DocsOverview
Other ways to solve the problem.
All documentation pages
Each option can be useful. The right choice depends on how much of the process it needs to cover.
Screen automation and RPA
Useful for: repetitive steps in a user interface. If the screen is the only way into a system, this is how you connect it.
Limit: it moves the click. Reusable business records, calculations, matching, exception handling, and supporting evidence require more work. Each workflow is maintained separately.
How Manuel handles it: we use screen automation as one way to connect when no API exists. The control is managed separately from the screen steps.
Generic workflow builders and low-code
Useful for: building almost any workflow, quickly, with a visual canvas. Enormously flexible.
Limit: flexibility still means you design and maintain the process. At production volume, a long chain of generic steps and repeated model calls can also increase the cost per item.
How Manuel handles it: the finance and supply chain controls arrive as working processes. You state the rule and Manuel runs it.
Document capture, OCR and IDP
Useful for: turning a page into fields. Modern tools do this well.
Limit: extraction is one early step. The team still has to identify the supplier, calculate the expected amount, compare it with the contract, and handle mismatches.
How Manuel handles it: extraction is one stage of seven. What comes out is a business object with provenance, enrichment, calculation state and validation status, ready for a decision.
A general-purpose AI assistant in the finance team
Useful for: interpretation, investigation, summarising, explaining. Genuinely useful, and your team should be using one.
Limit: a chat answer may lack a version, source history, replay, approval step, and owner. A later answer may differ without a record of the change. It can also add cost when a lookup or fixed rule would be enough.
How Manuel handles it: AI handles interpretation and investigation. Versioned rules handle calculations, matching, checks, decisions, and records.
A module in your ERP or core system
Useful for: controls that live entirely inside one system, on data that system already owns.
Limit: this problem is cross-system by nature. It needs the bank statement, the processor file, the contract, the master data and the ledger at once, and no single vendor's module holds all five. Changing one is an implementation project.
How Manuel handles it: connect the required records across systems. Rule changes are published as configuration.
Building it internally
Useful for: getting exactly what you want. A competent engineering team can build any single control here.
Limit: maintenance grows with each new control. Rules, matching paths, exception categories, retained records, replay, and approvals all need long-term owners.
How Manuel handles it: provide the shared control infrastructure so each new control starts from an existing base.
Outsourcing to a BPO or shared service
Useful for: absorbing volume quickly without hiring, and moving cost to a lower-cost location.
Limit: the team may still have weak version control, key-person dependency, and difficult audit work. Exceptions remain expensive when a case arrives without its calculation, match history, or supporting records.
How Manuel handles it: the exception carries its own context. Whoever resolves it, in-house or outsourced, starts with the answer already half-assembled.
Mature products in these categories can configure matching, handle exceptions, run at volume, and pass an audit. Compare how much of the full process each option covers and what it costs to maintain.