The situation
Every new client file starts with the same chore: read the reports, pull the indicators, check the news, and write it up. It is slow, repetitive, and the quality depends on who did it that day.
Documents · Finance
Upload a company's documents, get its profile back
Finance and advisory teams spent hours turning annual reports into a readable company profile. We built a workflow where they upload the documents and receive a clean, well-written profile page, assembled by four agents.
Illustration
Documents · Finance
- Upload
- Parse
- Three analyses
- Merge and clean
- Deliver
What changed
The documents are parsed by an AI document parser, chosen after we benchmarked it against a plain OCR. Three analysis paths run in parallel, and a fourth agent merges them, cleans the figures and the wording, and delivers one readable profile page.
How it works
5 steps, end to end
- 01
Upload
Up to four PDF files and an email address, from a simple web page.
- 02
Parse
An AI document parser turns the files into structured text.
- 03
Three analyses
KPIs, web research and document analysis run side by side.
- 04
Merge and clean
A fourth agent combines the three, cleans the data and writes the copy.
- 05
Deliver
One self-contained profile page, sent to the workspace or by email.
Where the AI stops
The model writes. The code decides.
The model
- Reads and summarises the documents
- Researches recent news on the company
- Writes the profile's narrative sections
The code
- Runs the three analyses in parallel
- Hands the aggregator only the three results
- Delivers one self-contained page
By the numbers
Counts read in the code we shipped. No invented results.
- 4
- agents in the workflow
- 3
- analysis paths in parallel
- 2
- parsers benchmarked before choosing
Built with
- n8n workflow
- AI document parsing (LlamaParse)
- Web research
- LLM writing
- HTML reports
Built for
The client behind it
AI consultingGranit AIAI consulting and automation: strategy, RPA, agents and custom tools for organisations.Read the client storyMore use cases
All use cases
- AI agentsA legal assistant that cites the law, or says it doesn't knowA legal assistant that only answers from the law it retrieved, and checks every article it cites.SAAFKA
- Voice AIA sales trainer that picks up the phoneTrainees phone a realistic AI property owner, then get a scored debrief with one action to work on.ConciergElite
- AI agentsA coach that reads the playbook so advisors don't have toA coaching agent that retrieves only the right sections of a large knowledge base, and keeps internal names out of its answers.SAAFKA
Contact
let's talk about your project
- Where we areMorocco, working worldwide
- Writehello@kaishift.com
- LanguagesFrench, Arabic, English
Write to usWe answer the same working day.