Former Apple and McKinsey employees from Kazakhstan have built an AI startup for presentations

Former Apple and McKinsey engineers launched the startup Perceptis, which allows users to create high-quality business presentations in minutes. The company has already raised $3.6 million in funding and is developing a product born from the founders’ personal experience of dealing with the disproportionate amount of time spent preparing slides. Read more in our interview.

Yersultan Sapar, City — Seattle, Chief Technology Officer, LinkedIn

How It All Started

Before starting the startup, I worked at Apple as part of the Special Projects Group. Within this team, we worked on several research and product initiatives. One of the projects we successfully launched was called Vehicle Motion Cues. This is a feature on the iPhone that displays dots along the edges of the screen to help reduce motion sickness while traveling. Through this work, we managed to bring the product to launch. It was one of the most significant and interesting projects in my career.

After completing this project, at the invitation of a close friend and co-founder, we began working on the startup. The initial idea for the project was named Whiteboard.ai. It was a free-form canvas where users could draw, use voice commands, and interact with the system, which in turn interpreted the semantics and connections between elements placed on the canvas. Essentially, it was an intelligent whiteboard capable of enhancing and developing the user’s visual ideas.

Initially, the idea was quite ambitious and existed at the concept stage without clear product detailing. As we worked, we gradually clarified which audience and specific tasks the product could be most useful for.

Launching the Startup

During my university years, I worked as a teaching assistant, helping professors with teaching and student interactions. When students asked questions directly in the classroom, prepared slides often did not match the specific situation, requiring us to explain the material on the fly or use a whiteboard for visualization. While I was adding explanations or writing code, part of the audience would lose focus. Additionally, lecture duration was limited to about an hour.

As a result, I developed a need for a more dynamic and adaptive way of explaining material—a more intelligent tool than traditional presentations.

My co-founder, Alibek Dostiyarov, approached this same area from a different angle. He worked at McKinsey for nine years, and a significant portion of his work, about 40%, involved creating slides that had to clearly convey the key message and served as the final result of their work. His goal was to accelerate and simplify the process of preparing such materials.

Thus, we approached the problem from different professional perspectives but arrived at the same conclusion: the need for a smarter tool for working with slides. With this idea, we began developing the product, trying to define its specific application and target audience. In April-May 2024, we received funding for the Whiteboard.ai project. However, after initial conversations with investors and receiving feedback, it became clear that we were focusing too long on developing the MVP and not interacting actively enough with customers.

As a result, we significantly revised the product concept: we abandoned the free-form canvas and moved to a more structured B2B solution aimed at consultants who need to prepare presentations and commercial proposals.

The reason for this change was that the initial audience for the free-form canvas remained insufficiently clear. Therefore, we focused on a more familiar and understandable area for us—the consulting industry.

One of the key problems for small and medium-sized consulting businesses is the high labor intensity of slide preparation, while the quantity and quality of commercial proposals directly impact company revenue. This was a specific business problem for which our solution proved relevant.

Further work was directed toward achieving product-market fit and transitioning to a more practical and understandable product.

Target Audience

During the accelerator program, the pace of work significantly accelerated. We began actively expanding our network, interacting more frequently with potential customers, strengthening our focus on sales, while simultaneously continuing product development.

In the early stages, the main clients were heads of consulting companies and heads of sales. For them, this was a labor-intensive task with a high demand for process acceleration without compromising quality. Therefore, the ideal client is either an individual specialist or a small team with a regular need for such materials and a fast decision-making cycle.

The main focus was on small companies, as they adopt new tools faster. They also often lack developed IT departments or proprietary AI solutions, which increases the value of external products.

Currently, the product has changed significantly compared to the initial version, but its key application remains in creating commercial proposals and business presentations. At the same time, we have reached a state-of-the-art level in generating business-quality slides, including materials for boards of directors and top management.

We implemented a freemium model: basic functionality is available for free, while extended capabilities are available in paid tiers depending on needs. Currently, the product is used by more than 200 companies of various scales—from individual users to small and medium-sized teams.

Among the users are also MBA students, corporate teams for internal communications, startups, venture funds, investment banks, and PR agencies—all those who regularly need high-quality presentations and previously spent a lot of time on this.

At the moment, this is effectively the final iteration of the product. Development was conducted in search of the optimal combination of technologies, understanding of consulting specifics, and the ability to create high-quality presentations.

Incubation Programs

At a certain point, we moved from a simpler idea, partially already existing in the market, to the next stage of product development. It was then that we received our first funding from The House Fund—a fund based in Berkeley that supports only UC Berkeley graduates. In our case, both founders were graduates: I completed my bachelor’s degree in computer science, and Alibek completed an MBA program.

After that, we went through an accelerator program, which proved useful for the early stage. Essentially, we were first-time founders, and we were surrounded by more experienced entrepreneurs, which greatly accelerated our learning and company development. This stage can be considered the formal start of the startup: we began moving away from the initial idea, forming a new product concept, strengthening our focus on sales and product development.

We participated in only one accelerator—The House AI Accelerator in the summer of 2024. It was a valuable experience: mostly experienced founders were around us, while we were at an early stage. We gained significant knowledge not only from the fund’s team but also from the environment and other participants. The program did not involve prize placements—all went through one process that ended with a demo day. After that, we began raising a Seed round and ultimately raised $3.6 million for scaling.

This was the only accelerator in our history, and it became an important part of our journey, for which we are truly grateful.

On Difficulties

The global problem lies in the need to constantly maintain focus. Against the backdrop of the popularity of AI coding, startups grow quickly in metrics, and strong information noise forms around them: growth, ARR, milestones, investment rounds, and new investors.

This creates a startup culture as a race for growth, where it is difficult to keep attention on real customer needs and on which technological trends are truly important. We also see that AI labs release products that can be significantly cheaper than analogs: for example, ChatGPT for $20 already offers a wide range of tools and use cases.

Therefore, the key task is to find a balance between product development and stable delivery of value to the customer amidst constant information noise. From the clients' side, this is expressed more in caution: they are rarely willing to immediately sign long-term contracts and more often choose monthly subscriptions, as the market changes quickly and new AI solutions can drastically change the competitive landscape.

This both complicates and makes the work more interesting—having to develop the product under conditions of high uncertainty. Speaking about product cases, one of the key tasks was generating high-quality slides. Tools like Figma and Canva have already changed user expectations, while PowerPoint is often perceived as outdated, even as the market actively transforms under the influence of AI.

An especially difficult task was achieving a level where slides are hard to distinguish from human work, with a focus on MBB standards—McKinsey, Bain, BCG—where presentations are used for real business decisions. This required many iterations and research and became one of the most challenging cases.

One of the key pieces of advice is to remain open to change. Processes constantly occur in the industry, not always related to a specific company.

The current situation differs from the classic SaaS world: AI agents can quickly copy functionality, and having a product itself does not guarantee a sustainable advantage. The barrier to copying is lowering, and the previous protection of solutions is disappearing, so it is important to adapt to this new reality.

We are in the role of early adopters of a technology that has not yet reached maturity—this is similar to the early stages of the cloud or mobile industries. As in the book "The Innovator's Dilemma," such technologies initially appear insignificant but over time become an industry standard, and judging by the dynamics, we are still at the beginning of this cycle.

Achievements

Our main achievement is that we have a large number of satisfied customers who use the product daily to create slides. We constantly monitor metrics and understand how effective it works.

This is especially valuable to us: a person who previously spent about 10 hours a week preparing presentations can now achieve a comparable result in 10 minutes. They can upload an existing deck and receive an updated version or create a new presentation based on a request, adapting the structure or material.

It is interesting to observe how users apply the product and how technology returns time to people. Essentially, they get the opportunity to "buy" time instead of routine work in PowerPoint. This is the main result, and we are also happy to be in the state-of-the-art segment of business slide generation.

It is also important that during our work, we met people from different industries. Consulting is a broad field where specialists often act as generalists rather than narrow experts, and every project is different. This allowed us to better understand different businesses, their products, and clients, which significantly broadened our professional horizons.

Plans

Our main goal is to ensure that anyone who needs to create presentations can use our product. All team work—from infrastructure and AI models to marketing—is built around this. We want people to no longer spend hours in PowerPoint but simply upload a request and receive a ready-made, high-quality deck.

At the same time, it is important that presentations are not monotonous or "plastic," as is often the case with AI solutions. They must maintain high quality and, if necessary, match the user's style. The combination of user expertise and the level provided by the product is the result we strive for. This is our polar star.

Currently, from a strategic perspective, we are in the scaling stage. We plan to launch an API so that third parties can use the product, as well as consider integrations, MCP servers, and add-ins to expand usage beyond the platform.

The main focus now is growth and scaling, while constantly monitoring how the AI market develops and where we should move next.