AI Assistance
Controlled AI support for document extraction, report commentary, portfolio analysis and operational review, with human approval, tenant governance and auditability built into the Prism operating model.
AI in investment operations should not mean uncontrolled automation, black-box decision-making or unchecked investor-facing output. In private capital, the cost of a wrong extraction, misleading summary or unsupported recommendation can be high. AI has to sit inside the operating controls of the platform, not outside them.
Prism uses AI assistance as a productivity layer around real workflows: interpreting documents, helping prepare first-pass commentary, answering portfolio questions, supporting structured review and reducing repetitive manual work. The purpose is to help teams move faster while keeping durable records, approvals and formal outputs under human control.
That distinction matters. Prism’s AI features are designed to assist investment, finance, operations and investor relations teams, not to replace judgement, bypass approvals or become the source of truth for financial records.
✓ Document extraction
✓ Commentary drafting
✓ Portfolio assistant
✓ Human review
✓ Tenant-level controls
✓ Usage governance
✓ Auditability
AI with control, not AI as control
AI should shorten workflows, not bypass judgement
Prism’s AI layer is designed to help users prepare, extract, summarise and review information faster while keeping formal decisions, financial records and investor-facing outputs under user control.
Assistive by design
AI can support first-pass extraction, commentary and analysis, but it should not become an unchecked authority over fund records, calculations, approvals or investor-facing output.
Grounded in platform context
AI features are most useful when they work from tenant-scoped Prism data, uploaded documents and user-provided context rather than generic assumptions detached from the operating record.
Reviewed before reliance
AI-generated outputs should be reviewed, edited and approved by users before they are relied on in reports, investment decisions, workflows or external communications.
Document extraction
Turn source documents into structured review material faster
Private capital workflows involve PDFs, fund documents, deal materials, investor files and operating records. AI assistance can reduce manual re-keying and help users prepare cleaner structured inputs for review.
Upload
Users provide the relevant document, file or text source within the appropriate tenant and workflow context.
Extract
AI assistance can identify relevant fields, summaries, figures or narrative content for structured review.
Review
Users check extracted values, correct errors, confirm assumptions and decide what should be accepted into the workflow.
Apply
Only reviewed and confirmed outputs should be applied to records, reports, analysis or downstream workflows.
Reporting assistance
Draft commentary without starting from a blank page
AI can help teams prepare first-pass commentary, narrative explanations and operating summaries from platform context, while reporting governance remains under user control.
Report commentary
Prepare draft narrative for quarterly reports, investor updates, board packs and internal operating reviews based on selected context and user prompts.
Review workflow
AI drafts should be edited, challenged and approved by the team before being included in formal reports or investor-facing materials.
Reporting connection
AI support works best when it connects into report templates, approval paths, export workflows and portal publication controls rather than sitting in a separate writing tool.
Portfolio assistant
Ask better questions of portfolio and operating data
AI assistance can help users interrogate portfolio context, identify areas for review and prepare operating summaries, while the underlying records and calculations remain governed by Prism’s data model and user review.
Portfolio summaries
Help users prepare first-pass summaries of portfolio movements, exposures, valuation changes or operating context for further review.
Review prompts
Support ad hoc questions around concentration, activity, workflow status, reporting preparation or documents that need attention.
Context checks
Help teams identify items worth reviewing, such as missing source data, unusual movements, document gaps or commentary areas that need human judgement.
Human decisioning
AI can surface information, but users remain responsible for decisions, interpretations, approvals and formal outputs.
Tenant AI governance
AI should be governed at tenant level
Enterprise AI use needs configuration, visibility and control. Prism is designed so AI behaviour can be governed within each organisation’s tenant rather than being treated as a generic platform-wide setting.
Admin control
Tenant administrators should control whether AI features are available, which workflows can use them and how AI usage fits the organisation’s internal policy.
Usage limits
AI usage can be governed through budget, throttle and usage controls so teams reduce the risk of accidental overuse, cost spikes or uncontrolled activity.
Provider configuration
AI provider settings and credentials should be controlled per tenant, with sensitive integration credentials protected rather than exposed casually to users.
AI providers and data handling
Use AI providers with clear processing boundaries
AI workflows may involve transmitting selected prompts, document content or platform context to configured AI providers. That means AI use should be transparent, governed and suitable for the organisation’s data policy.
Supported Providers
Anthropic, OpenAI and Google AI support
Prism is designed to support AI provider configuration using providers such as Anthropic, OpenAI and Google AI, subject to tenant configuration, commercial availability and the organisation’s data governance requirements.
Data Handling
Selected context only, not uncontrolled background submission
AI workflows should be triggered by explicit user action and limited to the prompt, document or workflow context required for the task. Prism’s AI approach is intended to avoid silent, general-purpose background submission of customer data.
Auditability and review evidence
AI activity should be visible, reviewable and explainable
When AI is used in sensitive workflows, administrators and reviewers need context around what happened, where the output was used and whether it was reviewed before formal reliance.
Who
Record which user initiated an AI workflow and which tenant or workflow context the request related to.
What
Capture the type of AI action, such as extraction, commentary drafting, portfolio analysis or document interpretation.
When
Preserve timing context so AI usage can be reviewed alongside reports, approvals, document actions and operational changes.
Outcome
Connect AI outputs to review, acceptance, edits, rejection or downstream workflow use where appropriate.
Connected AI workflows
AI is more useful when it connects to the operating platform
The real value of AI in Prism is not a separate chatbot. It is the ability to support specific workflows across documents, reporting, portfolio analysis, deals and governance.
Reporting and report packs
Use AI to support first-pass commentary and reporting preparation while keeping report publication under approval and review workflows.
Portfolio management
Use AI assistance to interrogate portfolio context, prepare summaries and identify areas that need deeper human review.
Approvals and governance
Keep AI-assisted outputs inside the same governance model as other sensitive workflows, with review, approval and audit evidence where required.
Prism Cloud or private deployment
AI governance should match the deployment model
Use AI assistance in managed Prism Cloud, or discuss standalone private deployment for organisations that need stricter control over hosting, provider access, network boundaries or customer-controlled environments.
Prism Cloud
Managed AI assistance inside Prism Cloud
Prism Cloud gives teams access to AI-assisted workflows within a managed platform environment, alongside reporting, documents, approvals, audit trails and tenant controls.
Private Deployment
AI options for customer-controlled environments
For organisations with stricter AI, data-control or hosting requirements, Perspect can discuss how AI-assisted workflows may operate within a local, private cloud or customer-controlled deployment model.
Why AI assistance matters in practice
Private capital teams do not need AI theatre. They need less manual re-keying, faster first-pass analysis, better document interpretation, cleaner reporting preparation and stronger operating discipline around sensitive data.
Prism’s AI layer is designed around that practical reality. It helps users move from documents and platform context to structured review material faster, while preserving the controls that matter: tenant governance, human review, approvals, auditability and deployment choice.
For investment, finance, operations and investor relations teams, the value is not replacing judgement. It is giving the team a better starting point, with the operating controls needed to use AI responsibly.
See how Prism uses AI assistance inside governed investment operations workflows
Explore document extraction, reporting support, portfolio context, approvals and governance controls in the live Prism demo environment.