Skip to content
Astorna

For technology and data

Has anyone tested whether the AI proposal is buildable?

An AI proposal can reach technology with the model and expected benefit already decided. The questions that determine feasibility often arrive later.

Where does the data live, and is it reliable enough for the job? What will integration require? How will output quality be measured against a meaningful baseline? What happens when the model is wrong, unavailable, or changed by its provider?

Who detects the problem, who owns the fallback, and who supports the system after launch? Have those questions been answered, or simply assigned to you? Until they are settled, the business case rests on assumptions about what your team can make work.

First check if this workshop is suitable for you.

Book a free 15-minute call

What would need to be true before anyone called it buildable?

Before committing engineering time, could you answer:

  • Can the required data be accessed reliably and used for this purpose?
  • Is the data complete and consistent enough, and who owns its quality?
  • What baseline or ground truth would show whether the output is good enough?
  • How would uncertain or wrong results be detected, routed, and explained?
  • What would need to change in existing systems, interfaces, permissions, and controls?
  • Could the system meet the required volume, speed, and cost in production?
  • What happens when the model changes or its provider is unavailable?
  • Who monitors performance, handles incidents, and supports the workflow after launch?

What this workshop helps you decide

Whether the proposal has a credible path to production, which assumptions need to be tested first, and whether the smallest useful test is worth your team’s time.

In a private ninety-minute workshop, we establish the AI concepts and control questions relevant to the proposal. Then we examine one workflow and the proposal around it: the intended use, data path, integration points, evaluation approach, failure handling, and production ownership.

Within two working days, you receive a one-page AI Feasibility Brief. It records the technical assumptions, unresolved dependencies, evaluation approach, failure and fallback requirements, ownership questions, and the smallest sensible test.

The workshop does not turn the proposal into an implementation design. It identifies what would need to be tested before the proposal becomes a production commitment.

What a buildable proposal would look like

The real question is not whether the model can produce a convincing output. It is whether the surrounding system can make that output usable, measurable, secure, and supportable.

You should be able to explain where the data comes from, how quality will be measured, what integration and controls are required, how failures will be handled, and who will own the system after launch.

Sometimes that leads to a narrow technical test. Sometimes it changes the proposal. Sometimes the production burden gives the firm a reason to leave the workflow alone.

Who this is for

  • Technology and data leaders responsible for assessing feasibility or production ownership
  • Teams handed a proposal or demonstration with unresolved questions about data, integration, evaluation, or support
  • Firms that want to test the assumptions before committing roadmap capacity or budget, or choosing a vendor
  • Someone who can bring one real workflow and explain how it operates today

How it works

  1. Start with a free 15-minute call

    Tell me briefly which workflow you have in mind. The call is simply to see whether the workshop is likely to be useful. It may not be.

    Book a free 15-minute call

    Prefer email? Write to hossein@initialblock.com.

  2. Book the workshop

    If the workshop fits, you will receive a private booking link to choose a date and time that suits you, and a few questions that help both of us prepare.

    Price
    €400
    paid when booking
    Workshop
    90 minutes
    Online, at your office, or at ours. You choose.

  3. Use the brief to test feasibility

    Within two working days, you receive a one-page AI Feasibility Brief. It records the preliminary assessment, what remains unresolved, and the smallest sensible next step.

    On whether the proposal has a credible path to production, the answer may be yes, no, or not yet.

What would you want the person testing the technical assumptions to understand?

AI, certainly. But also what happens between a model producing an output and a financial firm being able to rely on it: data access, integrations, controls, monitoring, failure handling, and support.

I have spent more than fifteen years building technology. For the past nine, I have been co-founder of Xeco Labs, where we build operational infrastructure for financial firms. I also have a master’s degree in Artificial Intelligence from the University of Amsterdam.

I have had to turn business and regulatory needs into production systems, then stay with those systems as clients grew and workflows changed. That is why I look at both sides: what the model can do, and what the surrounding system would need to make that useful, measurable, and supportable.

Some of the firms I have worked with through Xeco Labs

Index People, AssetCare, Carbon Equity, Bits of Stock, OAKK Capital Partners, RegLab.

If that sounds like the perspective you would want in the room, start with the free call.

Fifteen minutes to find out whether the workshop is likely to be useful for your workflow.

Book a free 15-minute call