EngagementsIT services

Find out what Gen AI changes about mainframe modernisation

For practice heads and delivery leads whose standard modernisation plan and estimating rules pre-date generative-AI tooling.

Engagement type
Solo Pod
Experts
1 lead
Hours
~80
Duration
5 weeks

What the client needed

A provider's standard mainframe modernisation plan was written before Gen AI tooling existed. It needed to know, honestly, where the tooling cuts effort — code understanding, test generation, documentation, data migration — and where it does not, and to revise its estimating rules and its customer-facing position accordingly.

The answer had to come from someone who has run modernisation programmes on COBOL and PL/I estates and used the tooling in anger.

The brief

The questions on the table

As the client wrote them, lightly shortened. Client and respondents are not named.

  1. Q1

    Phase by phase — assessment, code understanding, re-platform or re-architect, data, test, cut-over — where does Gen AI tooling reduce effort on COBOL and PL/I estates, by how much, and where does it not?

  2. Q2

    Which of our estimating rules are now wrong, and what should replace them?

  3. Q3

    What can we say to customers about AI-assisted modernisation that we can also deliver?

The hours

What the work involved

  1. 01

    Walked the current plan

    The provider's phase plan, effort factors and the assumptions behind each, with a view of the last three programmes it had run against the plan.

  2. 02

    Assessed each phase against the tooling in practice

    For every phase, what the tooling actually does on real COBOL and PL/I code — where it reads, where it generates, where its output still needs a senior engineer line by line — drawn from programmes the expert had led. Effort factors revised per phase, with the basis written next to each.

  3. 03

    Rewrote the phase plan

    The same phases, re-sequenced where AI-assisted code understanding moves work earlier, with new entry and exit criteria and the roles that change.

  4. 04

    Drafted the customer-facing position

    Two pages the provider could put in front of its own customers: what AI-assisted modernisation does, what it does not, and how the provider prices and guarantees it.

  5. 05

    Presented to practice leadership

    A 60-minute session with the practice leadership, questions taken, and a short list of follow-ups.

Deliverables

What the client received

  • A revised programme planPhases, sequence, entry and exit criteria, roles.
  • Updated estimating rulesEffort factors by phase, each with its basis and the cases where it does not apply.
  • A customer-facing positionTwo pages, ready to use.
  • A leadership presentationWith the follow-up list.

The terms

What an engagement like this costs

You buy hours in blocks and can change the expert at the end of any block, at no extra cost. Longer projects run in stages: pause after a block, pick up the next stage when you are ready.

Pricing in full →

Rate band
$65–$250 /hr
Senior specialists can be up to $500
Hour blocks
1 · 2 · 5 · 10 · 20 · 50 · custom
This engagement
Solo Pod · ~80 hours · 5 weeks

More in IT services

Re-planning a modernisation practice?

Describe what you need

For example: “Our mainframe modernisation estimates pre-date Gen AI tooling and we need them re-based before the next bid.”

We aim to come back with matched experts within 24 to 72 hours.