// competency

ECM & Document Management

Electronic document workflow, ECM systems, IDP and approval automation. Without forced Documentum migrations "because we have the budget this quarter".

// about the practice

What it is and who needs it

The ECM practice covers design and implementation of enterprise content management systems: contracts, orders, HR documents, technical documentation. Includes approval workflows, electronic signatures (QES), archival storage, ERP/CRM integration.

In 2026, the key addition is AI/IDP (Intelligent Document Processing) for automatic classification, data extraction and routing of documents without manual work by a secretary.

You need this if

  • Contract approval from request to signature takes 2+ weeks.
  • You have legacy ECM (Documentum, OpenText) with end-of-life support.
  • Regulator requires electronic signatures and audit trail.
  • Secretaries/lawyers spend >50% of time on manual document processing.
  • Transition from paper to electronic document workflow is approaching.
// our position

Why we do this differently

01

ECM migration projects almost never fail because of data — they fail because new approval routes do not handle real day-to-day exceptions.

The technical part is 30% of the difficulty. The other 70% is getting lawyers, finance and accounting to agree on unified workflows without 47 exceptions "for our case".

02

IDP without clean training data is a surprise generator. Data first, AI second.

Before deploying an AI document classifier, we audit the existing corpus: do templates exist, are types known, is formatting uniform. Without this, 30% of documents are misclassified.

03

Do not replace ECM "because it is old". Replace it when vendor support has ended or integration with new systems became impossible.

A working ECM that covers needs is the best ECM. Migration is a 12–18 month budget and risk of breaking working processes. Do it only when really needed.

// honest filter

When you need this — and when you don't

It is more honest to say "you do not need this yet" than to sell an engagement that will not deliver ROI.

✓ Need it

  • >500 documents/day in workflow
  • Legacy ECM with end-of-life vendor support
  • Regulatory requirements for QES/audit trail
  • Business side ready with power users in each department
  • Strong executive sponsor for the project

✗ Not yet

  • Replacing ECM "because it is old", with no actual problems
  • Small business <50 staff
  • No power users ready to spend 30% of time on the project
  • You want "AI document workflow in 2 months"
// process

How we run the engagement

01

Discovery & process mapping · 4–6 weeks

Audit of existing approval processes, identification of exceptions and shadow processes, mapping of document types and regulatory requirements.

02

Target architecture · 3–4 weeks

ECM platform choice, workflow design, ERP/CRM integration, QES integration, archive strategy.

03

Pilot on one document type · 2–3 months

Usually we start with contracts — highest business impact, visible exceptions. Pilot closes 80% of technical risks.

04

Phased rollout · 6–12 months

Phased — orders, HR documents, technical documentation. Each type — separate sprint with its own onboarding.

05

IDP layer (optional) · 3–6 months

Adding AI classification, data extraction from documents, automatic routing. Only after the basic workflow works reliably.

Project lead: Softline and IQusion (Megapolis.DocNet implementation, registry integration).
Brought in when needed: InBase (Scriptum for workflow), Nectain (Nectainium for AI document processing).

// anti-patterns

Typical mistakes we have seen projects fail on

Big bang ECM migration

Attempt to replace the entire legacy ECM in one project. Plan duration 12 months, actual 24. Old processes break before new ones stabilize.

What we do instead: parallel run of old and new ECM for at least 6 months. Migration is phased by document type.

Workflow without exceptions

They design the "ideal" workflow first. In production, 30% of documents turn out to be exceptions. Users bypass the system via email.

What we do instead: workflow with built-in exception handling (escalation paths, manual override with audit). Exceptions are not bugs but expected behavior.

AI/IDP without training corpus

Buy an enterprise IDP platform expecting it to "learn on its own". Six months later — 60% accuracy, meaning every third document is processed incorrectly.

What we do instead: minimum 500 labeled documents per type before production. For rare types — rule-based fallback.

// experience

Typical scenarios from our practice

No precise savings percentages — actual numbers depend on the client's starting point. Instead — concrete architectural decisions and organizational changes.

Administrative court · Megapolis.DocNet

Migration from legacy to Megapolis.DocNet

The Softline team moved the court archive to Megapolis.DocNet in 4 months. The main time was spent not on document migration but on agreeing new approval routes with judges.

Energy holding · 14 regional branches

Unified document workflow on UnityBase

Each branch had its own ECM. Unification via Scriptum (low-code BPM on UnityBase). 11 months. The hardest part — agreeing unified document type nomenclature across branches.

Bank · IDP for KYC documents

AI classification and data extraction from passports/IDs

Nectainium for automatic data extraction from customer documents. KYC processing time dropped from 15 minutes to 90 seconds. Accuracy 96% for Ukrainian documents, 87% for foreign.

// deep dives

Articles on this topic

Nine recent expert articles — from thematic overviews to specific architectural decisions.

// stack

Technologies we work with

ECM platforms

Megapolis.DocNet · Scriptum.DMS · UnityBase · OpenText · M-Files · Alfresco · SharePoint

IDP & AI

Nectainium · Microsoft Form Recognizer · ABBYY Vantage · Google Document AI · UiPath Document Understanding

Workflow & BPM

Scriptum · Camunda · Bonita · jBPM · Activiti

Electronic signatures

КЕП (Україна) · QES (EU) · DocuSign · Adobe Sign · Diia.Signature

Archival storage

Iron Mountain · AWS Glacier · Azure Archive · MinIO · S3-compatible storage

Standards

ISO 15489 · ISO/IEC 27001 · eIDAS · ДСТУ 4145 · MoReq2010

// frequently asked

Frequently asked questions

How does Scriptum differ from Megapolis.DocNet?

Scriptum is a low-code BPM platform on UnityBase from InBase, for building workflows and custom business applications. Megapolis.DocNet is a specialized ECM for electronic document workflow with built-in approval features. They partially overlap: for classic document workflow — Megapolis, for custom workflows — Scriptum.

How long does ECM implementation take?

Pilot on one document type — 2–3 months. Full implementation for a mid-size organization — 9–14 months. If you are promised "ECM in 3 months", that is likely base configuration without real integration.

Can paper document workflow be replaced entirely?

Technically — yes. Legally — in Ukraine yes for most document types since 2017 (with QES). In practice, 5–10% of documents remain in paper form due to counterparty or process specifics.

How to choose between cloud and on-premise ECM?

Cloud — for most cases (faster deployment, lower TCO). On-premise — when documents have restricted access (KSZI requirements), when regulator prohibits cloud, or when there are existing data center investments.

What is IDP and when does it pay off?

IDP (Intelligent Document Processing) — AI systems for automatic classification and data extraction from documents. Pays off from ~5000 documents/month per type. Below — manual processing is cheaper than IDP infrastructure.

// other competencies

Related competencies

Real projects rarely fit in one competency. See which other areas we work in.

Tell us about your document workflows — we will see where to simplify

30-minute discovery call with an ECM architect. We will discuss current state and options, no commitments.

All contacts