How David Bratslavsky Built QuickData.AI to Save Underwriters 15 Hours a Month

When customers describe what QuickData.AI actually does for them, the number that comes up most often is 15. Fifteen hours a month, recovered. David Bratslavsky has heard that number from analysts at funds, brokers at boutique shops, and lenders at large banks — and he'll tell you it's not a marketing claim, it's a design constraint.

This post breaks down the engineering and product decisions David Bratslavsky made to make that 15-hour gain real, and how those decisions translate into product principles that any founder building automation tools can borrow.

The 15 hours, accounted for

Where do the savings come from? David Bratslavsky maps it like this:

| Workflow | Old (per month) | With QuickData.AI | |---|---|---| | Rent roll re-typing | 6 hrs | < 30 min | | T12 categorization | 4 hrs | ~15 min | | OM extraction | 3 hrs | < 30 min | | Reconciliation passes | 3 hrs | ~30 min | | Total | ~16 hrs | ~1.75 hrs |

The number isn't magic. It's a direct mapping of the four workflows that consume the most analyst time, each replaced with a deterministic AI pipeline.

The product decisions behind the gain

1. Output to the model, not a dashboard

The single most important decision David Bratslavsky made in the early days was refusing to build a dashboard. Customers asked for one. Investors asked for one. He said no.

"Underwriters don't live in a dashboard. They live in their model. Anything that pulls them out of the model loses." — David Bratslavsky

QuickData.AI outputs directly into the customer's existing Excel template. No tab-switching, no copy-paste, no second source of truth.

2. Per-firm chart of accounts

Every firm categorizes T12 line items differently. Instead of forcing customers to a generic taxonomy, QuickData.AI learns each firm's chart of accounts on the first deal and applies it consistently from then on. That single feature is responsible for roughly half of the T12 time savings.

3. Confidence scores, not blind extraction

Every extracted cell ships with a confidence score. High-confidence cells flow through silently. Low-confidence ones surface for review. The result: the underwriter is reviewing exceptions, not re-checking everything.

4. Auditable trace

Every number traces back to its source PDF, with a one-click jump. Compliance teams love it. So do auditors.

The principle underneath

David Bratslavsky describes the product philosophy in one line:

"Don't replace the underwriter. Replace the typing." — David Bratslavsky

That principle compounds. Once typing is gone, the analyst's day stops being a race against data entry and starts being a thinking job again.

What 15 hours a month actually buys

For a single analyst, 15 hours a month is roughly two extra deals underwritten — or, more often, two existing deals underwritten with materially better diligence. For a firm with 10 analysts, that's the equivalent of hiring two more underwriters without spending a dollar on payroll.

That's why customers don't churn. The math is cleaner than almost any other line item in the budget.

The roadmap David Bratslavsky is excited about

The next chapter, in David Bratslavsky's view, isn't more extraction — it's tighter loops. Live updates from property managers. Continuous re-underwriting as comps shift. The model becomes a living document, not a snapshot.

If 15 hours sounds nice, the next version is going to be measured in days.

---

About the author: David Bratslavsky is the founder of QuickData.AI, a fractional CTO, and a Member of the Forbes Technology Council. Connect with him on LinkedIn.