Databricks' OfficeQA uncovers disconnect: AI brokers ace summary assessments however stall at 45% on enterprise docs

[ad_1]

There isn’t a scarcity of AI benchmarks out there right this moment, with well-liked choices like Humanity's Final Examination (HLE), ARC-AGI-2 and GDPval, amongst quite a few others.

AI brokers excel at fixing summary math issues and passing PhD-level exams that the majority benchmarks are primarily based on, however Databricks has a query for the enterprise: Can they really deal with the document-heavy work most enterprises want them to do?

The reply, based on new analysis from the info and AI platform firm, is sobering. Even the best-performing AI brokers obtain lower than 45% accuracy on duties that mirror actual enterprise workloads, exposing a essential hole between educational benchmarks and enterprise actuality.

"If we focus our analysis efforts on getting higher at [existing benchmarks], then we're in all probability not fixing the correct issues to make Databricks a greater platform," Erich Elsen, principal analysis scientist at Databricks, defined to VentureBeat. "In order that's why we had been trying round. How will we create a benchmark that, if we get higher at it, we're truly getting higher at fixing the issues that our prospects have?"

The result’s OfficeQA, a benchmark designed to check AI brokers on grounded reasoning: Answering questions primarily based on complicated proprietary datasets containing unstructured paperwork and tabular information. In contrast to current benchmarks that target summary capabilities, OfficeQA proxies for the economically beneficial duties enterprises truly carry out.

Why educational benchmarks miss the enterprise mark

There are quite a few shortcomings of well-liked AI benchmarks from an enterprise perspective, based on Elsen. 

HLE options questions requiring PhD-level experience throughout various fields. ARC-AGI evaluates summary reasoning via visible manipulation of coloured grids. Each push the frontiers of AI capabilities, however don't replicate every day enterprise work. Even GDPval, which was particularly created to guage economically helpful duties, misses the goal.

"We come from a reasonably heavy science or engineering background, and generally we create evals that replicate that," Elsen stated. " So that they're both extraordinarily math-heavy, which is a good, helpful activity, however advancing the frontiers of human arithmetic will not be what prospects are attempting to do with Databricks."

Whereas AI is usually used for buyer help and coding apps, Databricks' buyer base has a broader set of necessities. Elsen famous that answering questions on paperwork or corpora of paperwork is a typical enterprise activity. These require parsing complicated tables with nested headers, retrieving info throughout dozens or lots of of paperwork and performing calculations the place a single-digit error can cascade into organizations making incorrect enterprise selections.

Constructing a benchmark that mirrors enterprise doc complexity

To create a significant check of grounded reasoning capabilities, Databricks wanted a dataset that approximates the messy actuality of proprietary enterprise doc corpora, whereas remaining freely out there for analysis. The group landed on U.S. Treasury Bulletins, revealed month-to-month for 5 a long time starting in 1939 and quarterly thereafter.

The Treasury Bulletins test each field for enterprise doc complexity. Every bulletin runs 100 to 200 pages and consists of prose, complicated tables, charts and figures describing Treasury operations: The place federal cash got here from, the place it went and the way it financed authorities operations. The corpus spans roughly 89,000 pages throughout eight a long time. Till 1996, the bulletins had been scans of bodily paperwork; afterwards, they had been digitally produced PDFs. USAFacts, a corporation whose mission is "to make authorities information simpler to entry and perceive," partnered with Databricks to develop the benchmark, figuring out Treasury Bulletins as superb and guaranteeing questions mirrored practical use instances.

The 246 questions require brokers to deal with messy, real-world doc challenges: Scanned photographs, hierarchical desk buildings, temporal information spanning a number of reviews and the necessity for exterior information like inflation changes. Questions vary from easy worth lookups to multi-step evaluation requiring statistical calculations and cross-year comparisons.

To make sure the benchmark requires precise document-grounded retrieval, Databricks filtered out questions that LLMs may reply utilizing parametric information or internet search alone. This eliminated easier questions and a few surprisingly complicated ones the place fashions leveraged historic monetary data memorized throughout pre-training.

Each query has a validated floor reality reply (usually a quantity, generally dates or small lists), enabling automated analysis with out human judging. This design alternative issues: It permits reinforcement studying (RL) approaches that require verifiable rewards, much like how fashions prepare on coding issues.

Present efficiency exposes basic gaps

Databricks examined Claude Opus 4.5 Agent (utilizing Claude's SDK) and GPT-5.1 Agent (utilizing OpenAI's File Search API). The outcomes ought to give pause to any enterprise betting closely on present agent capabilities.

When supplied with uncooked PDF paperwork:

• Claude Opus 4.5 Agent (with default considering=excessive) achieved 37.4% accuracy.

• GPT-5.1 Agent (with reasoning_effort=excessive) achieved 43.5% accuracy.

Nevertheless, efficiency improved noticeably when supplied with pre-parsed variations of pages utilizing Databricks' ai_parse_document, indicating that the poor uncooked PDF efficiency stems from LLM APIs scuffling with parsing somewhat than reasoning. Even with parsed paperwork, the experiments present room for enchancment.

When supplied with paperwork parsed utilizing Databricks' ai_parse_document:

• Claude Opus 4.5 Agent achieved 67.8% accuracy (a +30.4 proportion level enchancment)

• GPT-5.1 Agent achieved a 52.8% accuracy (a +9.3 proportion level enchancment)

Three findings that matter for enterprise deployments

The testing recognized essential insights for practitioners:

Parsing stays the elemental blocker: Complicated tables with nested headers, merged cells and weird formatting continuously produce misaligned values. Even when given actual oracle pages, brokers struggled primarily attributable to parsing errors, though efficiency roughly doubled with pre-parsed paperwork.

Doc versioning creates ambiguity: Monetary and regulatory paperwork get revised and reissued, that means a number of legitimate solutions exist relying on the publication date. Brokers usually cease looking as soon as they discover a believable reply, lacking extra authoritative sources.

Visible reasoning is a niche: About 3% of questions require chart or graph interpretation, the place present brokers constantly fail. For enterprises the place information visualizations talk essential insights, this represents a significant functionality limitation.

How enterprises can use OfficeQA

The benchmark's design permits particular enchancment paths past easy scoring.

"Because you're ready to take a look at the correct reply, it's simple to inform if the error is coming from parsing," Elsen defined.

This automated analysis permits fast iteration on parsing pipelines. The verified floor reality solutions additionally allow RL coaching much like coding benchmarks, since there's no human judgment required.

Elsen stated the benchmark supplies "a very robust suggestions sign" for builders engaged on search options. Nevertheless, he cautioned in opposition to treating it as coaching information.

"No less than in my creativeness, the purpose of releasing that is extra as an eval and never as a supply of uncooked coaching information," he stated. "When you tune too particularly into this atmosphere, then it's not clear how generalizable your agent outcomes can be."

What this implies for enterprise AI deployments

For enterprises at present deploying or planning document-heavy AI agent techniques, OfficeQA supplies a sobering actuality test. Even the newest frontier fashions obtain solely 43% accuracy on unprocessed PDFs and fall in need of 70% accuracy even with optimum doc parsing. Efficiency on the toughest questions plateaus at 40%, indicating substantial room for enchancment.

Three instant implications:

Consider your doc complexity: In case your paperwork resemble the complexity profile of Treasury Bulletins (scanned photographs, nested desk buildings, cross-document references), count on accuracy nicely under vendor advertising claims. Check in your precise paperwork earlier than manufacturing deployment.

Plan for the parsing bottleneck: The check outcomes point out that parsing stays a basic blocker. Finances time and assets for customized parsing options somewhat than assuming off-the-shelf OCR will suffice.

Plan for exhausting query failure modes: Even with optimum parsing, brokers plateau at 40% on complicated multi-step questions. For mission-critical doc workflows that require multi-document evaluation, statistical calculations or visible reasoning, present agent capabilities might not be prepared with out vital human oversight.

For enterprises trying to lead in AI-powered doc intelligence, this benchmark supplies a concrete analysis framework and identifies particular functionality gaps that want fixing.

[ad_2]