Build Auto ABS Charts Without Waiting on an Analyst

Most Auto ABS chart requests start with a plain English question:

Can I see cumulative gross loss by months on book for 2024 vintages, split by issuer?

That should not require a new spreadsheet, a one-off analyst pull, or a week of back and forth over which denominator was used. The answer should be a chart you can inspect, change, and share.

That is what the AI Data Explorer tab is built for.

The Explorer now gives Auto ABS users two paths. You can build the visual yourself from the catalog, or, on Premium, ask AI to create the chart for you. Either way, the result stays editable. The dataset, metric, axis, series, chart type, and filters all remain visible on the left side of the screen.

That last part matters. A chart is only useful if you can tell how it was built.

Build it yourself

The manual builder is for users who know what they want.

Pick a dataset, pick a metric, choose the x-axis, add a series, and set filters. For example, the screenshot above uses:

• Dataset: Cohort (vintage curves)
• Metric: Cumulative Gross Loss %
• X-axis: Months On Book
• Series: Issuer
• Chart type: Line
• Filter: Vintage Year

The point is not to make users memorize every field in the underlying ABS-EE data. The point is to expose the chart controls in a way that matches how analysts talk about the market: vintage, MOB, issuer, FICO band, term, LTV, delinquency state, geography, and product type.

The current Auto ABS catalog includes cohort curves, coverage, geographic views, originations, snapshot metrics, and transition data. Across those datasets, the Explorer exposes 44 approved metrics and the dimensions that make sense for each one. Users can start from one of 33 templates or build their own chart from scratch.

This is the self-serve workflow: no SQL, no local extract, no custom notebook. You make the chart in the browser, adjust the controls, and keep moving.

Ask AI when you want a faster first pass

Premium users get the AI workflow on top of the same builder.

Instead of selecting every control manually, you can type the request:

Show cumulative gross loss by months on book for 2024 vintages, colored by issuer.

AI Data Explorer fills in the chart controls from that prompt. It chooses the dataset, metric, axis, series, chart type, and filters, then renders the chart through the same trusted metric catalog.

That means the AI is not freehanding an answer. It is filling out the chart builder.

After the chart appears, you can edit it directly. Change cumulative gross loss to 60+ DPD. Switch the series from issuer to FICO band. Add a product filter. Turn a line chart into a bar chart at MOB 24. The AI can help with those revisions too, but the controls never disappear.

For analysts, that makes the feature easier to trust. You are not left with a mysterious chart and a polished paragraph. You can see the exact query shape.

Why editable AI matters

Most AI chart tools make the same mistake: they treat the chart as the final answer.

In structured finance, the chart is usually the start of the work. You still need to check the cohort definition, confirm the denominator, change filters, compare one issuer against the group, and decide whether the visual tells the story clearly.

AI Data Explorer is designed around that reality.

If the first pass is close but not quite right, you do not have to start over. You can change the visible controls or ask AI to revise the view. The chart stays tied to the metric catalog, so the same definitions drive the visual whether it came from a manual template or a Premium AI prompt.

For Auto ABS, that distinction is important. "Loss" can mean gross loss, net loss, loss frequency, loss-to-liquidation, or a delinquency proxy depending on the analysis. A good data tool should make that choice obvious.

What Premium adds

The regular chart builder is useful on its own. Premium makes it faster and more flexible.

Premium users can:

• Ask AI to create a chart from plain English
• Ask AI to revise an existing chart
• Use the Premium Auto ABS view
• Export CSV from Explorer results
• Keep every AI-generated chart editable

The workflow is meant for the common moments when you know the question but do not want to click through every control. It is also helpful when you are exploring. Start with a rough prompt, see the chart, then tighten the filters.

For example:

Compare 60+ DPD by issuer for recent vintages.

or:

Show originations over time by FICO band.

or:

Switch this to a bar chart at MOB 24 and sort largest to smallest.

The AI handles the setup. The user keeps control.

A better way to answer one-off chart requests

Auto ABS analysis has a lot of small chart requests. Some turn into research notes. Most are quick checks:

• Is one issuer running above peers?
• Did the 2023 vintage season worse than 2022?
• Are subprime losses separating earlier in the curve?
• Which FICO bands are driving delinquency?
• Is a transition spike broad-based or issuer-specific?

Those questions should not wait for a custom extract. They should be answerable in the product.

AI Data Explorer turns the chart builder into the first stop for those questions. Use the templates when you want a familiar view. Build manually when you already know the metric. Use Premium AI when you want the system to fill in the controls for you.

The output is still a chart you can inspect. That is the whole point.

Try it in Auto ABS

Open the Auto ABS dashboard and go to AI Data Explorer. Start with a template, build your own query, or ask Premium AI for a chart in plain English.

If the first chart is not the one you want, edit the controls on the left. The workflow is built for iteration, because that is how ABS analysis actually works.