AcademyUnderstand

Technology Portfolio Intelligence

Why inventories decay, how to reconcile evidence from many sources into one trusted record, how to grade the evidence behind each attribute, what AI and agents change, and how to use portfolio intelligence for decisions on cost, risk and change.

U101Foundation13 slidesAbout 11 minutes

Start the course Download the PDF
1 / 13
PDF
AxiareteAcademy

U101Understand · Foundation

Technology Portfolio Intelligence

Building an evidence-based model of your applications, infrastructure and dependencies that stays current.

FormatKnowledge pack · 13 slides
ForArchitects and portfolio managers
LevelFoundation
Axiarete Academy · U10102 / 13

The premise

A portfolio model maintained continuously from evidence

A survey-based inventory begins to go out of date while it is still being compiled. Portfolio intelligence is the capability to answer questions about the estate from current evidence, whenever they are asked.

CurrentRefreshed from the systems themselves, by scan and event.
EvidencedEvery attribute carries its source and its evidence grade.
ConnectedApplications linked to capabilities, data, infrastructure and cost.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10103 / 13

Purpose

Eight questions it must answer on demand

1What do we run?Every application, service and platform, including unregistered SaaS and AI agents.
2Who owns it?An accountable business owner and a technical owner.
3What does it do?The capabilities and processes it supports.
4What does it depend on?Upstream and downstream systems, data and identities.
5What does it cost?Fully loaded run cost, not license fees alone.
6What data does it hold?Data classes, sensitivity and retention duties.
7How exposed is it?Vulnerabilities, end-of-support components, privileged access.
8What breaks if it changes?The blast radius of a change, migration or outage.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10104 / 13

Evidence

What each evidence source sees and misses

SourceSees wellMisses
CMDB and service catalogDeclared ownership and service relationshipsDrift and unregistered systems
Discovery and cloud APIsWhat runs, where, and how it is configuredBusiness purpose, ownership and SaaS
Network flowsActual communication between systemsOwning application, payloads, jobs outside the window
Source code and buildsFrameworks, libraries and internal dependenciesDeployed versions, configured endpoints
Identity and access logsWho signs in to what, and how oftenAnything outside SSO, and activity after sign-in
Finance and contractsCost, vendors and renewal datesDeployment and actual use
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10105 / 13

Method

From raw evidence to a trusted record

1CollectConnect sources read-only and keep the raw evidence.
2NormalizeApply common identifiers, names and versions.
3ReconcileMatch entities across sources; set source precedence per attribute.
4GradeRecord each attribute’s source, grade and last-verified date.
5RelateLink capabilities, dependencies, data flows and cost.
6MaintainDetect drift continuously and route exceptions to owners.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10106 / 13

Confidence

Grade the evidence behind every attribute

Each decision should state the grade it requires. Retiring an application needs observed usage data over a full business cycle.

UnknownNo evidence yet. Recorded as a gap to close.
InferredDerived from names, patterns or adjacent records.
DeclaredStated by an owner or entered manually in the CMDB.
ObservedSeen directly by discovery, telemetry, network flows or access logs.
ConfirmedTwo independent sources agree, at least one of them observed.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10107 / 13

Structure

Model the estate in connected layers

CapabilitiesWhat the organization does, independent of systems. It changes least often.
ApplicationsWhat delivers each capability, with owners and lifecycle stage.
Components and dataCode, libraries, databases, interfaces and data classes.
InfrastructureHosts, containers, platforms, networks and cloud accounts.
IdentitiesPeople, service accounts, AI agents and the privileges that link them to systems.
Why layersMost questions cross layers. Finding which capabilities rely on an end-of-support database means following the links.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10108 / 13

Dependencies

Six kinds of dependency to map

Observation must cover a full business cycle. Quarter-end and year-end processes stay invisible to a short sample.

Runtime callsAPIs, queues and service-to-service traffic.
Shared dataCommon databases, file transfers and extract jobs.
IdentityService accounts, single sign-on and shared credentials.
InfrastructureShared hosts, clusters, DNS, certificates and networks.
SchedulesBatch jobs that run monthly, quarterly or once a year.
PeopleKnowledge held by only a few individuals.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10109 / 13

Use

Decisions the model supports

DecisionThe question it answers
RationalizationWhere capabilities overlap, and what is unused or redundant
ModernizationWhat depends on a system, and how hard it is to change
SecurityWhat is exposed, and what an attacker could reach from it
CostWhat each capability costs to run, and who consumes it
ResilienceWhich paths are critical, and where single points of failure sit
Mergers and acquisitionsWhere two IT environments overlap, and what a carve-out must separate
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10110 / 13

AI and agents

What AI and agents change

Role of AIWhat changesWhat to require
Part of the estateModels, agents and AI features in SaaS products need records, as any application doesOwner, purpose, data accessed, tools called and the identity each agent uses
Evidence readerLanguage models extract owners, terms and purpose from contracts, tickets and runbooksGrade the output Inferred until evidence or the owner confirms it
Entity matcherModels match one application across sources that name it differentlyKeep each match with its reason, and review a sample before relying on it
Consumer of the modelAgents that assess change impact or triage incidents query the model, often through MCP serversReturn each fact with its grade and last-verified date; limit what each agent can read
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10111 / 13

Measurement

Measures of record quality and use

CoverageShare of applications seen directly by at least one automated source.
FreshnessMedian and 90th-percentile time since key attributes were last verified.
ConfidenceShare of key attributes graded observed or confirmed.
OwnershipShare of applications whose owner confirmed the record in the last six months.
Dependency coverageShare of critical applications whose upstream and downstream links are observed.
Time to answerTime to answer the eight questions for one application.
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10112 / 13

Pitfalls

Why inventories fail

Starting with surveys
Start with evidence, then ask owners to confirm
Waiting for completeness
Make it useful for one decision first
Running it as a project
Run it as a standing capability with an owner
Treating the CMDB as truth
Treat it as one source among several
Sampling a short window
Observe a full business cycle
Collecting without using
Tie every attribute to a decision that needs it
Technology Portfolio IntelligenceUnderstand
Axiarete Academy · U10113 / 13

Summary

Key takeaways

  1. Refresh the record continuously from evidence; use surveys only to confirm it.
  2. Reconcile many sources and grade every attribute by the evidence behind it.
  3. Anchor the model on business capabilities and follow dependencies across layers.
  4. Observe a full business cycle before declaring anything unused.
  5. Measure coverage, freshness, confidence and the time it takes to answer a question.
Next course · U201AI-Ready DataUnderstand · Practitioner
Technology Portfolio IntelligenceUnderstand

Further reading

Continue learning

U201 · UnderstandAI-Ready DataAxiarete AcademyAll courses and tracks

Bring this course to your leadership team

We run private working sessions that apply these frameworks to your own estate, data and pending decisions.

Request a Private BriefingExplore the Academy