Nutrient expands AI Assistant, automating multi-step doc workflows inside any utility

Nutrient, the doc intelligence firm for contemporary companies, introduced a significant replace to its AI Assistant, including a doc modifying agent that autonomously plans, executes, and adapts throughout multi-step duties. The up to date AI Assistant has entry to Nutrient’s purpose-built doc instruments for rendering, structure-aware extraction, kind operations, annotation, and redaction to finish end-to-end doc workflows inside any utility.

Enterprise groups evaluating doc brokers now face a build-versus-embed determination. They’ll assemble their very own stack by stitching collectively LLMs, doc parsing and rendering, software orchestration, and governance controls, then keep it as workflows and compliance wants evolve. Or they will embed a pre-built, policy-governed agent designed for manufacturing doc work from day one. First launched in early 2025, Nutrient AI Assistant is deployed straight contained in the Nutrient Viewer SDK customers already depend on, the place it advantages from high-fidelity doc rendering and purpose-built doc tooling moderately than brittle, one-off integrations.

With this launch, Nutrient introduces an agentic doc modifying agent to AI Assistant. This new reasoning layer is ready to plan and execute multi-step workflows utilizing Nutrient’s purpose-built doc instruments (rendering, structure-aware extraction, kind operations, annotation, and redaction), whereas being ruled by customized expertise — reusable activity definitions that encode domain-specific logic — and three-tier approval insurance policies (autonomous, confirmation-required, or prohibited). Builders don’t should sew collectively parsing, chunking methods, prompts, and brittle software calls. They merely embed the SDK, configure guidelines and permissions, and get a doc agent that operates with production-grade accuracy and compliance inside their product.

For document-heavy organizations, this implies an agent that interprets what workers want, applies pre-configured expertise and enterprise logic, and produces outcomes tailor-made to how they really function. For the builders constructing these purposes, it means an embeddable SDK with an autonomous modifying agent they will configure with their firm’s guidelines, workflows, and compliance necessities — linked to the LLM supplier of their alternative.

“Each enterprise we discuss to is attempting to unravel the identical drawback: they want AI that works with their paperwork, not simply reads them. However additionally they want it to run inside their very own utility, below their very own guidelines,” mentioned Jonathan Rhyne, co-founder and CEO of Nutrient. “We’re giving builders and the enterprises they serve a doc agent they will embed, configure, and belief — one which operates inside their purposes with the safety and management their organizations count on.”

Additionally Learn: AiThority Interview with Glenn Jocher, Founder & CEO, Ultralytics

What’s new: the doc modifying agent

• Agentic execution contained in the viewer: A reasoning layer plans work, selects instruments, and carries multi-step doc duties by way of to completion utilizing the viewer’s high-fidelity rendering and doc construction.
• Observe, confirm, adapt: The agent executes actions in sequence, observes outcomes, verifies outcomes, and adjusts as wanted throughout extraction, kind filling, annotation, and redaction workflows.
• Staged adjustments with approval: Edits are staged for overview and utilized completely solely after approval, holding people in charge of the ultimate output.
• Coverage-governed autonomy: Organizations outline what the agent is aware of (expertise), the way it operates (context), and what it will possibly do (approval insurance policies) – from absolutely autonomous actions to laborious prohibitions.
• Not a single-shot function: A tool-using system designed to execute doc work reliably inside the appliance the place customers already function.
• Constructed for document-heavy industries

AI Assistant is designed for industries the place accuracy, compliance, and auditability are vital, together with:

• Authorized: Extract legal responsibility caps and indemnification clauses from a contract, then stage redactions of consumer names for lawyer overview earlier than everlasting removing.
• Monetary companies: Learn a receipt and populate an expense report robotically. Extract bill line gadgets or redact checking account numbers earlier than audit sharing.
• Development: Render a blueprint, extract door dimensions from the second-floor plan, pull window schedules from tables, and spotlight fire-rated materials references.
• Healthcare: Extract affected person demographics from consumption types, summarize lab report findings with citations to particular pages, or redact PHI earlier than sharing with analysis groups.
• Authorities: Determine private identifiers throughout a public data request, stage redactions for overview, and apply everlasting removing on affirmation.

Your online business logic, embedded within the agent
Doc automation solely creates worth when it behaves the best way a enterprise operates. That is very true in regulated, document-heavy work the place “virtually proper” is a threat. Nutrient AI Assistant is constructed so organizations can apply their guidelines to each workflow and get constant outcomes throughout groups, paperwork, and purposes.

For builders, this implies they will ship that consistency inside their utility with out constructing and sustaining a bespoke doc agent stack. They embed the Viewer SDK and configure governance controls so higher-risk actions comply with prescribed approval paths, whereas lower-risk steps can run robotically. The result’s workflow automation that matches how the enterprise works and nonetheless suits the group’s compliance and overview mannequin as necessities evolve.

Finally, Nutrient AI Assistant helps corporations flip doc work from a bottleneck right into a repeatable, policy-aligned course of that improves velocity, reduces threat, and will increase operational consistency.

Embedded, not bolted on
Nutrient AI Assistant runs as a service that plugs straight into Nutrient’s Internet, iOS, Android, and React Native SDKs — the identical SDKs builders already use to view and edit paperwork. The AI interface is constructed into the viewer, with no context switching, no copy-paste, and no exterior utility required. Organizations can join AI Assistant to the LLM supplier of their alternative, together with OpenAI, Azure OpenAI, AWS Bedrock, Anthropic Claude, or self-hosted fashions, holding full management over the place knowledge is processed. On the backend, AI Assistant integrates together with your present knowledge layer and will be deployed standalone or paired with Nutrient’s server-side doc administration platform (Doc Engine) for centralized storage, collaboration, and model management. Organizations preserve full management over the place knowledge lives and the way the agent operates inside their infrastructure.