Nooli Living sells interior products such as glass walls, sliding doors and acoustic panels. After participating in Claro's AI Masterclass*, co-founders Marcus and Ulrika Kjellman began using personal AI assistants to streamline everyday work across the business. Within just a few weeks, AI had become a practical part of several recurring workflows.
The challenge
At Nooli Living, product expertise and close customer relationships are central to the business. As the company grows, many important tasks still require significant time and attention. The goal was not to replace expertise, but to reduce repetitive manual work and create more efficient internal processes, allowing the team to focus on higher-value activities.
What happened
Rather than relying on generic AI prompts, Marcus and Ulrika developed personal AI assistants adapted to Nooli Living's products, tone of voice and day-to-day operations. They quickly began applying AI to real business tasks: a live business dashboard that provides an overview of sales performance in real time, and improved product images, making it easier to produce consistent visual content. Instead of becoming a separate AI initiative, the assistants became integrated into existing workflows and supported work that was already part of the business.
The team continues to identify new opportunities where AI can simplify recurring tasks and improve internal workflows.
What they run now
Live sales dashboardProduct image editingPersonal AI assistants
“The dashboard gives us a much better overview of the business and helps us follow sales in real time.”
Marcus KjellmanCo-founder, Nooli Living
“The AI assistant has made it much easier to work with our product images. We're now working significantly faster, allowing us to dedicate more time to providing excellent customer service.”
*The AI Masterclass was the in-person predecessor of today's AI Accelerator.
Customer case
An agent structure for company and portfolio analysis.
Program
AI Accelerator, 4 weeks
Starting point
Daily AI use, no shared structure
Built
A shared knowledge base and a personal AI assistant per team member
Every day, AI agents at Inbox Capital follow the news flow around the portfolio companies and evaluate whether anything affects the firm's risk picture. What they catch is reported back every morning. Nobody on the team puts that overview together anymore.
Staying up to speed used to mean exactly that: working through the news flow manually and judging what was relevant.
The starting point
They were not new to AI
When the team started working with Claro, they were already using language models every day: for company analysis, presentations and legal review of contracts. Parts of the risk work and the quarterly analysis were semi-automated.
But the work happened in chat windows in the browser. Source material was pasted in and results were pasted out. Every analysis was handled manually and started from scratch, and prompts were personal: what one person had built, the others could not reuse.
This is where many organizations stop. The tools make individual tasks faster, but the way of working stays the same, and the quality of the analysis depends on who does it.
The goal
A systematic way of working with AI
When the team entered the AI Accelerator, they wanted to start working with AI systematically: to find out which tasks AI agents could take over from the team with maintained quality, and how each person could become more productive. The aim was a robust structure of agents that could relieve the team of analysis work and make it less dependent on individual people.
The work
Information first
The Accelerator runs over four weeks. In the joint working sessions, the team learns the patterns and sets up the first structures together with Claro. Between the sessions, the team puts in their own work: getting their material in place, testing what works in practice and building out their setups, with Claro available throughout for questions, troubleshooting and individual follow-ups.
For this team, most of the work was about information rather than AI. They reorganized how the firm's knowledge is stored: a clear folder structure, with shared knowledge files written in a form the agents can read and work from. Knowledge that used to live in people's heads, email threads and individual prompts became explicit and shared.
The structure they are building has two layers. At the base, the shared knowledge files: the firm's companies and risks, documented in one place. On top of that, a personal AI assistant for each team member, with workflows adapted to how that person works, all drawing on the same shared foundation.
This is why the analysis stays consistent: team members work in their own way, but their assistants use the same information, so the firm gets one picture of the portfolio. It is also why the structure keeps growing: when the team adds a company or a risk, every assistant works from it the next day.
Claro provided the method and the patterns. The team built everything themselves, on their own companies, risks and workflows, which means they can maintain the structure, extend it and adapt it as the portfolio changes.
Today
The whole portfolio, every morning
Today, the agents analyze the portfolio companies' quarterly reports, flag new risks, and evaluate every day whether the news flow changes the existing risk picture. The results come back every morning.
The first difference is time. Nobody needs to work through the news flow or a full mailbox to stay up to date on the portfolio. The agents catch what matters, weigh it against the current risk picture and report what has changed. The team spends their time reading and acting on the analysis instead of producing it.
The second difference is coverage. Analysis that used to depend on who did it, and was only feasible for parts of the portfolio, now runs at the same depth across all portfolio companies.
“In practice, we can now operate as a far larger investment organization than we actually are, without taking on more costs. We have even reduced costs compared to when we entered the project.”
CEO, Inbox Capital
The structure keeps evolving. New companies, new risks and new workflows are added as the portfolio changes, and the daily analysis no longer depends on someone having time to produce it.
Recognize the chat-window stage? That is exactly where the Accelerator starts.
