Buyer guide
How to choose an application modernization company.
Twelve questions that separate a firm that ships from one that decks. Use them on every proposal, including ours.
The short answer
Choose an application modernization company by what it will prove, not what it promises. Ask how it tests behavior, who reviews every change, which tools it uses and why, what it will not do, and who runs the system afterward. Score every proposal against the same 12 questions. We do not publish a ranked list.
Approach
Why is this a checklist and not a ranked list?
A list of the best application modernization companies, written by a company that sells it, is an advertisement. We would rather give you questions you can score a proposal against. The right firm depends on your stack, your size and your risk tolerance, and a ranking cannot know those.
Ace8 appears at the end as one set of answers to the same questions, unranked. Use the questions on us too.
The 12 questions
What should you ask a modernization partner?
Score each answer 0 to 2: no answer, a vague answer, a specific answer you can check. A proposal that scores under 16 of 24 needs more questions.
| # | Question | A good answer sounds like |
|---|---|---|
| 1 | How do you prove behavior is unchanged? | Tests first, run before and after, with results shared |
| 2 | What if we have no tests? | We write characterization tests around the code before we change it |
| 3 | Who reviews each change? | A named senior engineer, not a tool's summary |
| 4 | Which tools do you use, and where do they fail? | Specific tools, with known limits and the failures stated |
| 5 | How is AI used, and who is accountable for its output? | AI drafts, engineers decide and sign |
| 6 | What is the target platform, and could we leave it? | Open targets, no new lock-in |
| 7 | How is the work phased? | Phases that each stand alone, each reversible |
| 8 | How is the price set? | Scope agreed first, then a fixed price for a phase |
| 9 | What does the estimate leave out? | A plain list: licences, environments, your team's time |
| 10 | Who does the work, and where? | Named engineers, location and time zone stated |
| 11 | Who runs it after go-live? | A support plan agreed before cutover |
| 12 | Show a similar project, with numbers. | A real reference, or an honest 'not yet' and a sample run instead |
Warning signs
What are the red flags?
- A price quoted before anyone has looked at the code.
- A timeline with no testing phase.
- "AI does it all" with no named reviewer.
- No discussion of cutover or rollback.
- A new proprietary runtime or platform the firm owns as the target.
- References that cannot be named or checked, or customer counts with no source.
- Staffing that changes after signature, with no named engineers.
Types of firm
What types of firm will you meet?
Each type suits a different job. None is best across the board.
| Type | Tends to suit | Watch for |
|---|---|---|
| Large systems integrators | Very large programs with many workstreams | Junior staffing, long timelines, high rate cards |
| Offshore delivery shops | High volume, well-defined work | Review depth, time zones, thin domain knowledge |
| Cloud vendor professional services | Moves to that vendor's platform | Targets that favor the vendor |
| Tool vendors | Using their product at scale | Services sold to fit the tool |
| Specialist engineering firms | Focused Java or middleware work with senior staff | Smaller capacity, narrower scope |
Proposals
How do you compare proposals fairly?
Ask every firm to price the same sample: one application, the same target, the same assumptions about tests. Then compare the scope lines, not the totals. Two proposals at the same price can cover different work. Check what happens at each phase boundary, and who owns the risk when a phase runs over.
For cost context, see application modernization cost. A short paid assessment is a cheap way to test a firm before a large commitment.
Disclosure
How would Ace8 answer?
Ace8 is the AI-assisted application modernization practice of AceMQ. Our answers, briefly. Tests come first. AI tools do the bulk rewrites and senior engineers review every change; see our method. We publish our tool testing and its limits in the AI tools benchmark. Targets are open. Price follows agreed scope, starting with a fixed-price assessment. The AceMQ engineering team has delivered phased work such as American National's messaging migration, and AceMQ supports the stack after go-live; see supported after migration.
We have not yet written up an application-server or Java-upgrade project with numbers, so on question 12 we show a sample run, not a client reference.
Score each firm against the same questions: how it proves behavior is unchanged, who reviews changes, what tools it uses, how it prices, and who supports the system afterward.
The 12 questions in the table above. Start with testing and review, because they decide the risk.
Not on price alone. Compare scope lines and what each estimate leaves out. A low price with no testing phase moves the cost to you.
It depends on size and risk. Large programs may need an integrator's capacity. Focused Java work often suits a smaller firm with senior engineers.
No. A ranking from a firm that sells the service is not neutral. We publish questions you can use on any proposal, including ours.
Know the cost before anything moves.
A fixed-price assessment gives you the inventory, the risk score and a phased plan. Scope and cost are fixed before the migration starts.
Book a Fixed-Price AssessmentGet in touch
Talk to a modernization engineer
Tell us what you run today. An engineer, not a sales script, replies within one business day.