Expertise

What I deliver,
and what your team keeps.

Investigations, reconciliation tools, reporting packs, forecasts and integrations, scoped to the problem and the systems you use.

01 / Financial problem-solving & reconciliation

Trace and explain discrepancies.

Forensic work on data-heavy finance problems: trace transactions across systems, find the cause of a discrepancy and quantify the correction.

Outcome

A documented explanation of what went wrong, its effect on the numbers and the correction needed, with checked workings and an evidence trail.

Example deliverables
  • Forensic financial investigations

    Reconstruct transaction histories across ledgers, bank records, payment providers and source files. Test explanations for unexplained balances or missing receipts, and document the evidence, financial impact and resolution.

  • Data cleanup & corrections

    Identify duplicates, gaps, mapping errors and broken allocations across large datasets. Prepare checked correction files or journals for approval, with a record of the changes and control totals before and after.

  • Reconciliation workbenches

    Bank, payment-provider and ledger records in one repeatable matching process: defined rules, an exception queue and a record of how each item was resolved.

02 / Reporting, cash flow & forecasting

Reporting and forecasts traceable to source.

Management reporting and cash-flow forecasts built on source data, so every figure can be traced.

Outcome

Reporting your team can refresh and explain, with consistent definitions. Forecasts with visible assumptions, measured against actuals.

Example deliverables
  • Management packs & dashboards

    A repeatable pack: profit and loss, balance sheet, cash and KPIs, with budget or prior-period comparison and a route back to the supporting records.

  • Cash forecasts & scenarios

    Rolling forecasts, including a 13-week cash flow forecast, built from collection and payment timing. Adjustable assumptions, comparable scenarios, and forecast measured against actual.

  • Group & multi-currency reporting

    Entity reporting templates, account mappings, intercompany schedules and currency translation workings for a consistent group view.

03 / Connected systems & automation

Fewer manual steps between your systems.

Integrations and internal tools between your operational systems and the ledger, with validation at each step.

Outcome

Data moves between your existing systems without re-keying. Incomplete inputs are flagged, and people approve what reaches the books.

Example deliverables
  • Connected finance workflows

    Move approved data between sales, billing, banking and accounting systems through integrations or controlled imports, with validation, error handling and a log of each run.

  • Internal finance tools

    Purpose-built tools for exception handling, collections, approvals or recurring controls, with user access, status tracking and an audit trail.

  • Accounting-system migrations

    Data cleanup, mapping and migration support, with opening balances reconciled, a parallel-run plan and checks before the team changes over.

04 / Processes, controls & handover

Processes more than one person can run.

Documented rules, responsibilities and controls, so month-end and recurring work can be run, reviewed and handed over.

Outcome

Written procedures, approval responsibilities and working records, with walkthroughs so more than one person can run each process.

Example deliverables
  • Month-end plans & handovers

    Close calendar, task owners, working-paper templates and runbooks: what to do, when, and what evidence to keep.

  • Payment & supplier controls

    Review workflows for payment batches, supplier changes and procurement: duplicate and bank-detail-change checks, approval limits, and a record of who authorised each action.

  • Process reviews & procedures

    Process maps, finance policies and an action plan for control gaps or repeat audit findings, with an owner for each action.

AI & large language models

AI for defined finance tasks,
with review built in.

I set up local and cloud large language models (LLMs) inside finance departments, as tools for the team or as a repeatable process that runs each day.

01 / Set up & enable

An AI workspace for finance.

A configured local or cloud LLM environment with team access, approved document sources, reusable instructions and a user guide.

  • Local, cloud or combined deployment
  • Finance knowledge assistants with source references
  • Access rules, usage guidance and handover
02 / Build & improve

Tools built with your team.

Spreadsheets, scripts and internal applications built with AI assistance, with documented finance logic and tests against known results.

  • Reconciliation workbenches and reporting utilities
  • Integrations and recurring data preparation
  • Test cases and maintainable documentation
03 / Process & review

An AI processing workflow.

A recurring workflow that extracts document data, proposes classifications and routes exceptions, with review screens and an audit trail.

  • Invoice or remittance processing queues
  • Exception summaries and draft commentary
  • Validation rules and human approval steps
Local, cloud or both

Choose where the work runs.

The choice depends on your data, result quality, expected usage and who maintains the setup. We agree what each tool can access and where data is processed and stored.

Local
Models on your own infrastructure, for closer control over processing, if you can support the hardware and maintenance.
Cloud
Hosted models, chosen on the task, provider terms, data handling and usage cost.
Both
Local and cloud together, with rules for which data and tasks use each.
Controls

From source record to approved output.

Calculations and posting rules stay explicit. Testing uses representative finance data, source evidence is retained, and exception review has a named owner. Ongoing quality checks, running costs and support are scoped.

  1. Source records
  2. AI-assisted processing
  3. Checks & human review
  4. Approved output
Talk about AI in finance
Systems

Built around
the systems you use.

Your accounting software stays the ledger. I work on data quality, integrations, controls and tools around it, and on migration when a change is needed. Systems I have worked with:

Accounting and finance

  • Access Financials
  • QuickBooks
  • Excel
  • LibreOffice Calc

Payments and banking

  • Stripe
  • Payfast
  • Netcash debit orders
  • Lightspeed POS
  • ThriveCart
  • Bank and payment-provider APIs
  • SWIFT and SEPA payments
  • Crypto settlement and on-chain data

Data and tools

  • Python
  • SQL (PostgreSQL, SQLite)
  • VBA
  • PowerShell
  • TypeScript
  • R Shiny
  • Streamlit dashboards
  • Windows desktop tools

Business apps and AI

  • Google Workspace
  • Microsoft 365
  • Dubsado
  • Claude and OpenAI APIs
  • Local AI models (Ollama)

Hosting

  • Proxmox
  • Cloudflare
  • nginx
  • Self-hosted email
  • GitHub Actions
How engagements work
Contact

Describe the problem.

Send a short description of where the work gets stuck. I’ll reply with whether I can help and a suggested first step.

Get in touch