Skip to content
Pyrana

Solutions

Your process, built as an application. Your whole team uses it, in weeks.

Any process with data, rules and approvals can run as an application. Nobody has to learn to prompt: people work in pages built around the task, agents do the analysis behind them, and every action carries the person's authority, waits for the approval your policy names, and leaves a record before it runs. Nine examples below; the first application is live in about six weeks.

Business function owners

Nine examples

Each written as the work today, what the App does, and who approves what. Any other process follows the same shape.

AI, data, and IT owners

The platform

What is decided where, and why the foundational engineering is never redone.

Builders

The mechanisms

The checks on every action, how work runs, approval policy, your own code, provenance.

A day in the application

You work in a page built around your process. The agents do the analysis; you decide.

Reconciliation / Approval queue

Review

Approve, return, or edit what the policy routed to you.

Role: Reviewer

Journal entry · PMW intercompany adjustment

38,400 USD · proposed by the reconciliation workflow

6 of 6 policy checks passlane pmw-approvers

ApproveSend back

Accrual reversal · MWM materials

12,900 USD · proposed by the close workflow

6 of 6 policy checks passlane mwm-approvers

Approved · posted once

Journal entry · GLH freight reclass

4,150 USD · proposed by the close workflow

5 of 6 checks pass · missing supportlane glh-approvers

Sent back with a note

A queued request is visibly different from a posted entry. Nothing here has run.

The controller starts with what the policy routed to her overnight: proposals from the close workflow, each with its evidence and the checks it passed.

Illustrative. Pyrana Capital is a synthetic portfolio.

01Financial close and review

A month-end close with the evidence attached.

Ways inschedulecopilotpages
Approvalscontroller above 25,000 USD
Contextclose policychart of accountsprior-period commentary
Demo: financial analysis
The work today
Month-end runs on spreadsheets, email threads, and a controller who reads every entry because nobody else can. Variances surface late and commentary is written at midnight.
What the App does
A close application runs on a schedule: the workflow hands work to specialized agents for each entity, each reconciling intercompany balances, drafting accruals from open POs, and posting what falls inside its authority. Findings publish back to the application as versioned results. The controller reads a short list of exceptions on the review page, with the evidence attached, and approves there or through the copilot. The auditor reads one execution record.
Who approves what
Journal entries above the threshold wait for the controller. Intercompany mismatches route to the owning unit's controller. Everything below threshold posts with a receipt.
Pyrana Capital Finance / Monthly analysis / PMW · June 2026

PMW · June 2026

Monthly pack · published v2 by the copilot on behalf of J. Okafor

Published v2Add to chat

Revenue

$12.4M

-$2.5M year over year

Materials

-$188K

PPV breach, Mar and Apr

Labor

31.2%

of revenue · plan 30.0%

Findings

4

1 needs approval

Illustrative. Synthetic data.

02Collections

The aging report worked every morning, with one packet per dispute.

Ways inschedulepagesAPI from the AR system
Approvalscollections lead for write-offs and plans
Contextcredit policycontract termscustomer history
The work today
The 60-day bucket grows while the team works the accounts it can remember. Promises to pay live in inboxes. Escalations happen when someone notices.
What the App does
A collections application runs from a schedule each morning: the workflow matches remittances, drafts outreach from the customer's history and tone, schedules follow-ups, and assembles each dispute into one packet with the invoice, the contract clause, and the payment history. The team works the queue on the application's pages. Escalations route to the collections lead with the packet attached.
Who approves what
Write-offs and payment plans above policy wait for the collections lead. Customer-facing messages above a sensitivity threshold require a reviewer.
Action Items

Action Items

Identified by the Apps' runs, owned by people.

Board

Open

12

identified this week

In progress

3

Resolved

9

75% completion

Open

revenueHIGH

Rebuild APL account revenue to narrow the shortfall

Accept

In progress

collectionsMED

Work the 60-day bucket for Greenfield Agriculture

Resolved

closeDONE

Post intercompany adjustment PMW/MWM

Illustrative. Synthetic data.

03Procurement

Purchase requests with the policy applied as they are written.

Ways inpagescopilotAPI from the ERP
Approvalsprocurement lead above 50,000 EURvendor onboarding for new vendors
Contextprocurement policyvendor recordscontract terms
The work today
Buyers copy policy from a PDF into a request, approvers ask for context that was never attached, and preferred vendors are discovered after the PO is raised.
What the App does
A procurement application works on the buyer's page: as the request is written, an agent fills in the contracted vendor and price, checks the budget line and the DPA on file, attaches the quotes, and flags a request that exceeds the buyer's authority before it is submitted. The director approves as the policy names. Requests raised in the ERP arrive through the API and pass the same checks.
Who approves what
Spend above a threshold waits for the procurement lead. New vendors require the vendor-onboarding approvers. Cross-border purchases require a signed DPA on file.
Procurement / Approval queue

Approvals

Requests the policy routed to you, with the quotes and the budget line attached.

Role: Reviewer

Purchase request · Nordholm Logistics, Q4 freight

68,000 EUR · contracted rate applied · budget line 4410

Policy checks 5 of 5lane procurement-lead

ApproveSend back

New vendor · Aster Packaging

DPA on file · tax form missing

5 of 6 checks pass · tax form missinglane vendor-onboarding

Sent back with a note

Illustrative. Synthetic data.

04Marketing content

Content that carries its own review record.

Ways inpagescopilot
Approvalsreviewer approvers on publishcompliance approvers for regulated claims
Contextbrand guidelinesapproved claims libraryprior baselines
Demo: document redaction
The work today
Drafts move between tools, reviewers comment on stale versions, and the approved baseline is whichever file was emailed last.
What the App does
A content application holds drafts, review entries, and published baselines as versioned results. Writers draft on the application's pages with an agent, submit a review entry that persists, and a separately authorized reviewer approves. Publishing against a stale version is refused, never overwritten. Every earlier baseline stays browsable and citable.
Who approves what
Maker-checker on every publish. Claims that touch regulated language route to the compliance approvers.
Copy Engine / Review

Review

Drafts against the brand rules. Approve, return with a note, or edit.

Role: Reviewer

Amazon listing · scented serum, refresh

Draft v3 · brand rules applied · Writer: A. Reyes

Brand rules 8 of 8lane marketing-review

ApproveSend back

Website page · dry-skin collection

Draft v2 · regulated claim detected

Brand rules 7 of 8 · claim needs compliancelane compliance-review

Sent back with a note

Illustrative. Synthetic data.

05IT project management

Project status as a record, not a report.

Ways inpagescopilotAPIMCPschedule
Approvalsrelease manager for production
Contextrelease policyproject graphQA gate results
One operation, five doors
The work today
Status is assembled by hand from three trackers. Stale projects sit in review for weeks. Nobody can say who moved what, or why.
What the App does
An IT-PM application owns the project graph in a client extension with its own durable storage. A manager moves a card on a page, a colleague asks the copilot to move it, and a nightly workflow advances stale projects. All three reach the same action, pass the same checks, and write to the same transaction. Syncs to external trackers run as governed effects.
Who approves what
Transitions to production require the release-manager approvers. The nightly workflow waits for a person when the policy says so.
Project Portfolio

Project Portfolio

What is in flight, and how healthy it is.

25 tracked projects

Total projects

25

Value captured

$4.5M

28% of planned

Budget

$5.5M

of $10M planned · $4.7M under

Backlog1

PRJ-0015

Retail media measurement

PM J. Amara

+ Add to chat

Charter1

PRJ-0012

Supplier risk scoring

PM N. Bakhtiari

+ Add to chat

Execution2

PRJ-0006

ERP integration hardening

PM M. Oyelaran

Move to Production · awaiting release manager

Value realization1

PRJ-0018

Contract lifecycle rollout

PM C. Vandermeer

$240K captured

Illustrative. Synthetic data.

06Compliance and risk

Every risk tied to the plan it threatens. Every control with its evidence.

Ways inschedulepagescopilotAPI
Approvalsnamed executive for tolerance breachesrisk owners for proposalscontrol owner for exceptions
Contextplan objectives and tolerancesrisk registerpolicy librarycontrol matrix
See the enterprise risk demo
The work today
The risk register is a spreadsheet reviewed once a quarter. Owners explain away the numbers, a projected breach reaches the executive after it lands, and evidence for controls is gathered in the two weeks before the audit.
What the App does
An enterprise risk application ties each risk to the plan commitment it threatens and keeps the outlook against tolerance current. A weekly review computes the record from the data, then writes the outlook and the draft board pack. Agents propose drivers, refresh owner testimony and sweep for risks peers typically carry that are missing from the register; a reviewer agent checks every citation, and every proposal waits in the review queue for a person. Early signals sit outside the register until they mature or expire with a reason. When an outlook is projected to breach its floor, the exception goes to the executive with the options, the recommended one and the cost of waiting. The same platform runs control testing and policy scanning, with findings that cite the control, the policy unit and the source line.
Who approves what
Nothing on the register changes on its own: proposals wait for a person in the review queue. A projected breach of a tolerance goes to the named executive as an exception. An owner's explanation is checked against the data and recorded whether it holds or not. Control exceptions need the control owner; policy changes to regulated units need compliance sign-off.
Enterprise Risk / Overview

Risk overview

Meridian Specialty Foods · FY2027 plan · record current as of 18 Aug 2026

Review queue · 5

The gross margin floor has an open exception: a projected breach needs a decision. Three of four objectives on track.

Decide the margin exception

Exposure in view

$17M to $48M

Upside developing

$9M to $18M

Waiting on you

4 decisions

Exceptions this quarter

1

Objectives

Every plan commitment, its outlook against tolerance, the largest exposure on it, and the response in motion

  • Revenue $205MInside tolerance$199.2M vs $196M floorCustomer concentration · $8M to $19M · Diversification program · two new accounts by Q1 2027
  • Gross margin at or above 31%Exception · needs decision30.4% vs 30.0% floorDairy input cost break · $1.2M to $2.6M · Exception open · decision waiting on the CEO
  • New co-packing line by Q2 2027Inside tolerance+2 wks vs +6 wks limitKey-person dependency · 4 to 9 wks · Second commissioning lead in training

Illustrative. Synthetic data.

07Customer service and case resolution

Resolve the case with the policy and the history in view.

Ways inpagesAPI from the CRMcopilotschedule
Approvalsservice lead above thresholdproduct team for policy exceptions
Contextservice policyentitlementscase history
The work today
Cases bounce between queues. Agents on the phone search three systems for the order, the contract, and the last promise made. Escalations depend on who is on shift.
What the App does
A service application holds the case: the workflow assembles the customer's orders, entitlements, prior cases, and the applicable policy units, drafts the resolution, and issues credits inside its authority. The person on the case sees one page with the evidence attached and decides. Refunds and exceptions above policy queue for the approvers the policy names.
Who approves what
Credits above threshold wait for the service lead. Policy exceptions route to the owning product team. Customer-facing messages on regulated products require a reviewer.
Action Items

Action Items

Identified by the Apps' runs, owned by people.

Board

Open

12

identified this week

In progress

3

Resolved

9

75% completion

Open

revenueHIGH

Rebuild APL account revenue to narrow the shortfall

Accept

In progress

collectionsMED

Work the 60-day bucket for Greenfield Agriculture

Resolved

closeDONE

Post intercompany adjustment PMW/MWM

Illustrative. Synthetic data.

08Supply chain and inventory

See the shortage before the line does.

Ways inschedulepagesAPI from the ERPMCP
Approvalssupply lead for expeditessourcing approvers for supplier changes
Contextlead-time policysupplier termsdemand history
The work today
Planners reconcile forecasts, purchase orders, and supplier confirmations in spreadsheets. A late shipment is discovered when a work order cannot start.
What the App does
A supply application runs from a schedule: the workflow fans out per plant and per supplier, compares demand against confirmed supply, flags shortages and excess with the lead-time policy applied, and proposes expedites, substitutions, and reorder changes. The planner reviews the proposals on one page; changes to purchase orders post only inside the planner's authority or through the approval approvers.
Who approves what
Expedites above cost threshold wait for the supply lead. Supplier changes require the sourcing approvers. Safety-stock policy changes require compliance sign-off where regulated.
Dewsy Performance Intelligence / Overview
OverviewAmazonDTCPaid mediaSearchAction planData through Jun 30, 2026

Sales vs plan

19%

trailing

Blended return

0.77x

breakeven near 2.2x

Marketing efficiency

0.93

sales per ad dollar

Monthly net

-$334k

cash out vs in

AmazonConfirmed

The scented serum's low Amazon conversion is a traffic problem, not the listing

Paid TikTok traffic points at the serum's Amazon page, where the ad pixel attributes zero purchases. Sponsored clicks convert at about 10.2%, so the fix is the traffic. cxu 4e9a…

View the full diagnosis →

Illustrative. Synthetic data.

09HR and people operations

Onboarding that follows the policy, not the checklist someone remembers.

Ways inpagesAPI from the HRISschedulecopilot
Approvalsidentity approvers for access beyond profileHR and manager for compensation
Contextrole profilesonboarding policyregional requirements
The work today
A new hire's first week depends on tickets across HR, IT, facilities, and finance. Access is over-provisioned because nobody knows the exact profile for the role.
What the App does
A people application runs each hire as a governed process: the workflow provisions accounts to the access profile the role defines, schedules training that the policy requires, opens the equipment and badge requests, and reports what is done and what is waiting. Managers and HR see one page. Anything that grants access beyond the role's profile, or touches compensation, waits for the approvers the policy names.
Who approves what
Access beyond the role profile waits for the identity approvers. Compensation and contract changes require HR and the hiring manager. Terminations run the same process in reverse with the same record.
People Ops / Onboarding / Approvals

Approvals

Access and compensation changes wait for the lane the policy names.

Role: Reviewer

Access beyond role profile · S. Adeyemi, Finance Analyst

Requested: treasury workbench · not in profile finance.analyst

Role profile applied · exception requestedlane identity-lane

ApproveSend back

Offer adjustment · R. Chen, Senior Engineer

Compensation change · requires HR and hiring manager

Policy checks 4 of 4lane hr-and-manager

Approved · posted once

Illustrative. Synthetic data.

For finance leaders

The monthly review becomes a proof. Every run checked against your rules. You decide only the exceptions.

Financial close, collections and procurement run as applications. Every action carries the person's authority, waits for the approval your policy names, and leaves a record before it runs. An auditor can reconstruct any decision from the record alone.

  1. 1

    Proposed

    The close workflow drafts a 38,400 USD intercompany entry, with its evidence.

  2. 2

    Held

    It is above 25,000 USD, so it waits, unchanged, for the treasury approver your policy names.

  3. 3

    Approved

    K. Osei approves the exact entry. Nobody can edit it between the request and the decision.

  4. 4

    Posted once

    It posts once. A retry after a failure reuses the same receipt, so it can never post twice.

Illustrative

Go deeperApproval is a policy, not a boolean8 min

Same platform, any application

Any App is one full-stack application.

Every example above is the same configuration with different pages, agents and access. Change the configuration and the application changes; everything behind the pages comes from the platform.

app.yaml
one configuration

Any App is one full-stack application

Pages
which pages exist, what each may read, the copilot beside them
Operations
what it offers, declared once, reachable from every door
Agents
the specialists and workflows it uses
Data
its tool sources and named data calls
Knowledge
the cortIQ context it starts with and searches
Access
members, roles, and which entities each person may see
Results
the versioned results it keeps
Extension
its own service for plans and decisions, if it needs one

One App, the full stack: everything behind the pages comes from the platform.

Same platform, any application

Month-end close

Pages
month overview, reconciliation, approvals
Agents
an analyst per entity, a reconciliation agent
Access
each controller sees their own entity

Risk register

Pages
register, review queue, board pack
Agents
status outlook, proposals, pack assembly
Access
risk owners edit their own risks

Store operations

Pages
portfolio, store board, daily brief
Agents
a store analyst, a brief writer
Access
a manager sees their own store

Beyond the examples

Any repeatable process with rules, data, and approvals.

The examples above are a sample, not a catalog. The App configuration, the gate, the approval lanes, and the context engine are the same for a process the platform has never seen. Twelve functional areas where the same shape applies.

  • Finance and accounting

    Close, consolidation, accruals, intercompany, FP&A commentary, AP and AR, treasury, expense review.

  • Sales and revenue operations

    Pipeline hygiene, quoting and pricing approvals, renewal forecasting, churn signals, territory and commission reconciliation.

  • Marketing and eCommerce

    Channel-aware content, campaign performance diagnosis, catalog enrichment, promotion analysis, brand-rule review.

  • Customer service and success

    Case assembly and resolution, entitlement checks, credits inside authority, escalation with the evidence attached.

  • Procurement and sourcing

    Requisitions with policy applied, vendor onboarding, contract compliance, spend against budget lines.

  • Supply chain and logistics

    Demand against confirmed supply, shortage and excess detection, expedites and substitutions, supplier performance.

  • Manufacturing and quality

    Deviation and CAPA workflows, batch record review, SOP-bound checks where the policy unit and its source line matter.

  • Human resources and people operations

    Onboarding and offboarding to role profiles, policy-bound access, compensation changes with approval approvers.

  • Legal and contracts

    Clause review against playbooks, obligation tracking, matter intake, maker-checker on outbound documents.

  • Risk, compliance, and internal audit

    Control testing with captured evidence, policy change tracking, regulatory monitoring, audit-ready records.

  • IT operations and service management

    Project portfolio governance, incident triage, change approvals, SaaS control audits, configuration drift.

  • Data, analytics, and research

    Data quality monitoring, cross-source reconciliation, market and scientific research with cited findings, report generation.

Examples by industry

Industry-agnostic. Strongest where the rules matter most.

The rules, the data, and the approvals change by industry; the platform does not. Public, private, and PE-backed businesses, in regulated environments that need audit trails, data residency, and governance controls. Twelve industries where the platform applies.

  • Financial services and banking

    Close, reconciliation, KYC and onboarding, credit review, with records auditors and regulators can read.

  • Insurance

    Claims triage and adjudication support, underwriting workups, policy servicing under filed rules.

  • Private equity and portfolio companies

    Multi-entity close, portfolio reporting, value-creation tracking, one governance model across holdings.

  • Life sciences and pharma

    SOP-bound and validated processes, regulatory submissions support, epidemiology and evidence pipelines with provenance.

  • Healthcare providers and payers

    Prior authorization, revenue cycle, care coordination workflows under privacy and consent rules.

  • Manufacturing and industrials

    Quality and deviation management, supply planning, maintenance and asset workflows across plants.

  • Energy and utilities

    Asset and field work orders, regulatory reporting, outage and compliance evidence.

  • Retail, consumer, and eCommerce

    Channel performance, pricing and promotion decisions, catalog and content operations, launch diagnostics.

  • Hospitality and travel

    Multi-property close and procurement, revenue management, guest-service cases with approval authority that follows the org chart.

  • Technology and software

    Release governance, SaaS compliance checks, customer-success operations, usage and renewal signals.

  • Public sector and education

    Case management, grants and procurement under statute, records that must survive an audit or a records request.

How an engagement runs

Six weeks to the first App. Less to the second.

The timeline is typical for a first App and is confirmed per engagement. With clean data available, the App configuration itself typically stands up in about a week; the frontend experience, the access profiles, and the knowledge loaded into cortIQ take the rest.

  1. Week 1Define

    The process is named: the owner, the cycle time today, the systems it touches, and the approvals. SSO is connected.

  2. Weeks 2 to 3Bind and extend

    Access by role and the approval policy are written. The documents that matter are loaded into cortIQ, Pyrana's context engine. Systems are connected; your own code is added where the domain needs it.

  3. Weeks 4 to 5Pilot in production

    The application runs on real identities with real approvals and a real record. Cycle time and exceptions are measured against the baseline.

  4. Week 6Activate

    The published version is switched on. The identity model, roles, approvers and record are assets the second application starts with.

Agentic applications, not agents.

Bring the process that hurts. Leave with a written plan.

Talk to us

Book a demo

A 45-minute working session on your process. You leave with a written plan.

Read first

Demos

Recorded runs of real applications, in the pages people work in.