Automation vendor Make says visual orchestration is the key to managing AI agents in the enterprise

You can hardly move these days for announcements about agents and agentic technology — and with the topic currently cresting the "peak of inflated expectations" in Gartner’s hype cycle, it feels like the mood amongst enterprises is souring.

Fabian Veit, CEO of low code and automation vendor Make, was keen to address this elephant in the room from the very outset of the company’s recent Waves 2025 event in Munich, using humor to acknowledge the fatigue and provide a launchpad for more practical discussion:

2025 was promised to be the year of the AI agents — and the reason you can all be here today is because you understood that early. So you can now lean back and enjoy the day while your AI agents run the business without you. Oh wait… maybe not?

The laughter that followed suggested a collective feeling of relief that we weren’t going to hear any overblown claims — with the resulting enthusiasm showing that people didn’t just want to hear what was new, but what was real.

Seeing is believing

Using this goodwill, Veit quickly pivots from amusement to realism:

I'm sure that you have all felt the excitement around AI agents everywhere in the last year. Yet in many companies, the reality around using AI isn't as rosy.

He goes on to note that this reality has led Gartner to place AI agents at the very top of the hype cycle, and an informal show of hands confirmed that most attendees recognize the gap between promise and practice in their own organizations.

Veit’s message is that Make wants to help its customers cross that gap by applying the right balance of deterministic and agentic automation to the right kinds of problems:

Our goal at Make is to help you unlock fully automated business operations across your entire value chain — starting with very traditional automation and expanding into new forms of automation like AI.

So far, so aspirational — language that could have come from any vendor in the space. But what really set the tone for the day was how Veit framed the company’s distinct perspective:

Automation is powerful — but visibility is what really makes it scale.

That line captures Make’s long-held philosophy — that empowerment comes from the ability to see and understand the systems powering a business, whether at the point of creation or operation. It’s a belief that has driven Make’s highly visual approach since its founding — and seems to have remained a guiding principle as Make extends that philosophy into the new modalities discussed during the day’s sessions. In Make’s worldview, visualization is no longer just a preferred modality for solution creation, but a shared conceptual foundation for multi-modal collaboration between people and AI.

That idea of visibility as the unifying thread across automation and AI was picked up by Anton Danilov, Make’s VP of Product, who describes the practical implementation of this strategy as visual orchestration:

We designed visual orchestration around three key pillars: how you build, how you accelerate, and how you scale.

Each pillar, he says, represents a different stage in the journey — but all are united by Make’s belief in visualization as a route to clarity. From an agentic perspective, this philosophy is now being extended to fuse automation and AI into a single visual language that spans all three pillars, enabling customers to use and visualize AI consistently from design through to operations.

While Danilov’s three pillars represent the kind of lifecycle structure that makes sense for a platform company, I spent the day piecing together something slightly different — not how each feature served a stage of development, but how each one turned visualization into value for Make’s customers. And from this perspective, three bigger threads of value emerged in my mind as a way of making sense of the day’s announcements — visualization as a medium for creation, collaboration, for co-ordination.

Visualization as creation

The first of those threads — visualization as a basic tool of creation — found its best expression via Sebastian Mertens, Make’s Head of Applied AI, who explains how this philosophy has shaped the company’s take on AI agents:

We’ve seen the rise of agents — it’s probably the topic of the year. But our focus is to make them visual, intuitive and transparent.

That statement neatly sums up how Make views the world. The company isn’t abandoning its visual-builder roots to chase the latest AI trend. Instead, it’s folding agents into the same design language that has long defined its appeal.

Mertens goes on to describe how the new agent-building experience lives directly within the company’s existing visual Scenario Builder, allowing users to visualise, observe and guide each step of an agent’s reasoning:

You can build your agents directly in the scenario builder. You can then see them work and gain understanding in real time — including how the agent reasons and which actions it takes.

In this way, visualization remains at the heart of the creative act within the Make platform — and by embedding agent development inside the same canvas, the company aims to ensure that AI represents an extension of the existing approach already valued by its customers.

More broadly, the integration of the agent control plane into the platform — so that the platform, rather than individual LLMs acts as the nexus of agent control and governance — is consistent with the wider shift I’m observing in the industry. But with its visual-first philosophy, Make aims to ensure that this integration is both visible and understandable across every automation and asset it hosts.

Mertens also acknowledges that building and scaling agents can be difficult when there’s little prior art to draw on. To address this, Make is enabling easy sharing and reuse of agents, along with a growing library of ready-made 'standard' agents that customers can adapt and extend. It’s a pragmatic model that enables agents to be packaged and delivered to lower the barrier to experimentation — while keeping all of that activity within a governed, visual environment.

At the most fundamental level, therefore, Make continues to treat visual models as the core of creation, with agents simply becoming another layer in its visual toolkit — a natural extension of the principles that have guided its platform from the start.

Visualization as collaboration

If visual tools have long been the core of solution creation within Make’s worldview, then the rapid shift to natural language interfaces could easily have been seen as a threat.

Instead, Make’s CTO and co-founder, Patrik Šimek, was eager to show how the company’s new platform agent, MAIA, uses visual models to create a shared conceptual space — increasing understanding between people and agents. He explains:

Every step of the way we ask ourselves: how can we make building in Make more convenient, more visual and more fun?

Šimek explains that the answer to this question was inspired by the "magical" feeling that comes from the new wave of vibe coding tools that turn natural language into code — but also that he wanted to deliver that same magic without sacrificing safety or scalability:

It feels magical — but at the same time a bit intimidating. Is the code safe, will it scale and what if something stops working?

You already know that solutions in Make are scalable, secure and reliable — but imagine co-creating them with a partner who truly understands you. MAIA is your personal buddy enabling you to build automations or AI solutions from end to end — all through natural language.

Crucially, Šimek ties all of this back to Make’s foundational belief in the power of shared visualizations — even in a world increasingly defined by conversational interaction:

And all this comes with something essential to me — full visual transparency. You always see what’s happening, so you always stay in control.

In doing so, Šimek contrasts the opaque outputs of natural-language coding tools with the inherent transparency of Make’s platform, whose opinionated architecture and shared visual language, he believes, provide a structured framework for anchoring conversational interactions in concrete concepts — enabling their effects to be visualized and evaluated at every step. (A point I’ve previously made about the potential convergence of vibe coding and low-code tools.)

For me, this was a key takeaway — that Make sees shared visualizations as a way to make natural-language conversations more concrete and understandable to both sides, increasing the power of the platform by transforming its visual language from a tool for faster individual solution creation into a shared conceptual model for safe and transparent co-creation with AI.

Visualization as co-ordination

For me, the combination of agent builder and conversational agent — together with other expected announcements such as MCP support — seemed like enough to place Make broadly at parity with the best of today’s platforms. At that point I would have walked away satisfied.

But there was another strong thread of value discussed across the event — one which was unexpected but really leveled up Make's value in my opinion by shifting the focus of visibility from the micro to the macro.

As German Bernhart, Principal Product Manager at Make, explains, the challenge as organizations scale isn’t just to build automations faster — it’s keeping sight of how they all connect so that they can maintain enterprise-scale coordination over the rapidly escalating complexity of the overall landscape.

Incredible things are being built with Make — not just scenarios, but whole landscapes with hundreds of scenarios and all the systems connected to them. And as these landscapes grow, it’s getting harder and harder to understand the big picture, as well as all the dependencies. To prevent it from becoming a black box, we’ve invested a lot of time and energy in visibility here too.

Announced last year in closed beta and now fully released, Make’s Grid capability is the company’s answer to scaling its visual logic beyond individual scenarios and AI-augmented conversations and into enterprise-wide coordination challenges.

Bernhart describes Grid as an auto-generated, interactive map of the entire automation landscape — a tool that brings together traditional workflows, AI-powered automations and AI agents into a single visual model:

If you can see it, it’s much easier to understand it and scale it further. You don’t have to worry about breaking something when making a change because you can see all the dependencies — and that also helps you collaborate and align with management or clients.

Based on a ‘city planning’ metaphor, Grid auto-generates a real-time, living map of automation across the enterprise, in which related workflows are grouped into 'districts' that mirror business domains. Each district then shows the dependencies between processes, agents and underlying systems — providing the ability to both zoom out and zoom in to see both the big picture and the specific details. And — like with Google maps, for instance — there are also various layers that can be toggled on or off — helping users to focus on particular system dependencies, business areas, data flows or security, for example.

And there definitely seemed to be a customer buzz around the idea — with Bernhart commenting on the number of ideas the company had received from enthusiastic testers during its closed beta phase. One CIO I spoke to described it as “the missing layer” — the moment when he could finally scale low-code with confidence because he could see what was really going on. Yet others described it as a governance tool which has helped them optimize aspects of their IT estate from credit consumption to portfolio planning — without having to introduce painful bottlenecks that slow down the teams who build.

It fell to Anastasi Horňáková, Head of Product for Scalability, to use Grid as a way of closing the loop on CEO Veit’s initial statement that true scale demands visibility:

Scaling starts with a foundation of trust. You need visibility, you need governance and you need control. And that’s what Grid delivers.

My take

Make seems to be thriving under the ownership of process mining vendor Celonis following its acquisition in 2020.

The impressions I took away from the Waves event in Munich were of a noisy and enthusiastic community, a company with growing enterprise confidence in its platform, and a fully intact vision that has survived the acquisition and which continues to rest on the idea of visualization as a route to speed and safety — by way of understanding.

Everything the company announced stayed true to that founding vision but extended it in new ways — whether through the visual creation of agents, human-AI collaboration based on visual models or enterprise-scale coordination through enterprise-scale visibility.

Many of the announcements were solid but expected — an agent studio, a platform agent, MCP support — and are the kinds of developments that keep Make moving with the market.

But Make Grid stood out for me as both unexpected and unique.

It’s a deceptively simple idea, but in my view that topological view transforms Make from a straightforward low-code platform for faster development into an enterprise-scale coordination platform — one capable of helping organizations resolve the long-standing tension between bottom-up innovation and top-down oversight.

And while I — as a strategy and operating model geek — immediately saw the power of this visual landscape, the company’s appreciation for the 2,000 members of the Make community who were enthusiastic enough to join the closed beta phase suggests that this kind of visibility has a much wider potential audience.

To me, it didn’t feel like anything I’d seen before — something visual and playful, yet also vital and powerful. Visualization as grand-scale coordination — a new kind of strategic lens through which an organization can better understand itself.

It was also, perhaps, the clearest expression yet of Make’s founding philosophy — that seeing is understanding, and understanding makes anything possible.