Most AI pilots fail. Not because the AI doesn't work.
They fail because the data underneath was never ready. The warehouse has three different definitions of "customer." The CRM doesn't talk to the ERP. Half the ETL pipelines run on cron jobs that someone wrote in 2018 and nobody's touched since. There's a spreadsheet that Linda in accounting emails to six people every Tuesday, and that spreadsheet is the closest thing the company has to a single source of truth.
So a vendor pitches an AI platform. Leadership buys it. The implementation team spends four months trying to connect it to data that lives in 14 disconnected systems, none of which have documented schemas. The pilot produces garbage results because the training data is incomplete, inconsistent, and stale. The project gets shelved. The vendor blames the team. The team blames the vendor. Everyone blames "AI."
The AI wasn't the problem. The foundation was. And until someone actually diagnoses what's broken in the data layer, every AI initiative will follow the same path. The Data Readiness Assessment exists to give you that diagnosis: a structured, evidence-based evaluation of where your data infrastructure stands today and exactly what needs to change before AI can work.
The Integration Maturity Index: a credit score for your data infrastructure.
The IMI is our proprietary framework for measuring data and integration readiness. It produces structured, repeatable, and benchmarkable scores across four dimensions. Think of it less as a consultant's opinion and more as a diagnostic instrument.
Data Architecture & Quality
Schema documentation, sources of truth, and data freshness and completeness. How well-structured is your data, and can you trust it?
Integration & Connectivity
API documentation and versioning, integration architecture, and system dependency mapping. How do your systems talk to each other, and is it held together with duct tape?
Governance & Operations
Data ownership and stewardship, lineage tracking, SLAs, and incident response. When a pipeline breaks at 2 AM, does someone know about it?
AI & Analytics Readiness
Feature engineering capability, ML training data availability, MLOps maturity, and inference infrastructure. Can your organization actually use its data for intelligence?
Each dimension is scored independently, producing an overall IMI score that maps where you are on the maturity spectrum. The scoring is structured enough to benchmark against peers and track progress over time. Not just a consultant's impression of how things "feel."
Two and a half weeks. Discovery to roadmap. Here's the sequence.
Most assessments land at two and a half weeks. Smaller environments can finish in two; complex organizations with dozens of systems sometimes stretch to three. The cadence stays the same regardless.
Discovery: map the terrain before diagnosing anything
Your Conductor starts with the data architecture itself. Automated documentation analysis and dependency tracing map databases, warehouses, data lakes, ETL/ELT pipelines, and data flows. Manual review fills in the gaps that automation misses: the undocumented API endpoints, the FTP server nobody remembers setting up, the "temporary" Excel-based pipeline that became permanent in 2019.
Simultaneously, 4-8 stakeholder interviews surface the human side. Data producers, data consumers, and leadership each experience the data infrastructure differently. The engineer who built the pipeline knows where the bodies are buried. The analyst who consumes the data knows which reports they don't trust. The CTO knows which vendor is pitching them an AI platform next quarter. These conversations shape the assessment's priorities.
Diagnosis: profile the data, score each dimension
Now the IMI framework comes to bear. Data quality profiling samples core datasets for completeness, accuracy, freshness, and consistency. This isn't a full remediation; it's a diagnostic that quantifies the problem. A CTO who knows "the data quality is bad" is in a different position than one who can tell the board "our customer data has 34% incomplete records and our inventory data is an average of 72 hours stale."
Governance assessment runs in parallel: data ownership, stewardship, lineage tracking, access controls, and compliance posture. Then the AI readiness evaluation: can this infrastructure support ML workloads? Is there feature engineering capability? What's the gap between current state and production-grade AI? Each IMI dimension gets scored, producing a picture that's specific enough to act on.
Prescription: build the roadmap your engineering team can execute
Here's where this assessment earns its fee. Your Conductor isn't a strategist writing recommendations they'll never have to implement. They're a senior data engineer who has built warehouses, migrated databases, and deployed models. The roadmap they produce is scoped at a level where an engineering team can start sprint planning, not at the "consider modernizing your data platform" level of abstraction.
The roadmap sequences the work by dependency and impact. Fix the data quality issues in your customer table before building the ML feature pipeline that depends on it. Consolidate those three redundant reporting databases before layering a governance framework on top. Every recommendation includes estimated effort, technology choices where relevant, and a clear rationale tied back to the IMI scores. The final presentation walks your leadership team through the full picture: where you are, where you need to be, and what it takes to get there.
Seven deliverables. Zero shelf documents.
Everything is designed to be actionable. Your engineering team picks up the roadmap and starts planning. Your CTO presents the IMI scorecard to the board. Nothing sits in a shared drive gathering dust.
IMI Scorecard
Scores across all four dimensions with an overall maturity rating. Comparable, trackable, and specific enough to anchor a board-level conversation about data investment priorities.
Data architecture diagram
A visual map of your databases, warehouses, lakes, pipelines, and data flows. What exists, how it's connected, and where the gaps are. The kind of documentation most organizations wish they had but never built.
Integration landscape map
Every system-to-system connection documented: APIs, file transfers, manual processes, middleware, and message queues. Fragility points, redundancy, and missing connective tissue identified.
Data quality report
Profiling results for core datasets: completeness, accuracy, freshness, and consistency. Numbers your CTO can take to the board, not adjectives like "it could be better."
Governance assessment
Data ownership, stewardship, lineage tracking, access controls, and compliance posture. An honest answer to "who's responsible for this data?" (Often: nobody.)
AI readiness evaluation
Can your infrastructure support AI and ML workloads? Feature engineering readiness, training data availability, inference infrastructure, and MLOps maturity. The gap between where you are and where production-grade AI needs you to be.
Prioritized modernization roadmap
The primary deliverable. A sequenced plan with estimated effort levels, dependency mapping, and recommended technology choices. Scoped at a level where your team can start executing, not at the "consider modernizing" altitude.
Fixed scope. Clear price. A roadmap in three weeks.
Depends on organizational complexity and number of systems. Most assessments land around $50K.
Two and a half weeks is the sweet spot. Two for focused environments; three when the system count is high and stakeholder interviews need to be broader.
A senior data engineer-architect. Not a strategist, not a junior analyst. Someone who's built the systems they're assessing. The recommendations come from someone who'll have to live with them if you hire us to implement.
The assessment is the beginning, not the end.
The roadmap feeds directly into a Data Engineering & Platform Modernization engagement. The same Conductor who ran your assessment scopes and leads the build. Every architecture decision, every integration mapping, and every stakeholder insight carries forward.
For organizations with strong data foundations, the path might lead directly to Applied AI & Intelligent Systems. The assessment tells you which path is right.
Just need the diagnostic?
That's fine. Some clients use the IMI scorecard to build internal buy-in for a modernization budget. Others take the roadmap to their own engineering team. PE operating partners use it to inform board-level investment decisions. The deliverables stand on their own.
Four situations where the assessment pays for itself before the roadmap is done.
Companies whose data is a mess and they know it
Multiple databases, no documentation, ETL pipelines held together with cron jobs and hope. You know it needs fixing but you don't know where to start. The assessment gives you a starting point backed by evidence: here's what's broken, here's the priority order, and here's the effort level. Not opinions. Data.
Organizations preparing for AI adoption
You've heard the pitch. Maybe you tried a pilot; it went nowhere because the data wasn't ready. The IMI scorecard quantifies exactly what "ready" looks like for your organization and what it'll take to get there. It's the difference between "we need to improve our data" and "we need to fix these seven specific things in this specific order before any AI initiative will stick."
PE portfolio companies post-acquisition
The operating partner needs to understand the technology health of a newly acquired company. What state is the data infrastructure in? Where are the risks? What investment is needed? The IMI scorecard gives the board a quantified, structured picture they can compare across the portfolio. Not a narrative; a diagnostic.
CTOs inheriting legacy systems
New CTO walks into a company with a decade of accumulated technical debt in the data layer. They need an independent assessment of what they're dealing with before building a credible plan. The assessment is the first-90-days deliverable that buys them credibility with the board and gives their team a clear starting line.
Great fit
- Organizations whose data lives in silos with no single source of truth
- Companies preparing for AI adoption that need to know what's ready and what's not
- PE portfolio companies that need a structured, comparable assessment of data health
- CTOs inheriting legacy data systems who need an independent diagnostic
Not the right fit
- Companies that already have a modern, documented data platform
- Teams that just need a dashboard or report on existing clean data
- Organizations unwilling to address what the assessment finds
- Projects where the entire budget is under $40K