There is no such thing as a scarcity of AI benchmarks out there at the moment, with fashionable 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 based mostly 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, in keeping with new analysis from the information 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 vital 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 precise issues to make Databricks a greater platform," Erich Elsen, principal analysis scientist at Databricks, defined to VentureBeat. "In order that's why we have been wanting 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 clients have?"
The result’s OfficeQA, a benchmark designed to check AI brokers on grounded reasoning: Answering questions based mostly on advanced proprietary datasets containing unstructured paperwork and tabular knowledge. Not like current benchmarks that target summary capabilities, OfficeQA proxies for the economically precious duties enterprises truly carry out.
Why educational benchmarks miss the enterprise mark
There are quite a few shortcomings of fashionable AI benchmarks from an enterprise perspective, in keeping with Elsen.
HLE options questions requiring PhD-level experience throughout numerous fields. ARC-AGI evaluates summary reasoning via visible manipulation of coloured grids. Each push the frontiers of AI capabilities, however don't replicate each day enterprise work. Even GDPval, which was particularly created to judge economically helpful duties, misses the goal.
"We come from a fairly heavy science or engineering background, and typically we create evals that replicate that," Elsen mentioned. " In order that they're both extraordinarily math-heavy, which is a superb, helpful job, however advancing the frontiers of human arithmetic will not be what clients try 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 job. These require parsing advanced tables with nested headers, retrieving data throughout dozens or a whole bunch of paperwork and performing calculations the place a single-digit error can cascade into organizations making incorrect enterprise choices.
Constructing a benchmark that mirrors enterprise doc complexity
To create a significant take a look at of grounded reasoning capabilities, Databricks wanted a dataset that approximates the messy actuality of proprietary enterprise doc corpora, whereas remaining freely obtainable for analysis. The staff landed on U.S. Treasury Bulletins, printed month-to-month for 5 many years 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, advanced 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 many years. Till 1996, the bulletins have been scans of bodily paperwork; afterwards, they have been digitally produced PDFs. USAFacts, a company whose mission is "to make authorities knowledge simpler to entry and perceive," partnered with Databricks to develop the benchmark, figuring out Treasury Bulletins as perfect and guaranteeing questions mirrored lifelike use instances.
The 246 questions require brokers to deal with messy, real-world doc challenges: Scanned photographs, hierarchical desk buildings, temporal knowledge spanning a number of studies 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 might reply utilizing parametric information or net search alone. This eliminated easier questions and a few surprisingly advanced ones the place fashions leveraged historic monetary information memorized throughout pre-training.
Each query has a validated floor fact reply (usually a quantity, typically dates or small lists), enabling automated analysis with out human judging. This design selection issues: It permits reinforcement studying (RL) approaches that require verifiable rewards, just like how fashions prepare on coding issues.
Present efficiency exposes elementary 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 pondering=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 fighting parsing fairly 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 share level enchancment)
-
GPT-5.1 Agent achieved a 52.8% accuracy (a +9.3 share level enchancment)
Three findings that matter for enterprise deployments
The testing recognized vital insights for practitioners:
Parsing stays the basic blocker: Advanced tables with nested headers, merged cells and weird formatting ceaselessly produce misaligned values. Even when given precise 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 persistently fail. For enterprises the place knowledge visualizations talk vital 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 in a position to take a look at the precise reply, it's straightforward to inform if the error is coming from parsing," Elsen defined.
This automated analysis permits fast iteration on parsing pipelines. The verified floor fact solutions additionally allow RL coaching just like coding benchmarks, since there's no human judgment required.
Elsen mentioned the benchmark supplies "a extremely robust suggestions sign" for builders engaged on search options. Nevertheless, he cautioned towards treating it as coaching knowledge.
"At the very least in my creativeness, the aim of releasing that is extra as an eval and never as a supply of uncooked coaching knowledge," he mentioned. "If you happen to tune too particularly into this setting, then it's not clear how generalizable your agent outcomes could be."
What this implies for enterprise AI deployments
For enterprises at the moment deploying or planning document-heavy AI agent programs, OfficeQA supplies a sobering actuality test. Even the most recent 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 quick 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 beneath vendor advertising claims. Check in your precise paperwork earlier than manufacturing deployment.
Plan for the parsing bottleneck: The take a look at outcomes point out that parsing stays a elementary blocker. Funds time and assets for customized parsing options fairly than assuming off-the-shelf OCR will suffice.
Plan for laborious query failure modes: Even with optimum parsing, brokers plateau at 40% on advanced multi-step questions. For mission-critical doc workflows that require multi-document evaluation, statistical calculations or visible reasoning, present agent capabilities is probably not prepared with out important 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.
