📋 FSI — Post-Activity Review

FSI (After Action Review) is a product that allows your organization to create institutional memory from the events you experience. Each record consists of three mandatory data points: what was done, what could have been done, and lessons learned. As these records accumulate in categories, artificial intelligence finds patterns between them, generates in-service training curricula and exams; most importantly, it audits a decision your organization is about to make based on your past lessons.

Because the product is defined by these three elements. "What was done" is a record; the real value lies in "what could have been done" and "lessons learned" — that's what forms the institutional memory. A system that only records the event is an archive, it doesn't produce a learning organization. Only the title is optional; if you leave it blank, the system will generate one for you.

Yes, every record must have at least one category. A record without a category cannot enter the training curriculum or be scanned by the decision auditor. Categories are hierarchical and you can mark multiple categories on a record — an event often concerns more than one task. If you don't have a category yet, the form will redirect you to the category addition screen with a single click.

Yes. Visibility descends: a participant registered in the "Logistics" category will also see the curriculum of sub-branches such as "Fueling". The reverse is not true — a lesson written in a sub-branch does not automatically rise to the parent category without approval. Otherwise, every specific lesson would climb to the root and the executive's curriculum would become a pile of hundreds of items.

Yes. You can add visuals, PDFs, Word documents, Excel spreadsheets, text, audio, and video; you select the files within the form while creating the record. Artificial intelligence analyzes the attachments and appends a summary to the record — the answer to the question of "under what conditions was it done" is often in the photo or equipment inventory.

There are four roles. The Owner manages everything. The Administrator sees almost everything the owner sees; they can invite and remove members, and manage categories and exams. The Contributor only fills out the FSI form, selects from existing categories, and only sees the records they entered. The Training Participant only sees the exams opened to them and their own results.

No. Their access ends, but the FSI records they entered remain within the organization. Knowledge belongs to the institution, not the individual.

No. The field in the record is the lesson for that single event. The tab is for lessons extracted independently of the event or from multiple events; organization managers and artificial intelligence write there. Selecting a category is also mandatory for lessons in the tab.

No. The lessons generated by artificial intelligence are born unapproved and cannot enter the curriculum without human approval. Otherwise, the model would teach the organization its own inferences as "rules".

When you say, "We are about to make decision B on issue A", it generates reservations and actionable recommendations by looking at your organization's own body of lessons learned. Every reservation is based on a real record and shows its basis. If your body of knowledge is empty, it doesn't say "no risk"; it honestly says it couldn't look at any records.

Yes, you can select a category — this is optional. If you don't select one, the entire institutional memory is scanned. If you do select one, the audit is limited to that branch (and its sub-branches), and it is stated that a filter has been applied. You can also add files to the decision audit.

The Owner or Administrator can create a curriculum covering as many categories as they wish — a drill may concern logistics, communication, and security simultaneously. You give each exam a name and the exam receives a unique address; participants recognize the exam by this address. At least two approved lessons are required in the relevant categories to generate a curriculum.

There are three levels: Easy (recall), Medium (applying what has been learned to a new situation), and Difficult (analysis, scenario evaluation, "what could have been done" reasoning). The participant chooses the difficulty; their manager can say, "work until you get a score of X on level 3". Questions are regenerated with each attempt, so someone retaking the same exam will encounter different questions.

From the Exam Results tab. Only the Owner and Administrator can see it. There is a pass/fail summary per person, followed by a full table with exam name, category, difficulty, number of attempts, and score columns. The difficulty column is deliberately displayed: a 90 on level 1 is not the same as a 90 on level 3.

Yes, from the Document Outputs tab. You can download FSI records, Lessons Learned, Decision Audits, and Exam Results sets as PDF, Word (.docx), Excel (.xlsx), Markdown, HTML, CSV, JSON and plain text. PDF or Word suit report attachments, Excel or CSV suit tabular work, and JSON suits transferring data to another system.

No. Each workspace is isolated; categories, records, and the body of knowledge are tied to the workspace and cannot be associated with another organization's categories. Your records are processed by artificial intelligence using our embedding model on our own servers.

On-premise deployment is foreseen in the design: the system is installed on the organization's own server on its network, personnel access it via a browser, and data does not leave the organization. A local open model is required for this deployment form, as external model services cannot be used in a closed network. Contact us for this deployment form.

There is no fixed fee for the product itself; artificial intelligence usage (analysis, curriculum generation, exam evaluation, decision auditing) and storage of attachments are deducted from your wallet based on your usage. You choose the quality level of the model to be used from the workspace settings.