When a resource is good enough to be used far from the person who created it, it stops being a deliverable and becomes infrastructure.

The same principle in other contexts

The logic was later applied outside TOTVS, in companies of different sizes and stages. At Fidela, visual resources helped explain a Brand Loyalty mechanism that runs through leads, opportunities, customers, retention, upsell and referral. At Cogny, the institutional presentation in English brought international presence, clients and results together in a single narrative for sales conversations outside Brazil.

Enlarge

Enlarge

Different contexts, with the same underlying decision: treat the presentation as a tool someone else will use, not as a piece someone will admire.

In the end, none of these decks was made to be beautiful. A slide has one job: to guide understanding. It helps both sides of the room at once: the presenter, to organize and deliver the message; the audience, to receive it without unnecessary effort. That is why it works better without excess, supporting the speech instead of competing with it. The presenter is the protagonist, not the slide. A beautiful slide means nothing if the audience did not understand what it needed to understand.

And now, with AI?

This picture has changed a lot in recent years. The slide library no longer has to wait for someone to open the right file and look for the template: it can be consulted and applied by an AI assistant that knows the brand's rules.

Today this is within reach of any team. Claude already comes with a catalog of ready-made skills, created by Anthropic and partners for areas such as legal, operations, HR and customer support, and it lets each company create its own.

At UpFlux, the company already had a reference deck. The design team created a Claude skill with the presentation standard. A skill is a package of instructions and templates that the assistant loads when it receives a task of that kind. In practice, anyone could ask for a presentation and it came out with structure, hierarchy and identity already in place.

The most interesting effect was not speed. It was the shift in the design team's priorities. Once the standard no longer depended on human hands, designers stopped spending hours fixing alignment and turned to what AI does not deliver on its own: the piece that makes the room stop, the image that sums up the argument, the moment of surprise that an important proposal deserves.

A Harvard experiment with Boston Consulting Group helps explain this balance. On tasks within AI's reach, consultants who used it delivered more, faster and with higher quality. On a task outside that reach, they made more mistakes than those who worked without it. The machine speeds up what it already knows how to do. Noticing where it goes wrong is still human work.

But a gain on the task does not automatically become a gain on the calendar. A study that followed workers across Denmark found no measurable change in hours worked or earnings two years after ChatGPT arrived. What changed was how work was organized, with new tasks of reviewing and integrating AI. Time saved only shows up when someone decides where to reinvest it, and the moment of impact is a good place to put it.

So how much time is actually saved? There is still no rigorous study that measures slide production alone; the numbers in circulation usually come from software vendors. The best evidence measures neighboring tasks. In an experiment published in Science, 453 professionals produced typical work documents, such as reports and announcements: those who used ChatGPT took 40% less time and delivered material rated 18% better. In the BCG consultants study, tasks within AI's reach were completed 25% faster.

• What AI takes on: applying the standard, building the structure, fitting content into the template and keeping dozens of presentations consistent.
• What people keep: deciding what to say, choosing the evidence, creating the moment of impact and answering for the result in front of the client.

The logic of the 2021 library still holds. What changed is the format of the accelerator: before, a living deck that people navigated; now, a system that AI consults. In both cases, the groundwork is the same. Someone has to define the standard, organize it by use and keep everything up to date. A poorly written skill only spreads inconsistency faster.

That is why the role of design moves up a level instead of disappearing. Those who used to build the slide now define the rules AI follows and save their energy for the pieces where finish changes the decision.

THE STRATEGIC IDEA's suggestion for learning to build a skill: writing a good skill does not require being a developer, and Anthropic teaches the way for free at Claude Academy. Two courses are a good starting point:

How to start a visual accelerator on your team

1. Map the uses
   A diagnosis, a sales proposal, an onboarding and a retrospective are different journeys, with different needs. See the mandala below: it brings together 24 uses to get started.
2. Build a navigable index
   One of the biggest time sinks is when people cannot find what they are looking for, designers included, scrolling for hours through a deck. A library that requires manual searching ends up depending on whoever built it again.
3. Add instructions
   Let your deck travel through the company on its own; it will reach places you never imagined. Put the best way to use it inside the material itself. If your rollout requires training, the accelerator becomes a bottleneck. Your instructions also help AI become one of its users.
4. Define an owner and an update cadence
   An outdated template spreads wrong information very efficiently. The artifact has to be revisited, which is why we recommend a live link (Google Slides or PowerPoint). Its file size also stops being a problem that freezes the teams' computers, since each person downloads, or copies and pastes, only the slide they need.
5. Listen to the queue
   The requests that keep coming in show the next template that needs to exist. This gives you a consistent backlog and tells you exactly where your deck can evolve.
6. What if the team already uses AI?
   Translate the library into instructions the assistant can follow: the standard, good and bad examples, and the point where a designer needs to step in.

Enlarge

Knowing the moment is not enough: you also have to choose the form. The second mandala organizes 24 visual structures into six families, from the simplest way to compare two options to the scene that opens a story. Several of them came straight from the TOTVS library index, such as the screen mockup, the map, the one-pager and the case template.

Enlarge

Before and after: who does what

A summary of how presentation work is split between people and machines. The date matters: AI tools change within months, and this table reflects September 2026. What tends to last is the logic of the split, not the task list.

The limit of the template

No slide library replaces the decision about what to say. A misused template flattens the message and gives a weak argument the look of a finished one. That is why the system works best with judgment next to it:

• which problem opens the presentation?
• what evidence supports the argument?
• what decision is expected from the room?

Enlarge

A ready-made slide saves an hour. A slide system changes who can speak clearly inside the company.