One backlog in. One white-labeled report out.
Every engagement follows the same repeatable path. Upload a client’s backlog of any size, score it unattended, curate the findings, and deliver a report under your own brand — then re-score to show the improvement.
- Step 1
Set up the client and project
Create a client workspace for the end-client, then a project for the engagement. Each client's projects, uploads, and reports stay isolated from every other client's — so a single agency can run dozens of engagements side by side without anything bleeding across.
- Step 2
Upload the backlog
Export the backlog from the client's tool and drop the file in — .csv and .xlsx, with no file-size or row limits, backlogs of 100,000+ stories included. Vindex detects the delimiter and sheet and auto-suggests a column mapping (Jira, Azure DevOps, Asana, and Monday exports are recognized). Only Summary is required; mapping Description and Acceptance Criteria sharpens scoring, and mapping Epic, Author, or Status unlocks extra report sections.
- Step 3
Score the backlog
Before the run you see the exact cost — “Score 1,240 stories · 1,240 credits · balance 5,000.” Confirm, and the job runs unattended on the same pipeline that powers Vindex's real-time integrations: close your laptop and check the progress page later. Every story is automatically de-identified before any AI model sees it.
- Step 4
Review and curate the findings
The review workspace shows every scored story worst-first, with the plain-language findings, the engine's reasoning, and its recommendations. Exclude junk rows so they don't skew the aggregates, add analyst notes in your own voice, and curate which recommendations make the cut — you stay the expert; Vindex does the first pass.
- Step 5
Build and deliver the report
Compose the report from the sections you want, apply your branding, and generate an immutable version — delivered as a PDF, a self-contained HTML file, and an Excel data export. Hand the file to the client directly; they never need a Vindex account or login.
Then run it again
After the client revises their stories, run another scoring round on the updated backlog. The report shows the improvement round over round — concrete, measurable progress you can put in front of a stakeholder, and a natural reason for the engagement to continue.
You only pay for stories that score.
Credits are reserved when a run starts and settled when it finishes: successfully scored stories are consumed, and every failed or undispatched story is refunded to your balance automatically. The exact cost is always shown before you confirm a run — so a review is something you can quote a client with a margin.
See the setup & workflow guide for the full detail on credits, mapping, and rounds.
Safe to run on real client backlogs.
- Every story is de-identified before any AI model sees it — personal identifiers become placeholders like [NAME].
- Vindex never trains models on your data or your client's data.
- Failed scores are never charged — the credit reservation is refunded automatically at settlement.
- Your clients never need a Vindex account; you deliver the report file directly.
Read more on the Security & Trust page.
Common questions
- Do my clients need a Vindex account?
- No. You deliver the report file directly — a PDF, self-contained HTML, or Excel export. Clients never touch Vindex.
- Is there a limit on backlog size?
- No. CSV and Excel uploads have no file-size or row limits; backlogs of 100,000+ stories are in scope. Cost is bounded only by your prepaid credit balance.
- What happens to stories that fail to score?
- They are never charged. Credits are reserved when a run starts and the reservation for any failed score is refunded automatically at settlement.
- How does re-scoring work?
- After the client revises their backlog, you run another scoring round on the updated file. The report shows the improvement round over round.