Legacy System Assessment

When a prospect has an existing system — particularly an old or undocumented one — Telos can run a first-pass assessment to establish feasibility: what it would take to rebuild, re-platform, or support the system. This is not a full blueprint engagement; it is a scoped, low-commitment exercise designed to answer the question: is this worth pursuing, and at roughly what scale?

When to use this

• A prospect is evaluating whether to rebuild a legacy system
• No documentation exists and the system's complexity is unknown
• The prospect needs something concrete to present to a board or leadership team before committing to a full engagement

The assessment process

1. NDA and code access

Telos signs a non-disclosure agreement and receives a copy of the source code. The code is used solely for analysis purposes — no reuse, no retention beyond the engagement.

2. Software walkthrough

Before any code analysis begins, Telos asks for a walkthrough of the live system. This can take 1–4 hours depending on complexity. The walkthrough gives the AI a frame of reference — the language of the system, its information architecture, and the main user flows. It materially improves the quality of subsequent analysis.

3. First-pass code analysis

Telos runs a targeted analysis — not a full deep-dive, but enough to understand the breadth of the system. The output is a set of scope tickets: discrete, sized units of work representing the modules, functions, or domain areas identified in the code.

4. Subject matter interview

Where possible, Telos interviews someone with deep knowledge of the system — the original developer, a long-tenured support engineer, or a key business user. Code analysis surfaces what the system does; interviews surface why, and where the hidden complexity lives.

5. Feasibility output

The assessment produces a plain-language summary covering:

• Breadth of scope (number and nature of modules identified)
• Complexity indicators (language, architecture, data model legibility)
• Indicative time and cost range for a rebuild or re-platform
• Recommended next steps

This output is suitable for internal use; it gives leadership a concrete basis for a go/no-go decision on a full engagement.

Notes on legacy languages

AI tools perform best with modern, widely-used languages. Older or proprietary languages (e.g. Progress, COBOL, RPG) will be less familiar to the models. In these cases:

• Analysis scripts are tuned in advance to guide the AI on what to look for
• Human subject matter experts become more important as a complement to code analysis
• The first-pass may be more limited in depth, and that should be communicated upfront

What this is not

This is not a project plan, a fixed-price quote, or a commitment to build. It is a feasibility signal — enough information to decide whether a full blueprinting engagement makes sense.