Make your first use case live, for a year, for £49,500 (fixed price)*
Optical Character Recognition gives you text. Intelligent Document Processing (IDP) gives you so much more. It lets you describe what you want it to do, in natural language, and then leave it to do the rest. Pull the required data from any document with any layout, typed or handwritten, in multiple languages, and then automatically trigger the processing of that data in your chosen downstream system. IDP includes a confidence score with any data it extracts, so that you can choose to involve a human to review it, if you want to.
Public sector organisations typically continue to have multiple processes that require paper or digital document interpretation. Here are some examples of use cases for which MuleSoft Intelligent Document Processing could be used in the Public Sector.
Choose your sector
Invoices, credit notes, purchase orders, goods receipts and applications for payment.
Fleet, uniform, forensics
Planning & building, waste, social care
Consumables and agency
Works and stations
Labs and field ops
Estates, labs, catering
Interpreters, escorts, prison facilities
Call-off terms, renewal dates, and the monthly performance reports that arrive as PDFs.
Collaboration agreements
Outsourced services
Provider and supplier contracts
Works and concessions
Framework call-offs
Research and supply contracts
Tagging and prisoner escort contracts
Certificates of conformance, test certificates, insurance cover and ISO certification.
Vehicle and equipment
Insurance and ISO
Device and calibration
Station inspections, near-miss and shift reports
Trade certification and disease control
Lab and equipment certs
Security systems and fire safety
Right to Work, DBS, professional registration, licences, training and competency.
Vetting and DBS
DBS and registration
Clinical registration
Driver CPC and licences
Right to Work checks
Right to Work and sponsorship
Prison officer vetting and training
Applications and referrals, plus the statements, payslips and ID inside them.
Statements and disclosure
Planning, adult and children's social care
Referrals and correspondence
Permits and consents
Authorisations, evidence capture and transcription
Admissions and transcripts
Case issue, court bundles, skeleton arguments
Drawings, O&M manuals, as-builts, leases and historic case files.
Estate drawings
Property and leases
Estates O&M manuals
As-builts and drawings
Site plans and manuals
Campus drawings and manuals
Paper prison and probation case files
Ashford Borough Council runs Port Health checks for imports coming in via Dover and the Eurotunnel. Every consignment arrives with paperwork that somebody had to read and check by hand before it could be cleared. This is what happened when that step was automated with MuleSoft Intelligent Document Processing.
£3.5 million
expected operational savings in year one
30 → 4 minutes
expected document checking time, and the ability to run several jobs at once
Happier staff
less time on paperwork, more time using their expertise
>100% productivity gain in week one
Expected to now meet government targets
>97% of checks rated easy or very easy
Ashford's figures are their own, including the word "expected". Nothing here is claimed about your own volumes or savings.
A fixed-price start on your own documents. We take one document workflow end to end, and you choose where the extracted data lands: surfaced and validated in the Salesforce platform, or pushed straight to one of your downstream systems.
1 Step one
We take a single document type that matters to you and build the extraction in MuleSoft IDP: fields, confidence thresholds and the exception path.
2 Step two, whichever suits you
Every document lands as a structured record in your org, with confidence scores, a review queue and a full audit trail. Your people confirm the read, then search, report and act on it.
The extracted record is handed straight to whatever system runs the next step. Your system receives data it can act upon, not another document.
£49,500 is a fixed price for taking one document workflow live and running it for twelve months. It covers discovery and field design for a single document type, the extraction model built in MuleSoft Intelligent Document Processing, confidence thresholds and the exception path, and one destination of your choosing: either the structured record and validation in Salesforce, or the handover of that data to one downstream system. It also covers testing against your own documents, documentation and handover to your team.
It assumes one document type and one destination: one review queue if you choose Salesforce, one downstream system if you choose the push. It excludes any change required inside your downstream systems, further document types, and the migration or back-processing of historic records. Those are quoted separately.
It is designed for documents of standard complexity: a few pages long, with the same kind of information in each one, such as invoices, certificates or application forms. Documents that need additional validation, or that are of higher complexity, may not fit within the Quickstart. Where that is the case, we scope and quote them separately.
It is a fixed price, not a guaranteed outcome. Accuracy, handling time and savings depend on your own documents, volumes and processes. Any figures we put in front of you will be measured on your documents, in your systems, once the workflow is live. Nothing on this page is a commitment to replicate another organisation's results.
The Quickstart is the place to start. If your documents are harder to read, or you want the data checked and acted on once it is out, there are two further fixed-price options. Like the Quickstart, each takes one document workflow end to end, on real documents. What separates them is how hard the document is to read, and how much is done with the data once it is out.
£74,500
One document workflow where the data is checked by your people, then put to work in your downstream system (for example Salesforce, Microsoft, AWS or Oracle).
£99,500
For document types that are long, poorly scanned or hard for AI to read. Our Sensor application prepares them before extraction.
Any numbers we put in front of you will come from your own systems, measured on your own documents. Get in touch and we will start on one workflow that matters to you.
Prefer email? Write to us at info@cospiretech.com