Deal sourcing
The AI-ready data foundation for private equity
Everything your fund knows, ready for AI.
RIXA installs and maintains a private equity knowledge layer inside your own environment. Deals, investment cases, DD and portfolio reports, structured and linked to their source, so your AI works from your firm's data.
Your documents
Portfolio reporting331
- Monthly reports
- Board packs
- Board minutes
- Budgets
Investment documents47
- IC memos
- Business plans
- Investment theses
Due diligence18
- Financial DD
- Commercial DD
Your knowledge layer
| Company | Revenue | Adj. EBITDA | Net debt | Employees | Investment case | Status |
|---|---|---|---|---|---|---|
| Brantholm Systems ABVertical software | 316.7 | 66.7 | 147.4 | 212 | Two add-ons by year 3 | Approved66 sources |
| Kvarnvik Logistik ABContract logistics | 863.9 | 89.2 | 224.2 | 807 | No client above 20% | Draft71 sources |
| Tallvik Industriservice ABIndustrial services | 441.6 | 58.3 | 103.7 | 482 | Margin 12% to 13% | Approved70 sources |
| Havstrand Marin ABMarine services | 259.3 | 38.8 | 100.0 | 171 | Revenue +10% a year | Approved69 sources |
Read by your AI tools
- Which pricing initiatives are behind plan?
- What supports the EBITDA bridge since entry?
- How does Kvarnvik's client concentration compare with past deals?
Claude, ChatGPT or Gemini, through MCP. Every figure they read carries its document and page.
Works with
- Claude
- ChatGPT
- Gemini
Connected during installation
- Microsoft 365
- Data rooms
- CRM
Every fund can buy the same models. Only yours has your data.
Models improve for everyone at once, so they are not an edge. Your deals, investment cases, DD findings and years of portfolio data are. Structured once, that record grows with every quarter and every deal.
The problem
Your fund's knowledge is in files, not in a form AI can use
AI has reached private equity fast. The data underneath it has not moved.
- 82%1
- of GPs used AI by the end of 2024, up from 47% a year earlier.
- 8%2
- of GPs rate their data maturity as high.
- 70%2
- are strained by manual, Excel-based work.
- Report_Aug26.pdfOutlookRevenue, August 2026
- Board pack Q2.pdfSharePointNet debt, Q2 FY26
- KPI_2026_v3.xlsxSharePointKPIs, 2026
- CIM_Project_Ek.pdfData roomDeal passed, reason
- IC memo.docxSharePointThesis assumptions
- Financial DD.pdfData roomDD findings
File names are illustrative.
- 1Allvue GP Outlook Survey 2025.
- 2Allvue GP Outlook Survey 2026, n=102.
How it works
A knowledge layer, installed and maintained for you
RIXA does not replace your AI tools or your document systems. It sits between them, inside your own environment.
Connect
We read your documents where they already live: Microsoft 365, data rooms and your CRM. Nothing is migrated.
Structure and approve
Deals, investment cases, DD and portfolio reports become one structured record. Your investment team approves it, and every data point links to its document and page.
Use with your AI
Claude, ChatGPT or Gemini read the record through MCP. We set up the workflows your team uses and keep it current.
Integrations
Works where your team already works
Your documents are read where they already live, and your approved data reaches the tools your team uses every day.
Documents
SharePoint
Read documents where they already live. Nothing is migrated.
Word
Read IC memos and DD reports, page by page.
Excel
Read models and KPI files. Pull approved figures back into your sheets.
PowerPoint
Read board packs. Build slides from approved, sourced figures.
Communication
Outlook
Pick up monthly reports and board packs as they arrive.
Slack
Ask about a deal or a company and get a sourced answer in the channel.
CRM
HubSpot
Keep the deal pipeline and contacts in step with the record.
Salesforce
Keep the deal pipeline and contacts in step with the record.
Connections are set up during installation, and which ones your fund gets is agreed in the pilot. Product names and logos belong to their owners.
Use cases
Questions your AI can answer from your own data
From sourcing to exit, with every answer traced to its source.
Due diligence
“How does this target's customer concentration compare with what we found in earlier deals, and how did those deals turn out?”
Holding period
“Which portfolio companies are behind on pricing initiatives in their value creation plan, and what did the board say about it?”
Exit
“What evidence in our board packs and monthly reports supports management's EBITDA bridge since acquisition?”
Example questions. The answers depend on your data.
Inside RIXA
The foundation, as your team sees it
Every company on one set of definitions
Revenue, gross profit, EBITDA, adjusted EBITDA, net debt and employees for each company, read to the same month. Every figure opens the monthly report behind it.
- Adjusted EBITDA follows the fund's own definition, the same for every company.
- A quarter shows Approved only once a person has signed off its figures.
- A missing month shows as not reported and is left out of totals, never estimated.
Every document behind the dataset
Monthly reports, board packs, IC memos and adviser reports in one list, each showing what was read from it and whether a person approved it.
- Filter by fund, company, source, type and format.
- Tag documents, for example the entry case or an annual review.
- See what is stored but not yet read.
Each company next to the case it was bought on
Each thesis assumption from the IC memo sits next to the figures the company reported, with the pages behind it, ready for your AI to read.
- Each assumption comes from the IC memo, with its plan target.
- Next to it, the latest figure on file. RIXA does not grade it: your team, or your own AI, does.
- Any source opens the quoted line, its page and who confirmed it.
Demo dataScreens from the product on its demo portfolio. Companies, people and figures are fictional.
Ownership and security
Your environment, your data
You own the environment, the database and the data, and keep them if we part ways. RIXA sets it up and keeps it running.
- Deployed inside your own environment, on your permissions.
- You control RIXA's access, and can withdraw it at any time.
- Your data is not used to train AI models.
- Data use terms and a data processing agreement are signed before anything is connected.
FAQ
Questions
- Do we need to migrate anything?
- No. Your documents stay in Microsoft 365, your data rooms and your other systems. RIXA reads from them and keeps the structured result in your environment.
- Which AI models does it work with?
- Claude, ChatGPT and Gemini, through MCP, an open standard for connecting AI to data. You choose the model and can switch later without rebuilding the data.
- Who owns the data?
- You do. Your fund controls the environment, the database and everything in it, and keeps them if we part ways. RIXA keeps the software and the general data model.
- How is this different from Copilot on SharePoint, or portfolio monitoring software?
- AI on your folders rereads the files for every question, so answers vary and carry no source per figure. Portfolio monitoring covers reported KPIs, not deals, investment cases or DD. RIXA keeps one approved, sourced record of all of it.
Pilot program
Apply for the pilot
We are running pilots with a small number of Swedish PE firms. We start with a few portfolio companies and deals, and the documents you already have. You see the foundation and your AI working on your own data before you commit to anything.



