Incredible raises $2M to bring ‘vibe computing’ voice-first desktop assistant to enterprises | a deep dive interview

Incredible, a Swedish company building a voice-first desktop assistant for Mac and Windows, has raised $2 million (19 million SEK) in previously undisclosed funding since March 2024, bringing its total pre-seed round to $2.7 million (26 million SEK). The company, which has signed enterprise design partners including global supply chain firm CellMark, aims to bring 'vibe computing' AI desktop assistants to enterprises, addressing the finding from a Gartner survey that only 28% of AI use cases fully succeed and meet ROI expectations.

Become a member of GB MAX to gain exclusive access to the industry and to the most influential global B2B leadership community in the business of gaming, entertainment, and tech. Join now https://go.gamesbeat.com/gb-max/ and also get a VIP ticket to GamesBeat Next Nov 2-3, SF . Incredible https://www.incredible.one , a Swedish company building a personal voice assistant for Mac and Windows, has raised $2 million as it seeks to take “vibe computing” AI desktop assistants to enterprises. The Jönköping, Sweden-based company raised 19 million SEK ~$2 million in previously undisclosed funding since March 2024, bringing its total pre-seed round to 26 million SEK ~$2.7 million . The capital, raised as the company signed its first enterprise design partners, adds existing venture firms and angel investors with early stakes in some of the most closely watched technology companies, including Anthropic, Lovable, Legora, and Einride — among them Daniel Langkilde, a two-time Anthropic investor. The raise follows a year of listening. Across more than 300 discovery calls with companies trying to put AI to work, the team kept hearing the same thing: AI agents are powerful, but getting them to actually run inside a real organization is hard. The technology had outrun the way people use it. This friction is reflected globally: a https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-says-artificial-intelligence-projects-in-infrastructure-and-operations-stall-ahead-of-meaningful-roi-returns Gartner survey https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-says-artificial-intelligence-projects-in-infrastructure-and-operations-stall-ahead-of-meaningful-roi-returns of 782 infrastructure and operations leaders, published in April 2026, found that only 28% of AI use cases fully succeed and meet ROI expectations — and among leaders who succeeded, most attributed it to integrating AI into existing workflows and systems rather than to the underlying model. Welcome to Vibe computing Incredible’s solution is Vibe Computing—which the company called “a paradigm shift” materialized in a highly functional desktop assistant for Mac and Windows. Instead of navigating dense app ecosystems, hunting through menus, or managing endless windows, users simply hold a hotkey on their keyboard and speak in plain, natural language. Research from Stanford University https://news.stanford.edu/stories/2016/08/stanford-study-speech-recognition-faster-texting found speech input to be up to three times faster than typing for text entry, drastically lowering the friction of human-to-computer communication. That’s why Incredible went with voice-first for its solution. Crucially for enterprise IT, Incredible’s Vibe Computing model operates directly at the user interface level. This means enterprise workers can instantly benefit from cutting-edge AI automation even on older, legacy enterprise systems, completely bypassing the need for complex API integrations or expensive backend data migrations “For forty years we’ve adapted to the computer. We learned its menus, its shortcuts, its rules,” said Philip Alm, founder and CEO of Incredible. “Voice flips that. You describe the outcome you want, and the computer does the work.” That thesis is already proving out in the field. Incredible has signed a select group of enterprise clients as exclusive design partners, who use the assistant daily across their real workflows ahead of a wider release.”Global supply chain company CellMark, for example, has deployed Incredible to employees across Sweden, France, Germany and the United States through its AI Taskforce. “It’s not about taking the work away from us—it’s about boosting everyone,” said Håkan Enhager, VP Global IT and Digital at CellMark. “We are seeing higher accuracy and better results across the board.” “This additional capital allows us to continue building out Incredible with our enterprise design partner program,” Alm added. “We are proving that when everyone has an assistant doing the heavy lifting, the entire company moves faster. This momentum sets the foundation as we prepare to raise our official Seed round at the end of this year and release Incredible to the public.” The capital comes from operators, existing venture capital firms, and angels with a track record of backing category-defining companies. Incredible’s investors include: ● Magnus Emilson , one of Sweden’s most active tech angels and founder of Visit Group ● Rikard Steiber , Silicon Valley tech executive and AI Angel, founder and CEO of GoodTrust, former President at HTC Vive, and ex-Global Marketing Director at Google ● Daniel Langkilde , co-founder and CEO of Kognic ● Henrik Landgren ex-McKinsey and Spotify and Oliver Hildebrandt , co-founders of Gilion ● Fabian Hielte, Neudi & C:o ● Viktor Nord , active tech investor and co-founder of Adway ● Swedish venture firms Starbright Invest and Spintop Ventures “Every major computing epoch is defined by a breakthrough in user interface. Voice paired with autonomous agents is the next frontier,” said Rikard Steiber, a former HTC Vive executive, in a statement. “Incredible is building the interface that will define the next decade of workplace productivity.” “I come from autonomous systems, where the real challenge is earning enough trust to let a machine act on its own. Enterprise agents face the same problem, and Incredible understands that the interface is what unlocks it. That’s why I backed them,” said Daniel Langkilde, CEO of Kognic, in a statement. The ambition reaches beyond Sweden. The founding team is currently in the Bay Area, meeting early enterprise customers and US investors ahead of a planned Seed round, with an eye toward a future US presence. Origins Incredible was cofounded by Philip Alm and David Bacelj. They started on their work in natural language processing in 2022, working on text transformers and then large language models. They decided to build a system where LLMs can communicate with apps like enterprise resource planning systems ERPs , invoicing systems and more. “We built that out over four years into framework architectures that we designed in house, and then last year we started going out and commercializing that technology as Incredible,” said Alm. The company raised its pre-seed round in 2024, and then extended the round more recently. Back in December, the company rolled out a product launch targeting a few thousand AI enthusiasts. Out of that, it got 150 paying customers. It found that only 5% of the companies that were onboarded were able to work with autonomous agent technology. So it studied how people actually work day to day. “And it turns out it’s very different from what we think about when we’re in the AI tech bubble,” Alm said. “So we built this concept of vibe computing, a twist on vibe coding .” It’s where you talk to your computer and it pretty much everything for you, he said. “You get to use the most natural interface of all with very low latency voice,” Alm said. “It’s a very natural experience for the user. It’s easier for actual workers to adopt, and they are in verticals that are not necessarily AI enthusiasts and maybe don’t adopt AI that easily or that fast. Those are the best ones” for Incredible. He said the company found that the platform is taking root with people in industries like manufacturing, logistics, where you might have individuals in the teams that they might not have the largest IT departments or the biggest AI competence in the teams. “They are very keen on automating, making their processes more efficient. There’s clear ROI, lots of processes, lots of data entry — these type of things. So those have been the best,” said Alm. The company plans to raise a seed round later this year. The aim is to engage users with the product and grow from there. The company has a team of nine people. The company has come a long way to get to market without a large team or a lot of funding. Making the pitch Now the company is going to market. Incredible has a usage-based model with pricing on its web site . The pricing is tied more to the value that the company perceives that you’re getting from these products. “So someone who’s just getting started, we want them to get comfortable and be able to build a habit. And then once you’re starting enterprise-grade workflows and larger tasks that are really valuable, the charge is usually very small proportional to the value that you’re actually getting,” Alm said. For the underlying AI tech, the platform is agnostic to the type of LLM being used. I asked if token costs for AI processing were an issue now, and Alm said not really. Back in 2023, this was more of a problem. But now the token processing is more efficient, he said. “Cost has been very not much of an issue. So we’re able to both have a pretty good margin on our side, and the users get a lot of usage for their money in this platform,” Alm said. “We have a pricing model that goes with how much ROI we think they’re getting.” Getting started Alm said it’s easy for non-technical people to get started with Incredible. They can get new kinds of work done in the first 10 minutes or 20 minutes, he said, because it’s easy to use. You do, however, have to trust Incredible enough to have it learn from what you do on the screen. “You’re talking to your computer and just showing it what you do. So it can see it your screen, you can talk to it about it. You can show it things on your screen and then talk to it about that, and then it does it for you,” Alm said. “The biggest value use case so far has been large data entry. Things like take this data from here and put it into this system, and it works in legacy systems, things that are installed on your computer, things that aren’t necessarily are connected to the internet.” He said it still works in those cases because it uses your computer to click around in the background while you’re doing other things. It can handle tasks for accounting like collecting receipts, analyzing them, categorizing them bookkeeping the entries and so on. “It’s the stuff that we do that is repetitive but doesn’t require much cognitive thinking,” he said. “It’s just we need to collect all this data, we need to parse it somehow, we need to put it into a system and then send it away to someone. That’s pretty useful and very high ROI typically. But then users also love using it for looking up items in World of Warcraft or prices for items. You can ask it tasks that would take you a long time to do, he said. “Hey Incredible, can you take my latest 100 connections on LinkedIn and put them in an Excel sheet for me?” It will do that and perhaps show a plan if it’s useful. Then it does it in the background. Alm showed how Incredible was working on this task as we spoke. “You throw a task at it, and then you will see what it’s doing,” Alm said. “If I say something like, ‘Hey Incredible, reach out to 500 people. Pitch me for their podcasts.’ It might have to ask me something like, ‘What should the pitch be?’ Then it would start finding podcasts.” Before it sends off the pitches to the podcasts, it will ask for your approval. So it keeps you in the loop as it acts on your behalf. Opening up to more users Since the product is a business-to-business B2B product, it will start first with invite-only access. You can sign up on the website and get access within a couple of days of requesting access. You can buy so many licenses for a fixed fee per month. With each one so far, the company has spent time making sure the users are getting something out of it. I asked him if gaming people could use this. He noted that he has been looking up some info while he’s been inside a game. He makes the requests and stays inside the game. But it’s more focused on enterprise workflows. I asked if I had 80 browser tabs open, would it know what tab I am referring to with a given request. Alm replied that it knows everything running on your computer. “It’s just running on your computer. It has full knowledge of every app that’s installed, your browser, every local app, and you can also connect it to around 3,000 app integrations directly in the app that are currently built in. So if you want to make it more efficient at a specific task, you can do so,” he said. I asked him how long it would take to sift through 20 years of emails to find the person that I have emailed the most over the last 20 years. He said, “It shouldn’t take more than a few minutes. It’s not the AI that’s the bottleneck there. It’s going to be the download speed of your email.” But he noted Gmail has built-in queries that you can already do without the AI. So AI might simply use the built-in Google tools. “So one of the main issues we had in 2023 was people wanted to do things with their data, and they had 50,000 CRM records, right? There’s 50,000 companies or leads that they’re working, and you need to be able to work with all of that data. So in 2023, we started building what we call agentic AI models, which are fine-tuned open-source models working with code on large data,” Alm said. “That’s exactly the type of tech that we’re still running today. It practically means that the AI can pull down 50,000 CRM records and then run analytics on it. It doesn’t have to run inference on all of the all of the things in there to be able to extract things, acknowledge, filtering, understanding, and analytics, and everything. It can pull a lot of insights just from the data set without actually running inference on it, like putting it into a model, which helps a lot” on costs. What’s it useful for? This answer made me think of a lot of things that I could find this useful for. Alm said he thought about how to use the tool to parse relevant info from his 10,000 or so contacts. He figured maybe 1,000 of them were very relevant to him. He said, “I can ask it to open a tab in my browser with each one, so I can just press send, and then insert a message, and prep everything for me. So all I have to do is review that tab with the person, and then press send, and it has all the details for whatever we’re doing. As a human, I would never be able to do that.” So, yeah, that’s what an AI agent means to Alm. He also likes “vibe computing,” which is like vibe coding, or having AI do your programming. Normal workers aren’t into vibe coding as much as they want the AI to do something else for them. “When you’re actually sitting down and trying to do the work, we’re systematically trying to remove each one of those tasks and making it easy as possible to understand how to interact with the stack, how to build the first habit of using AI, and then immediately trying to hook them with voice because it turns out voice really is the most human interface we have,” Alm said. “Then we put that into high ROI workflows.” What it does and doesn’t do Alm said that Incredible isn’t into generating images, video, etc. It can do that, but there’s better tools for it. But it can learn from you the 50 things you have to do on a computer when you get an order for something and have to process it. Once you teach Incredible to do that, and you test it, then Incredible can do that task for you, Alm said. “You can say, ‘Hey, fix this order.’ If it doesn’t know, it will ask you to teach it how to fix the order, and you proceed to do so,” Alm said. You can find out what it does in a two-week free trial that starts today. Individual plans start at $50 a month. When you request access and then you’re approved, it will guide you with voice through the actual onboarding flow, so you’ll get to see it, try it, and then there’s going to be a point during sign up where you’ll be able to request access. “We’re going to start rolling out access daily to those who have requested,” Alm said. “We’re going to measure that funnel. We have a team that’s very good at product-led growth, so that’s the direction we’re heading in.” The challenge for some potential customers is they don’t trust it or they don’t trust that the AI will do things accurately enough. But you can see what it does on your screen, and before it sends something off on your behalf, it asks you to review it. The other problem that can result is that the AI may not understand what you’re asking because you’re being too vague. “Then you press teach, and it records your screen and talks to you, and you talk to it. So I go through my process. ‘Hey, this is how I take the invoice. I go into this system. I put it here. When this happens, I have to go to this system first and look at this information,'” he said. “This is in the worker’s head, but it’s never written down.”