Phase 0
Technical & Operational Baseline Review — a short, focused engagement that establishes a clear, factual view of your technical estate before anyone proposes a fix, a product, or a project.
Every engagement starts here. Most problems are misdiagnosed before they are mis-solved.
Analysis only — no tools sold, no vendors recommended, no long-term commitment. Many clients stop after Phase 0, and that is a valid outcome.
- Years inside production networks and systems
- 20+
- African countries with delivered engagements
- 16
- Practice areas under one accountable team
- 8
- Google Global Cache in East & Central Africa
- 1st
The review is conducted by engineers who have run NOCs, originated AfriNIC resources, rebuilt carrier cores, secured bank connectivity, and shipped production software — not by analysts reading your documentation back to you.
01 — Why Phase 0
Why Phase 0 Exists
Most technical initiatives fail not because the solution was wrong, but because the problem was misunderstood. A team spends a year replacing the access network when the churn was caused by a duplicate NAS entry in RADIUS. A bank buys a modelling platform when the real problem is that nobody can reproduce last quarter’s provision. A business migrates to the cloud and discovers its costs doubled because nobody owned the architecture.
By the time these become visible, the budget is committed and the decision is political. Phase 0 exists to intervene before that point — it answers one question only:
What is actually happening in our network and systems today?
Until that is clear, strategy is guesswork — and guesswork at infrastructure scale is expensive.
02 — Coverage
What This Review Covers
Eight review tracks spanning every practice range. We scope to the tracks that match your estate — you do not pay attention to areas that are not your problem.
Networks & ISP Operations
The access, core, and operational layers — reviewed by engineers who have run operator networks, not read about them.
- Physical and logical topology, access, aggregation, and core design
- Internet edge, upstream mix, peering and transit economics
- BGP, MPLS, SD-WAN, and route-hygiene review (IRR, RPKI)
- Satellite and hybrid connectivity integration where applicable
- Failure domains, single points of risk, and capacity headroom
- NOC workflows, monitoring effectiveness, escalation and change discipline
Enterprise IT & Productivity
The estate the business actually works in — identity, endpoints, servers, and the licences funding it all.
- Microsoft 365 tenant posture and licensing right-sizing (Business through E3, E5, E7)
- Active Directory and Entra ID health, hybrid identity, and conditional access
- Server and virtualization estate: patch state, documentation, and lifecycle risk
- Endpoint management, compliance baselines, and device hygiene
- ERP and line-of-business system fit — including Odoo and Dynamics estates
- Unified communications, telephony, and the support model behind them
Cybersecurity & Exposure
What an attacker sees, what you would detect, and what you could actually do about it.
- External attack surface and exposure review — the internet-facing truth
- Identity security: MFA coverage, privileged accounts, and admin tiering
- Segmentation and management-plane isolation across network and cloud
- Detection reality check: what is logged, what is monitored, what would page someone
- Backup, immutability, and ransomware recovery posture
- Compliance gap read against ISO 27001, Kenya DPA, GDPR, or PCI DSS as relevant
Cloud & Infrastructure
Architecture, operations, and the bill — assessed together, because they are the same subject.
- Account, project, and network architecture across AWS, Azure, and GCP
- Hybrid connectivity and routing symmetry between cloud and on-premises
- Container and Kubernetes estate review — or readiness, where it is being considered
- Cost visibility: tagging, allocation, waste, commitment coverage, and egress exposure
- Resilience: multi-AZ design, backup posture, and whether failover has ever been tested
- Operational ownership — who actually runs each system, and what happens when they leave
Software, APIs & Integration
The glue between systems — where silent failures quietly cost money every month.
- Integration inventory: what talks to what, over which contract, with whose ownership
- Payment, billing, and provisioning chains — idempotency, retries, and failure handling
- Reconciliation coverage: where drift between systems would go undetected
- Manual and repetitive operational work that should be automated
- Technical debt and key-person risk in internally built tooling
- Delivery practice: version control, testing, deployment, and documentation discipline
Data, Analytics & BI
Whether the numbers in your dashboards can be trusted — and what it costs to produce them.
- Data source inventory, lineage, and the pipelines feeding reporting
- Warehouse, lake, or lakehouse architecture review — including Hadoop estates
- Data quality: where records are dropped, duplicated, or silently transformed
- Metric definitions and semantic layer — whether one number has one meaning
- Reporting estate rationalisation and self-service readiness
- Platform and compute cost against the value actually being produced
FinTech & Credit Risk
For banks and lenders: whether the provisioning model survives validation, and whether anyone can reproduce it.
- IFRS 9 ECL model review: PD/LGD/EAD approach, staging logic, and SICR criteria
- Reproducibility test — can last period’s provision be regenerated from source data?
- Model documentation, validation status, and outstanding audit or supervisory findings
- Forward-looking overlays: scenario design, weighting, and the evidence behind both
- Spreadsheet risk: undocumented macros, broken references, and key-person dependency
- Payment and mobile-money integration integrity, including wallet-to-ledger reconciliation
AI Readiness
An honest read on whether AI would help you yet — and what has to be true first.
- Use-case triage: which candidates are real, which are baselines in disguise
- Data readiness — the precondition most AI programmes discover too late
- Deployment options: hosted APIs vs open-weight models on your own infrastructure
- Data-sovereignty, privacy, and regulatory constraints on model use
- Governance readiness: model risk, explainability, and audit expectations
- Team capability and the training required to operate what gets built
Every track links to the practice behind it — but Phase 0 is diagnosis, not a sales route into those practices.
03 — Methodology
How Phase 0 Is Conducted
This is an analysis-only engagement. We are not positioning a product, and nothing in the review is written to make a follow-on project look necessary.
- No tools sold
- No vendors recommended
- No architecture imposed
The focus is on evidence, not opinions — configurations, telemetry, and incident history over what the documentation claims.
Activities typically include:
- Structured technical interviews with the engineers who actually run the systems
- Configuration, diagram, and documentation review against live reality
- Metric and telemetry sampling — monitoring, logs, billing, and performance data
- Incident and outage pattern analysis to find repeat causes, not one-off symptoms
- Read-only access where you permit it; nothing is changed in your environment
04 — Deliverables
What You Receive
- A clear technical baseline — the estate as it actually is, documented
- Identified risks and blind spots, with the evidence behind each finding
- Observations grouped by impact and urgency, so triage is obvious
- A short executive summary written for the board, not for engineers
- A factual foundation for decision-making that outlives the engagement
- Direct answers to the specific questions your team brought into the review
No roadmap is pushed unless requested. The deliverable is yours to act on with us, with another partner, or internally.
05 — Process
How the Engagement Runs
Short, time-bound, and built to disrupt your operations as little as possible.
- 01
Scoping Conversation
A direct technical conversation to establish which review tracks apply, what questions you need answered, and what access is appropriate. No charge, no obligation.
- 02
Evidence Gathering
Structured interviews with your team plus configuration, telemetry, and incident review — scheduled around operations rather than through them.
- 03
Analysis & Correlation
We correlate what the systems show against what the team reports and what the incidents prove, separating symptoms from root causes.
- 04
Findings Walkthrough
A working session covering the baseline, the risks, and the prioritised observations — with your engineers free to challenge every finding.
- 05
Your Decision
Act internally, take the report to another partner, or engage us for what follows. All three are normal endings, and we say so up front.
Delivered remotely or hybrid, with on-site work where the estate genuinely requires it. Duration and commercials are scoped to your environment and discussed privately.
06 — Boundaries
What Phase 0 Is Not
- Not a sales pitch
- Not a compliance audit
- Not a transformation program
- Not a long-term commitment
- Not a tooling proposal
Phase 0 stands on its own. If the honest finding is that your estate is sound and needs nothing, that is what the report will say.
07 — Outcomes
Typical Outcomes
Organizations completing Phase 0 usually gain:
- Immediate clarity on where attention is actually needed — and where it is not
- Reduced internal disagreement between teams who were each describing a different symptom
- Confidence to pause, proceed, or reprioritize initiatives, with evidence behind the call
- A defensible technical position to take to the board, the regulator, or the auditor
- Avoided spend on projects the baseline shows would not have solved the problem
Many clients stop here — and that is perfectly acceptable. A review that prevents an unnecessary project has already paid for itself.
08 — Audience
Who Phase 0 Is For
Phase 0 is most valuable when:
- Systems are business-critical and outages or inefficiencies are genuinely costly
- Leadership wants clarity before committing further budget or headcount
- Teams disagree about the root cause and the debate has stopped being productive
- An estate has been inherited, acquired, or inadequately documented
- A major decision is imminent — a migration, a rebuild, a vendor change, or an audit
It is not necessary when systems are trivial or non-critical, and we will tell you if that is what we see.
09 — Questions
Questions We Are Usually Asked
The practical concerns that decide whether this is worth your time.
- We already know our problems. What would this add?
- Often it confirms what your engineers have been saying — which is itself valuable when leadership needs an independent voice. Where it adds most is in the gaps between teams: the issues nobody owns because they sit between the network, the systems, and the billing stack.
- How disruptive is this to our operations?
- Deliberately minimal. The work is interviews, document and configuration review, and telemetry sampling — scheduled around your operations. We do not change anything in your environment, and read-only access is granted only where you consider it appropriate.
- Will this turn into a proposal for a large project?
- Not by design. Phase 0 is analysis-only and the report is written to be actionable by anyone — your team, another partner, or us. If the honest answer is that nothing needs doing, the report says that.
- Who from our side needs to be involved?
- Typically the engineers who run the systems day to day, whoever owns the commercial or operational outcome, and — briefly — the executive sponsor. The heaviest demand is a few hours of your technical team’s time.
- What about confidentiality and access to sensitive systems?
- We work under NDA as standard and scope access to the minimum required — frequently read-only, and often nothing more than exports and configurations your team provides. Sensitive material stays within agreed boundaries.
- Our estate spans several of these areas. Is that a problem?
- It is the reason the practice exists. The failures that hurt most rarely respect the boundaries between network, cloud, security, data, and applications — reviewing them as one system is the point.
- What does it cost, and how long does it take?
- Both are scoped to the estate and the tracks in play, and are discussed privately in the scoping conversation. That conversation carries no charge and no obligation.
10 — Engagement
Duration & Engagement Model
- Short, time-bound engagement
- Minimal disruption to operations
- Remote or hybrid delivery
- Under NDA as standard
Details are discussed privately.
Closing
11 — Closing
Phase 0 exists to replace assumptions with facts.
Everything that follows — or does not follow — becomes clearer once this baseline is established. The conversation that starts it is free, technical, and short.
Request a Phase 0 Review
Let's establish a clear picture of your current technical environment — then decide what, if anything, needs to change.