Part I — The Enterprise Foundation

Finding the Signal in the Sprawl

Part I has covered the IT landscape: the three layers, legacy systems, cloud and SaaS sprawl, data readiness, integration patterns, technical debt, and the coexistence imperative. The practitioner who has absorbed this foundation now faces the question that actually matters: given all of this complexity, where do you start?

The answer is Signal Capture — the first pillar of the LegacyForward framework. Signal Capture is the discipline of finding the real problem worth solving before committing to the solution. In the context of enterprise IT sprawl, it means identifying the AI use case that has the clearest business case, the most accessible data, the lowest integration risk, and the most measurable outcome.

What You Will Learn

• The Signal Capture framework applied to enterprise AI initiative selection
• The five criteria that identify high-signal AI opportunities in complex IT environments
• How to distinguish signal (genuine value opportunity) from noise (technically interesting but strategically marginal)
• The one-page AI opportunity brief that focuses initiative scoping

10.1 What Signal Means

Signal, in the LegacyForward sense, is the combination of three things: a real problem that matters to the business, data that exists and is accessible, and a user who will act on the AI's output. Remove any one of the three and the initiative fails — either because the problem does not justify the investment, because the data cannot support the AI, or because nobody uses the output.

Most organizations have more AI ideas than they can fund and more data than they can govern. The role of Signal Capture is to identify which ideas have all three components — problem, data, and user — and to rule out ideas that are missing any of them before investing in scoping, much less implementation.

10.2 Five Criteria for a High-Signal AI Opportunity

Criterion 1: The problem is specific and measurable. "Improve customer satisfaction" is not a specific problem. "Reduce time-to-first-response on tier-2 support tickets from 48 hours to 8 hours" is specific and measurable. Specific problems can be scoped. Measurable outcomes can be used to determine whether the AI initiative succeeded.

Criterion 2: The data exists and is accessible. Using the data readiness framework from Chapter 5, the data required to train and operate the AI is accessible (not locked in an inaccessible legacy system), of sufficient quality (not requiring months of cleanup before use), and governable (the organization has the right to use it for this purpose).

Criterion 3: The integration path is tractable. Using the integration patterns from Chapter 6, the data access and write-back integrations required are achievable with the organization's current integration infrastructure. The integration work is sized and on the project plan.

Criterion 4: A user exists who will change their behavior. The AI's output will be acted on by a specific person or system. That person or system has been identified, consulted, and has confirmed that the output would change their behavior or decisions. AI that produces outputs nobody acts on is a research project, not a product.

Criterion 5: The value is worth the cost. The value of the problem being solved — in time saved, errors reduced, revenue generated, risk mitigated — justifies the cost of building and operating the AI. This does not require a precise ROI calculation, but it does require a honest comparison of expected value and expected cost.

10.3 The AI Opportunity Brief

Before scoping any AI initiative, document the opportunity in a one-page brief:

Problem statement — what is the specific problem? Who experiences it? What does it cost the organization?

Current state — how is the problem currently handled? What does the manual process look like? What are its limitations?

Proposed AI capability — what would the AI do? What inputs would it consume? What outputs would it produce?

Data inventory — what data is required? Where does it live? Is it accessible? Is it of sufficient quality?

Integration requirements — what systems does the AI need to read from? Write to? What are the integration mechanisms?

Success metrics — how will success be measured? What is the baseline? What is the target?

Risks — what are the top three risks? What is the mitigation for each?

This brief is the basis for the go/no-go decision on proceeding to full scoping. It surfaces the five criteria in a format that business stakeholders, architects, and data practitioners can all engage with.

10.4 From Signal to Strategy

The Signal Capture process applied to a portfolio of AI ideas produces a priority stack: the ideas with clear signal at the top, the ideas missing one or more components below, and the ideas that are technically interesting but strategically marginal at the bottom.

The priority stack is not a final roadmap. It is the input to the strategy work in Part II — where investment decisions, organizational readiness, and portfolio prioritization shape which signals get acted on and in what order. Part I has given you the map. Part II gives you the compass.