What AI Doesn't Know About Soil, Water, and Climate — And Why It's Critical

Artificial intelligence has revolutionized many fields, and landscape design is no exception. It can generate incredible concepts, visualizations, and even optimize layouts. But there's something AI cannot know or feel — the unique, living conditions of your site: soil type, actual water balance, and microclimate. And it is precisely these factors, in my experience, that are critically important for a successful and long-lasting landscape project.

The Importance of Understanding Soil: The Foundation of Everything

My many years of experience in landscape design, starting in 1997 with the company 'Dom i Sad' (House and Garden), have taught me that soil is not just the ground beneath your feet, but a complex ecosystem on which plant life depends. AI can obtain data on the average soil type in a region, but it cannot 'feel' it, assess its mechanical composition, pH, fertility, and drainage properties specifically on your site.

For example, sandy soil will drain water too quickly, requiring frequent watering and additional fertilization, while clay soil, conversely, retains moisture, which can lead to root rot. Loam, in my experience, is the most versatile, but its characteristics can also vary greatly. A typical mistake I often encountered: planting moisture-loving plants in sand or drought-resistant ones in heavy clay, which inevitably leads to their demise after one or two seasons.

Site Water Balance: More Than Meets the Eye

Water is life, but its excess or deficiency can be fatal for a landscape. AI can obtain information about the average annual rainfall, but it won't see how water drains from your roof, where localized flooding occurs after heavy rain, or what the groundwater level is on the site. I remember once on a site with a high groundwater level (about 0.5 meters from the surface), a client, inspired by beautiful pictures, wanted to plant trees whose roots cannot tolerate constant waterlogging. Without a drainage system, this would have been a failed project.

AI also won't account for the site's topography, which affects water runoff, nor will it sense air humidity. All these are dynamic parameters that require observation and comprehensive analysis, not just statistical data.

Microclimate and Climate Zones: Nuances AI Doesn't See

AI works excellently with general climate zones (e.g., USDA). It knows that frost-resistant plants are needed for zone 5b, and heat-loving ones for 9a. But what about microclimate? These are unique, localized conditions created by buildings, fences, large trees, or even topographical features. For example, a south-facing corner of a house can create a 'heat trap' where the temperature will be several degrees higher than in an open field. Conversely, a tall two-meter fence can lead to stagnant cold air and the formation of a 'frost pocket.'

I have personally observed how plants, perfectly suited for the region as a whole, suffered or died due to an incorrect assessment of the site's microclimate. AI cannot see exactly where on your site the sun shines from morning till evening, or where it's always shady, where the main wind blows, and where eddies form. This requires personal presence and observation at different times of day and year.

Why Data Isn't Experience: The Difference Between AI and a Landscape Designer

AI is, in essence, a very powerful tool for data processing. It operates with facts, figures, and images. But it lacks sensory experience: it cannot feel the soil's moisture, perceive the scent of blooming plants, hear the rustle of leaves, or feel the coolness of shade. My 20 years of experience allow me not just to analyze data, but to foresee how the landscape will 'live,' how it will change in one, five, or ten years. This is intuition, based on thousands of completed projects and overcome challenges.

AI cannot understand the 'spirit of the place,' the aesthetic that arises from the harmony of all elements, not just their logical combination. It doesn't communicate with the site owner, doesn't grasp their preferences, dreams, and lifestyle, which are then embodied in a unique design.

Specific Mistakes Avoidable with a Human Approach

In my experience, the most costly mistakes in landscape design are precisely related to ignoring the actual site conditions. These can include:

• Plant demise: Choosing species unsuitable for soil acidity or moisture levels, leading to costly replacement or complete replanting.
• Drainage problems: Incorrect assessment of groundwater levels or site slope, causing flooding of lawns, paths, and foundations. I've seen projects where the water diversion system had to be completely redone after rainwater runoff started eroding pathways.
• Ineffective space utilization: Ignoring wind loads or sun exposure, resulting in gazebos being in drafts, and favorite plants 'burning' in the sun.
• Budget overruns: Attempting to implement an 'ideal picture' from AI without adapting to existing conditions often leads to extensive earthworks, soil replacement, or the creation of complex engineering systems that could have been avoided with proper planning. For example, sometimes it's simpler to change a path's configuration by 1 meter than to build an expensive retaining element.

How to Plan Your Site Considering Real Conditions with formspace.design

I created formspace.design with the idea that it would be a powerful tool for designers, not a replacement. AI in formspace.design (module 2.1 'AI Concept' and 3.1 'Design by Photo') can quickly generate stunning ideas based on uploaded photos and general data. But this is just a starting point that significantly saves time in the initial stage.

My approach is as follows: first, you upload your site plan (module 1.x 'Plan/Zones/Objects') or photos (3.1) and receive several AI concepts (2.1). Then, armed with these ideas, you proceed to the most important part — 'grounding' them in reality. Conduct a soil analysis (you can even do a simple express test by rolling a 'sausage' or a full laboratory analysis of pH and mechanical composition), observe the movement of the sun and wind on your site throughout the day and in different seasons, and assess the groundwater level. Use this data to adjust the proposed design in our planning modules, choosing plants and materials that are perfectly suited for your unique microclimate and soil. For example, if AI suggested plants requiring a lot of moisture, but you have sandy soil, I would replace them with drought-tolerant ones or provide for more frequent watering and the addition of water-retaining components to the soil. It is in such synergy of experience and technology that truly successful projects are born. You can read more about how I use AI in my work in the article “Why I Decided AI Should Assist Designers, Not Draw For Them”.

FAQ

Can AI ever completely replace a landscape designer?

In my opinion, not yet, and unlikely in the foreseeable future. AI is a powerful tool for information processing and idea generation, but it lacks sensory experience, intuition, and the ability to comprehensively analyze unique natural conditions. My experience shows that these qualities are critical for a successful landscape, as every site is a living, constantly changing system.

How accurate is the climate data AI uses?

AI uses extensive databases on climate zones (e.g., USDA) and general meteorological summaries. However, it cannot account for the unique microclimate of your specific site: shade from a neighboring building, a wind corridor, or localized flooding. This requires personal observation and experience that cannot be obtained from general data.

How can I assess the soil type on my site myself?

For a basic assessment, you can take a handful of moist soil and try to roll it into a 'sausage.' If it rolls easily and bends, it has a lot of clay. If it crumbles, it's sand. Loam holds its shape but isn't as pliable as clay. For the most accurate data, especially if you're planning serious plantings or a large garden, it's best to conduct a laboratory analysis.

What if AI suggests plants that are not suitable for my climate?

This is a perfectly normal situation. Use AI's suggestions as a starting point for inspiration. Then, armed with information about your soil and microclimate, find species with similar decorative properties but better suited. In formspace.design, you can easily replace objects and plants on the plan, adapting the design to reality.

What is the main difference between 'data' and 'experience' in the context of landscape?

Data are facts and figures: average temperature, soil pH, rainfall. Experience is the ability to interpret this data, see their interconnections, predict consequences, and make decisions based on thousands of similar situations that AI cannot yet 'live through.' It's like the difference between a vast library and the wisdom of an old gardener who knows every corner of their garden.

Should I trust AI in landscape design at all?

You should trust it, but wisely and with an understanding of its limitations. AI is an excellent assistant for speeding up routine tasks, generating creative ideas, and visualization. It saves you time and opens up new possibilities. But the final decision, based on a deep understanding of the site and its peculiarities, should always remain with a human. formspace.design was created precisely for this synergy, so you can combine the speed of AI with the precision of your experience.

How has my 20 years of experience influenced the creation of formspace.design?

My extensive experience in landscape design, starting with 'Dom i Sad' in 1997, taught me what problems arise during design and construction when real conditions are not considered. That's why I focused on creating a tool that doesn't just draw pretty pictures but also helps systematize the process, leaving room for human input to make critically important adjustments based on site knowledge. It's a tool for practitioners, not just an image generator. For example, in the article “Real Client Case Study: From Brief to AI Visualization”, I show how this works in practice.

Conclusion

AI is an incredible assistant, but it won't replace your vision, your knowledge, and your intuition when it comes to a living, constantly changing space. I, Vladimir Vyborny, am deeply convinced that the future of landscape design is a synergy of advanced technologies and a deep human understanding of nature. Use formspace.design to generate ideas, experiment, and visualize, and then 'ground' these ideas on your site, considering its unique features. Your experience and knowledge are the key to creating the garden of your dreams that will delight you for years to come.