Example agent design
Comprehensive Cash Flow Modeling for Financial Services
ITECS helps lenders and advisory firms assemble comprehensive cash flow models (balance sheet, P&L with EBITDA, working-capital roll-forward, and borrowing-base availability) from a completed field exam and prior reports, which already supply most of the inputs. This is a starting design. An AI Blueprint adapts it to your goals and the systems it will connect to.
The goal is not a prettier spreadsheet. It is a governed modeling workflow that turns existing exam output and borrower financials into a defensible forecast professionals review and sign, so senior judgment goes to the findings instead of the data assembly.
- Owner
- The professionals who own and sign the model
- Reads
- Field exam output · Historical financials · Working-capital detail · Debt and borrowing base · Management assumptions
- Never writes
- Client deliverables without human review.
- Typical cost
- AI Blueprint $4,500–$7,500, credited toward the build. Typical first build $18,000–$35,000, or start local at $4,500–$9,500. Agent Operations after launch $2,500/mo for one production agent. See the first engagement
The workflow today
Model Assembly Buries the Senior Judgment Clients Pay For
Advisory and lending teams often rebuild comprehensive models by hand each engagement, even when a fresh field exam and prior reports already contain most of the inputs. The assembly work crowds out the analysis.
A governed modeling agent should assemble the statements, reconcile the math, and flag assumptions so professionals start from a defensible draft and move straight to scenarios and findings.
Illustrative model readiness
How Much of the Model Already Exists
Most of a comprehensive model is derivable from exam output and prior reports; the remaining inputs are where senior judgment belongs.
Risk matrix
Priority by evidence
Field exam output
ReadyRoll-forward, dilution, and eligibility already structured
Historical financials
ReadyBalance sheet, P&L, and EBITDA history available
Working-capital roll-forward
DerivedAR, inventory, and AP assembled from exam detail
Forward assumptions
Input neededWhere professionals apply judgment and scenarios
Model build time
~70% less
Assembly shifted to review when exam output is clean
What leadership sees
- Reuses exam output and prior reports instead of rebuilding from scratch
- Keeps assumptions visible and source-traceable for review
- Separates derivable assembly from the judgment that needs a human
The workflow
The workflow today and with the agent
Planning comparison only. Scope, timing, and outcomes vary by data, workflow, approvals, adoption, and implementation conditions.
- Model build
- TodayReassembled by hand each engagement
- With the agentAssembled from exam output with traceable inputs
- Time to first draft
- TodayDays of spreadsheet work
- With the agentHours, then professional review
- Scenarios
- TodaySeparate workbooks and versions
- With the agentShared assumptions with sensitivity views
- Governance
- TodayAssumptions buried in cells
- With the agentVersioned, source-traceable, and reviewed
Outcome hypothesis to validate
The value is speed from engagement to advice: a defensible model in hours, more engagements per professional, and senior time spent on findings instead of assembly.
Not a reported client result or guarantee.
With the agent
What Cash Flow Model Builder does
Each capability is designed to produce evidence for the people who already own the credit, advisory, or finance decision.
Statement forecast assembly
Build the integrated balance sheet, P&L with EBITDA, and working-capital roll-forward from exam output and historical financials.
- Integrated three-statement structure with traceable inputs
- Working-capital roll-forward derived from exam detail
- First-draft model in the firm's format for review
Borrowing-base and availability math
Connect the forecast to eligible collateral and advance rates so availability moves with the model.
- Borrowing-base availability tied to forecast collateral
- Covenant and liquidity checks across the projection
- Reconciliation against the field exam and prior reports
Scenario and sensitivity
Turn the base model into scenarios stakeholders can interrogate without rebuilding the workbook.
- Upside, downside, and stress scenarios with shared assumptions
- Sensitivity on margin, collections, and advance rates
- Draft commentary with source-backed rationale
Scenario
Illustrative cash flow modeling scenario
Illustrative planning scenario, not a reported client result. An advisory firm needs to deliver a comprehensive model quickly after a field exam, without rebuilding the statements by hand each time.
- Starting point
- Models are reassembled in spreadsheets per engagement, even though the recent exam and prior reports already contain most of the inputs.
- Scoped outcome
- ITECS scopes a modeling agent that assembles the statements from exam output, reconciles the math, flags assumptions, and produces a first-draft model and scenarios for professional review.
Systems it connects to
What the agent reads
Discovery confirms authoritative systems, data quality, access, and governance before any production workflow is proposed.
- Field exam output
- Collateral roll-forward, dilution, eligibility, and turnover detail from the most recent exam.
- Historical financials
- Balance sheet, P&L, EBITDA, and trial-balance history for the borrower or company.
- Working-capital detail
- AR, inventory, and AP detail used to build the roll-forward and liquidity view.
- Debt and borrowing base
- Debt schedule, advance rates, reserves, and covenant terms from the loan agreement.
- Management assumptions
- Forward revenue, margin, capex, and collections assumptions provided by professionals.
Workflow
Read-Heavy, Write-Controlled Financial-Services Intelligence
The system connects approved data, explains risk, prepares recommendations, and routes sensitive actions for human approval.
01
Ingest
Read approved exam output, historical financials, and working-capital detail.
02
Assemble
Build the integrated statements and working-capital roll-forward with traceable inputs.
03
Reconcile
Tie the model back to the field exam, prior reports, and the borrowing base.
04
Model
Generate scenarios and sensitivities with shared, source-backed assumptions.
05
Approve
Route the draft model and commentary to professionals for review and sign-off.
Controls
Read broadly, recommend carefully, keep humans in control
Financial-services AI becomes trustworthy when it preserves assumptions, source data, approvals, and confidentiality boundaries.
- Read-heavy
- Reads the systems listed above through access scoped to this workflow.
- Write-controlled
- Writes nothing listed under Never writes in the spec at the top of this page. Anything it drafts waits at the approval points below.
- Approval points
- Recommendations, drafts, and exceptions go to a named reviewer before anything changes.
- Audit trail
- Each run keeps its inputs, sources, output, and the reviewer's decision.
In this design
- The agent assembles and recommends, but professionals own and sign the model.
- It does not deliver a model or commentary to clients without human review.
- All assumptions, inputs, and reconciliations remain visible and source-traceable.
- Role-based access separates engagement, credit, and client-facing views.
Security
Security for Financial Services AI Workflows
Cash flow models contain confidential borrower and company financials. ITECS scopes access so each role sees only the engagements and data it is permitted to use.
- Selected business or enterprise tier data terms, tenancy, and retention verified during scoping
- Versioned assumptions and source-traceable inputs
- No autonomous client delivery of models or commentary
- Audit history for model versions, assumptions, and approvals
Deployment fit
Where this design can run
- Local AI Agents
- Runs in one employee's own Claude, Codex, or Copilot workspace, over files that person already has access to. It moves to a hosted agent when it needs scheduling, multi-user access, audit trails, or an SLA.
- Hosted AI Agents
- Runs in a governed Linux environment ITECS operates, on frontier models through enterprise API agreements, connected to exam output and borrower financials through controlled integrations, with human review and an audit trail.
Typical first engagement
Start with an AI Blueprint
The AI Blueprint sets the actual build rung and its fixed-fee not-to-exceed. The rows below show the typical path and the catalog range for each step.
- FirstAI BlueprintAgent Discovery & Technical Specification. Credited toward the build.$4,500–$7,500
- Typical first buildSingle-Workflow Production Agent$18,000–$35,000
- Or start localLocal Agent SprintOne workflow · 1–2 data sources$4,500–$9,500
- After launchAgent Operations, one production agent$2,500/mo
Prices from the published AI pricing catalog. Every guided or delivered agent build is archived in the ITECS client vault: workspace configuration, skills, prompt libraries, data-source maps, and usage documentation.
Customize this design
Bring your goals and your systems. The AI Blueprint decides the scope and the systems, not this example, and ends with a fixed-fee not-to-exceed for the build.
FAQ
Cash Flow Model Builder FAQ
How is this different from a modeling template?
A template still has to be filled in by hand. The agent assembles the statements and working-capital roll-forward from exam output and historical financials, reconciles the math, and flags assumptions, so professionals start from a defensible draft.
Does it replace the analyst or advisor?
No. It removes the assembly work so professionals spend time on scenarios, findings, and judgment. Professionals own, review, and sign every model.
Why does so much of the model already exist?
Once a field exam is complete, the roll-forward, eligibility, and working-capital detail supply most of the inputs a comprehensive model needs, so the model is largely an assembly and reconciliation task.
Can it produce scenarios for stakeholders?
Yes. The agent builds upside, downside, and stress scenarios on shared assumptions, with sensitivity on margin, collections, and advance rates, and draft commentary for review.
What is needed for a modeling pilot?
A focused pilot typically starts with the firm's model structure, a recent field exam, historical financials, the debt and borrowing-base terms, and the forward assumptions professionals normally apply.
Related examples