Service
AI
PPP uses AI across all of our work, from film to code. Most organisations aren’t there yet. They have bought AI licences, but the tools can’t access their data, and people still use them to rewrite emails. We connect AI to your systems, train your people to use it in their own work, and build tools for the tasks and bottlenecks that eat up their time.
Why you need us
Your AI can’t see your business
A model that can’t reach your CRM, finance system or documents will only ever give generic answers. We connect it through one secure login, so it works with your numbers and follows the access rights you already have.
Habits change slower than tools
A new tool doesn’t change how people work. Seeing it solve their own task does. So we train on real work, not generic prompts, and stay until the new way is the normal way.
The problem too small to fund
Every business has one: a spreadsheet updated by hand, a quote flow stuck in an inbox. Too small for an IT project. That’s the size our Chalmers engineers build for, with AI writing much of the code.
AI Strategy & Adoption
We start with where the hours actually go, not with which tool to buy. That tells us what to connect, who to train and what to build first. Then we agree who owns it, so it doesn’t stall after launch.
We used ourselves as guinea pigs. Nine agencies share one agent setup. Some of it worked straight away, some didn’t. The hours it freed up went back into time with clients.
MCP & AI integrations
MCP is how AI reaches your systems: CRM, accounts, documents. We put every connection behind one portal with your company login, so people sign in once and only see what they’re allowed to.
The portal also shows who used what, and what it cost. We run the same setup for our own MCP server, Klartext MCP, which connects Fortnox bookkeeping to Claude and ChatGPT.
AI Training
We train on your team’s own tasks. A salesperson learns to prepare a meeting with it, a controller to close the month. Once people see AI solve their own work, they keep using it.
Leadership gets a workshop on where to start. Teams get hands-on sessions and skills they can reuse. Afterwards we follow up on hours saved and tasks moved, not on how many people attended.
AI Development
One bottleneck, one tool. A quote flow stuck in inboxes, a spreadsheet the business runs on. The engineer who scopes it builds it and looks after it, using AI to write most of the code.
ProfileXpress handled quotes in personal inboxes, so customers never knew where their order was. We built a customer portal with a quote and order module, synced with Fortnox. Now customers can check for themselves.
AI Agents & Automation
Some work follows the same pattern every time: find the information, check it, summarise it, write it up. That’s what agents are for. Each gets one job, every step is logged, and a person makes the call.
Luminar Ventures evaluates many startups every month, and most of the research followed that pattern. We built a team of AI analysts to handle it. The same team now gets through more opportunities without hiring, and every decision stays human.
Cases & clients we’re proud of
/https%3A%2F%2Fpeoplepeoplepeople.group%2Fwp-content%2Fuploads%2F2025%2F02%2Fcase-hover.jpg)
/https%3A%2F%2Fpeoplepeoplepeople.group%2Fwp-content%2Fuploads%2F2025%2F03%2Fcase-hover.jpg)
Let’s start something fun.
Chatty CEO
