How much time do knowledge workers spend searching for information?

McKinsey Global Institute's 2012 report The social economy estimated that the average interaction worker spends nearly 20% of the working week looking for internal information or tracking down colleagues who can help with specific tasks.

How much time could AI save legal and other professionals?

In Thomson Reuters' 2024 Future of Professionals report, a survey of more than 2,200 legal, tax and accounting, and risk and compliance professionals, respondents predicted AI would free up four hours per week within the next year and 12 hours per week within five years. These are predictions, not measured savings.

Do AI legal research tools still hallucinate?

Yes. A 2024 Stanford RegLab and HAI study found that Lexis+ AI and Ask Practical Law AI produced incorrect information more than 17% of the time, and Westlaw AI-Assisted Research more than 34% of the time, despite using retrieval-augmented generation. That is why every AI answer should be checked against its cited source.

What concerns do organisations have about putting confidential data into AI tools?

In Cisco's 2024 Data Privacy Benchmark Study of 2,600 privacy and security professionals, 27% said their organisation had banned generative AI applications for the time being, 63% had limited what data can be entered, and 48% admitted entering non-public company information into generative AI tools.

AI document Q&A: key statistics and benchmarks for 2026

This page compiles published statistics relevant to AI document search and document Q&A for business teams. Every figure below links to its source, and we have kept only figures we could trace to a reputable public source. We update this page when new data is published.

Time spent searching for information

The average interaction worker spends nearly 20% of the working week looking for internal information or tracking down colleagues who can help with specific tasks (McKinsey Global Institute, The social economy, 2012).
In medium to large businesses, contract data is held in an average of 24 different systems (World Commerce & Contracting and Icertis, 2023 Contract Management Benchmark Report).

Market growth

The global intelligent document processing market was forecast to grow from USD 1.1 billion in 2022 to USD 5.2 billion in 2027, a compound annual growth rate of 37.5% (MarketsandMarkets, September 2022).

Expected time savings

In a survey of more than 2,200 legal, tax and accounting, and risk and compliance professionals, respondents predicted AI would free up four hours per week within the next year and 12 hours per week within five years (Thomson Reuters, Future of Professionals report, July 2024). These are respondents' predictions, not measured savings.

Accuracy

Retrieval-augmented generation (RAG), the architecture used by document Q&A tools, reduces hallucinations but does not eliminate them. A Stanford study found that Lexis+ AI and Ask Practical Law AI produced incorrect information more than 17% of the time, and Westlaw AI-Assisted Research more than 34% of the time, on a benchmark of legal queries (Stanford HAI, May 2024). The peer-reviewed version was published in the Journal of Empirical Legal Studies in 2025.
The same article notes that earlier research found general-purpose chatbots hallucinated between 58% and 82% of the time on legal queries.

The practical takeaway: an AI answer is only as useful as your ability to check it. Cited answers, which link each claim to the exact document and page, make that check fast.

Data privacy and security

From Cisco's 2024 Data Privacy Benchmark Study of 2,600 privacy and security professionals across 12 geographies (January 2024):

27% said their organisation had banned generative AI applications altogether for the time being.
63% had established limitations on what data can be entered into generative AI tools, and 61% had limits on which tools employees can use.
48% admitted entering non-public company information into generative AI tools.
69% cited threats to the organisation's legal and intellectual property rights as a concern, and 68% the risk of information being shared publicly or with competitors.

Pricing (October 2026)

Published list prices, not survey data. Third-party prices come from our comparison of AI document Q&A tools, last verified July 2026; Docutrix prices are from our pricing page.

Individual tools: roughly USD 10–20 per month (for example, ChatPDF Plus at $19.99/month)
Docutrix Starter: USD 30/month for 2 users; Pro: USD 100/month for 10 users
Docutrix Business: USD 20/user/month (minimum 5 users)
Enterprise plans: custom-priced

This page is maintained by the Docutrix team. If you are a researcher or journalist and would like to discuss specific statistics, contact us at support@docutrix.com.

Last updated: October 2026.