How Coformaco uses A.I. to prepare files, reporting and dealflow smarter

Project Overview

Who is Coformaco?

Coformaco is the family office of Conny Vandendriessche, co-founder of Accent Jobs and House of HR. From Roeselare, the team invests in entrepreneurs and companies, under the tagline "Funding Dreams." It's a small, experienced team where almost everyone works as a one-person team. There are no classic departments — each team member handles their own files, from dealflow and due diligence to financial reporting and board work.

That's exactly where the pressure showed up. Information on holdings is scattered across emails, notes, Dropbox, accounting software and reports. Preparing for a first meeting or a board meeting meant a lot of manual digging. Coformaco wanted to know where A.I. could genuinely lighten that load.

The challenge

Conny Vandendriessche and Bo Decloedt came to Flowkify with a two-part question. They wanted to bring the team along on A.I., but also wanted to know where the biggest wins were.

"How do we use A.I. across our family office and our holdings, so we sit down better prepared and faster, without losing control?"

There was also doubt about which tools to use. ChatGPT, Copilot, Gemini or Claude? The team works in Microsoft 365, but also uses Apple devices and Dropbox. The choice had to fit how they actually work.

Project Delivery

The approach

We started with an A.I. Opportunity Mapping — a workshop where we map out every repetitive and manual task together with the team, then prioritize the A.I. opportunities by value and complexity.

The workshop took place in Roeselare, in-house, with nine participants from across the team. Since Coformaco doesn't work with classic departments, we split into clusters instead: personal efficiency, IT, financial reporting, meeting preparation, and follow-up. Per cluster, we brainstormed pain points, existing initiatives and A.I. opportunities. Afterwards, the groups pitched their concepts and we plotted everything on a value-versus-complexity matrix.

Beforehand, each team member filled in an intake form with their workflows and bottlenecks, so no workshop time got lost warming up.

1 · A.I. Opportunity Mapping: five top opportunities

The workshop surfaced five top opportunities. They center on a shared data layer, automated financial reporting, and preparation tools for board meetings and first meetings with a target. Each concept got a goal, expected KPIs, boundary conditions and a defined first step.

We also identified three quick wins: an A.I. notetaker, a mail agent and a coaching format. Plus a backlog per cluster, so the team knows what's still on the shelf after the project wraps up.

The benefit: the team no longer chooses by gut feeling, but from a shared, prioritized list.

2 · Training part 1: the foundation

The first session focused on daily work. Half a day, in-house, with exercises on the team's own files.

• The framework: what A.I. is and isn't, and why your own judgment still matters.
• The Copilot and ChatGPT dashboards: choosing models, uploading documents, finding past chats back.
• Prompting: the seven-step model for role, goal, context, audience, parameters, tone of voice and extra context.
• Custom instructions: B1 reading level, no emojis, informal tone, accounting for Belgian law, never making things up when an answer is unknown.
• CustomGPTs and Copilot Agents: the difference between a brainstorm partner and a pre-programmed agent with a fixed output format.
• Working with projects and folders, so context per holding stays intact.

The benefit: everyone left with working instructions and their own way to structure files.

3 · Training part 2: from user to builder

A few weeks passed between session 1 and 2, so the team could first get hands-on themselves. The second session built on that, with Claude as the central tool.

• Context engineering: building a library of reference files, such as ICP, tone of voice, writing frameworks, example outputs and process descriptions.
• Claude Skills: the problem that A.I. starts from scratch every conversation, and how to fix that once.
• Seven concrete skills built specifically for Coformaco.
• Plugins and connectors: Excel, PowerPoint and MCP connections to existing systems.
• A.I. tools in practice: Leexi as a notetaker, Gamma for presentations, Lovable for quick prototypes.

The seven skills tie directly into the opportunity mapping:

1. prospect-one-pager: pulls together website, NBB, LinkedIn, emails and notes into a structured one-pager with red flags and investment criteria.
2. meeting-voorbereiding-eerste-gesprek: builds a meeting template with conversation structure and questions per phase.
3. rvb-rapport-samenvatting: fills in the Board Master Template with value drivers, improvement points and deal status.
4. exit-value-analysis: calculates an indicative exit value using Coformaco's own methodology.
5. kwartaalrapport-participatie: fills in the standard reporting format, including P&L, cash flow and commentary.
6. red-flag-analysis: flags falling margins, rising debt and deviations from budget.
7. document-analyse-lang: pulls the points that matter to an investor out of an annual report, due diligence report or contract.

Every skill carries the same hard rule: never make up numbers. Whatever's missing gets flagged as "information not available, manual follow-up required."

The benefit: the team now builds new skills on its own, without needing an outside party each time.

Results

In one project, Coformaco moved from "we want to do something with A.I." to a concrete, shared plan. The opportunity mapping delivered five fully worked-out top opportunities, three quick wins, and a backlog per cluster, each with a defined first step. The two trainings brought nine team members up to the same level, from prompting to building their own skills.

The biggest change shows up in preparation. Where a first meeting or a board meeting used to start with digging through emails, Dropbox and Excel, the team now starts from a structured one-pager or a filled-in board template. Meeting write-ups were also flagged as a quick win, so knowledge from conversations lands in the file instead of scattered notes.

Flowkify keeps supporting Coformaco as its A.I. partner, with a focus on the data layer and further building out the chosen top opportunities.

Curious what A.I. could mean for your family office or investment holding? Get in touch with Flowkify for a tailored A.I. scan.