Box expands AI‑driven content management across key workflows
An enterprise content cloud is accelerating its use of metadata, automation, and AI‑powered security to modernize workflows across multiple sectors
Box Inc., a California-based cloud content management provider, is turning the complex world of contracts into a proving ground for its next phase of AI-powered growth.
Rather than chasing generic productivity tools, the company is doubling down on contract lifecycle management (CLM), using metadata extraction and automation to help legal teams surface risk, standardize language, and untangle thousands of documents at scale.
The strategy hinges on treating contracts not as static PDFs but as rich, unstructured data sets. By applying AI to extract key fields, classify agreements, and trigger workflows, Box aims to give law firms and in-house legal departments a way to modernize document-heavy processes without ripping out their existing systems of record.
“On the legal side, where we’ve seen a lot of traction is in the CLM space, because a big portion of that is the ability to extract metadata and relevant information from a contract and effectively create structure from unstructured data,” Ravi Malick, global chief information officer at Box, told TechJournal.uk in an interview during his recent visit to London.
“That lets you categorize contracts, kickstart workflows, and in many cases automate contract creation.”
Malick added that law firms are also looking for a more scalable and secure way to handle content while tapping AI to interrogate what is buried in their files.
“We’ve seen a lot of value in just being able to modernize the way that law firms manage all of their content and do it in a way that’s scalable and secure, and that lets them leverage AI to do queries and understand contracts better,” he said. “Metadata extraction at scale means you can more effectively understand what’s in potentially thousands of contracts – where your risk points are, what the standard language is, and execution and expiration dates.”
That emphasis on unstructured data is now crystallising in Box Extract, the company’s metadata extraction engine, and in new workflows that route AI-derived insights directly into legal processes.
Instead of asking lawyers to manually tag documents or search across fragmented repositories, Box wants to standardize how content is ingested, interpreted, and pushed to downstream tools.
Behind the scenes, the company is updating traditional enterprise content management (ECM) for an AI era, with metadata extraction, processing, and integration forming what Malick described as the three core tiers. Box Extract now sits at the center of this framework, supplying structured data that feeds automation across legal, finance, and operations.
Malick said many organisations are now re‑evaluating legacy ECM tools as they adopt AI at scale.
“Traditionally, that has lived in legacy ECM platforms, and a lot of customers have been looking at how they modernize those platforms – how they get better scalability and cross‑functional capabilities,” he said. “Extraction of metadata is a major component of the ECM space.”
By delivering all three tiers within its content cloud, Box aims to give enterprises a cleaner, more automation‑ready alternative to juggling multiple repositories.
Automation engines for enterprise content
Malick said his remit across internal IT, security, and external strategy gives him a direct view of how enterprises are rethinking workflows as AI matures.
At BoxWorks 2025 in San Francisco in September, the company introduced Box Extract and Box Automate as part of a suite of agentic tools for tasks such as invoice processing and contract review.
Malick expects AI agents to increasingly perform discrete tasks within those flows. He said what differentiates Box from connector‑heavy competitors is its ownership of the content layer and the context surrounding it.
“What’s necessary for these platforms to be successful is to have the context and the unstructured data available,” he said.
Instead of competing head‑on with core systems like customer relationship management (CRM) or human capital management (HCM), Box sees its sweet spot in the layers that surround them. Malick stressed that leading CRM and HCM vendors remain key partners, not targets.
“We’re not looking to compete directly with CRM or HCM platforms,” he said. “What we’re trying to do is fill the gaps and provide additional capabilities where those areas may fall short. A lot of times, CRM or HCM is part of a longer process, and that process is filled with content and context. That’s where Box is complementary, helping complete the process and enhance the value of the systems of record.”
Content intelligence across the enterprise
Box is extending automation across the enterprise by funneling cleaner, structured data into its integration ecosystem. With Box Extract interpreting previously opaque documents, the company’s more than 1,500 integrations — supported by high‑volume processing capabilities — can now operate with greater accuracy and underpin more advanced, AI‑enabled workflows.
That combination is resonating well beyond the legal world. Malick pointed to sectors such as power generation and high‑tech manufacturing, where workflows around content frequently involve metadata extraction and population. Whether an organisation is managing technical drawings, maintenance records, or safety documentation, the same pattern emerges: someone must understand what is in each file before meaningful automation can begin.
Over time, Box expects more of that understanding to come from AI agents that can interpret context, reason across multiple documents, and trigger actions without constant human supervision. That vision is driving internal change as much as it shapes the product roadmap.
Building an AI‑first platform
Box’s 2026 plans include strengthening its position as an AI‑first enterprise content platform, with a focus on redesigning internal processes to leverage emerging agentic AI capabilities. Malick said the company is evaluating where AI can have the most meaningful impact across both operations and the product roadmap, framing this as a central part of its near‑term priorities.
He said organisations often limit the value of AI when they apply it to legacy processes without questioning whether those processes still make sense.
“If you just layer AI on top of what you have existing, you will dilute the value of it,” he said. “It is very much about redesigning and, in many cases, scrapping existing processes, starting from scratch and asking whether a process should even exist.”
Teams across Box are now reviewing ongoing projects with this perspective, with some initiatives being reframed or retired. Malick said the company expects tangible progress in the coming months as both the product and internal operations continue to evolve.
As the company scales its AI roadmap, Malick said external partners will play a larger role in complex deployments, particularly in regulated industries such as legal, financial services, and health and life sciences.
“We’ve been looking to partner with some of the larger players in the markets as well as enhance our current partnerships with smaller players,” he said.
Shielding content with AI‑driven security
Security is also a core focus of Box’s enterprise push as organisations grapple with the accelerating volume of AI‑generated content. Box Shield — launched in 2019 to help organisations classify sensitive files and guard against common content risks — established the baseline.
Last month, the company launched Box Shield Pro, an AI‑powered expansion designed to secure unstructured content at a far greater scale.
Ben Weiner, product marketing manager at Box, said the pressure on security teams has intensified.
“Organizations have to balance facilitating and empowering users with keeping sensitive content in only the right hands,” he said. “The velocity of content is exploding, with generative AI and AI agents delivering automated productivity beyond anything we’ve seen before.”
Box Shield Pro introduces three significant advances:
AI Classification Agent that applies sensitivity labels automatically using context‑based analysis.
Ransomware activity detection that flags suspicious behaviour in Box Drive and enables compromised sessions to be halted and restored.
AI Threat Analysis Agent (arriving next year) that summarises alerts to help teams prioritise threats more efficiently.
“Incorporating AI into core security functions like classification and threat protection lets us expand and streamline how customers protect their most critical and sensitive content,” Weiner said.
Box Shield Pro is available as an add‑on for existing Shield customers and reflects Box’s view that securing unstructured data is foundational to any enterprise AI strategy.