Financial AI Analytics Implementation Forum: Practitioners Sharing What Works
What the programme covers and how it is structured
The gap between pilots and production
A lot of teams have run a predictive analytics pilot. Far fewer have moved one successfully into daily financial operations.
This forum exists specifically for that second group, and for teams trying to get there. The conversations are grounded in what actually happened, not what the vendor deck promised.
Format and atmosphere
Sessions alternate between short practitioner talks and open discussion. Speakers are working professionals, not platform advocates.
Orlaith Fennelly, a credit risk analyst who spent eighteen months integrating a forecasting layer into a mid-size lending operation, opens the forum with a frank account of what she would do differently.
Topics the community wanted covered
Attendees from previous forums shaped this agenda through a shared input process. The topics that came up most consistently were model explainability for non-technical stakeholders, handling missing financial data without corrupting model outputs, and monitoring model performance after deployment.
All three get dedicated time here.
Who this forum fits
People who find value here tend to be hands-on: they write queries, review model outputs, or sit in the meetings where implementation decisions get made.
Executives looking for a high-level overview of AI in finance will find the sessions too granular. That is intentional.
Shared notes and follow-up
A collaborative document capturing key discussion points gets distributed to all attendees within 48 hours. It reflects what was actually said, not a sanitised summary.
Facilitator Declan Ostrowski hosts an optional online follow-up session three weeks later for anyone still working through implementation questions.
Stage-by-stage breakdown
Forum Programme
-
Opening Talk: One Implementation, Honestly Described
- Speaker
- Orlaith Fennelly, Credit Risk Analyst
- What she covers
- The eighteen-month arc of integrating a cash flow prediction model into a live lending workflow, including two significant rollbacks and what caused them.
- Duration
- 35 minutes plus questions
-
Working Session: Explainability for Financial Stakeholders
- Format
- Small group discussion with a structured prompt. Groups share back to the room.
- Focus
- How to communicate model outputs to credit committees, CFOs, and auditors without oversimplifying or losing accuracy.
- Duration
- 45 minutes
Preparation note
Bring one real example of a model output you have had to explain to a non-technical audience. Anonymised is fine.
-
Practitioner Panel: Monitoring After Go-Live
- Panellists
- Three practitioners from banking, insurance, and fintech backgrounds
- Topics
- Drift detection approaches, retraining schedules, and what metrics actually matter in financial prediction contexts
- Duration
- 50 minutes
-
Open Floor: Unresolved Problems
Attendees bring their current blockers. The room works on them together. No prepared slides, no product pitches.
This session consistently runs over time because the problems people bring are specific and the discussion is genuinely useful.
-
Close and Follow-Up Logistics
- Notes distribution
- Within 48 hours by email
- Follow-up session
- Online, three weeks post-forum, hosted by Declan Ostrowski