Case Study Agile Minds — Autom8 measured over 7.5 months
Before offering Autom8 to our clients, we deployed it on Agile Minds SRL. Over 7.5 months (August 2025 to March 2026), a single director, 8 departments covered, every task measured. The figures that follow are ours — not marketing projections.
Acceleration
Time saved
Per week
Value/cost
Why this case study?
When an SME asks us what AI can really do for them, the most honest answer often starts with: "let me first show you what we did at our own organisation". Before selling Autom8, we wanted to test it at our own expense, on our own company, with real stakes — not in a demo environment.
The setup is straightforward: Agile Minds SRL, a Belgian IT consultancy, one operational director (Patrick Impens), eight business functions to run (Executive, Sales, Finance, Marketing, HR, IT, Legal, Admin). Measurement period: 1 August 2025 to 15 March 2026, that is 7.5 months. Tools: n8n for orchestration, Claude API as the AI engine, 68 specialised agents spread across the 8 Autom8 modules. Every task produced was timestamped, every deliverable counted.
The figures that follow are therefore measurements, not projections. They come with a clear limitation that we detail further down: they were obtained in our specific context (IT SME, founder familiar with the tools). Your reality will be different; the order of magnitude, however, is being confirmed progressively with our first clients.
The six figures that stand out
Each one measures a different angle of the same reality: a one-person SME holding the workload of five.
Acceleration factor
What a classic team delivers in 13 months, Autom8 delivers in 1 month. For an SME, that means holding an annual plan within a single quarter.
Time saved per task
Average across 23 measured tasks. One hour of human work delivers the equivalent of fourteen hours of value. Not because we work faster — because we delegate the mechanical part.
Deliverables produced
SEO articles, configured agents, legal documents, workflows, procedures, architectures. Over 7.5 months. For an SME, that is the equivalent of two years of classic output.
Departments covered
Executive, Admin, Legal, Sales, Finance, Marketing, HR, IT: no blind spot. For an SME, that means AI is not confined to marketing or support — it structures the whole organisation.
Market-equivalent value
What the same deliverables would have cost at external consultant rates over the period. An honest benchmark for comparison — not a turnover figure.
Real ROI
~EUR 98,000 of value for ~EUR 2,550 of real cost (n8n subscriptions, Claude API, ancillary tools over 7.5 months). For each euro spent, 38.5 euros of useful output.
Source: Agile Minds internal case study, March 2026 — 7.5 months, 23 tasks measured, 8 departments, 365 deliverables counted. Methodology detailed below.
Detail by main domain
We group the 23 measured tasks into three main domains. Time is expressed in person-days (p-d) — a full working day of 8 hours.
Content & Documentation — detail
- 33 SEO articles (~1500 words, 3 languages): 20.6 p-d human to 0.5 p-d Autom8 (-98%)
- 9 pages of T&Cs structured: 10 p-d to 0.8 p-d (-92%)
- 48 agent prompts across 8 departments: 42.5 p-d to 1.5 p-d (-96%)
- 28 pages of architecture (technical and functional): 20 p-d to 1 p-d (-95%)
- Full Business Plan across 3 scenarios: 10 p-d to 0.8 p-d (-92%)
- 24 Startup documents (pitch, one-pagers, memos): 15 p-d to 0.6 p-d (-96%)
Automation & Workflows — detail
- 17 productive n8n workflows: 11.3 p-d human to 1 p-d Autom8 (-91%)
- 9 MCP connectors tested and documented: identical proportional saving
- Data architectures (DuckDB, Metabase, ETL): major reduction of the scoping time
Business Operations — detail
- 5 operational playbooks (onboarding, N1/N2 support, follow-up, escalation): 6.3 p-d to 0.5 p-d (-92%)
- HR procedures (D0 to D90, Belgian social compliance): -92%
- Financial reports (monthly close, budget watchdog): from D+3 to D+1
Source: Agile Minds internal case study, March 2026 — all measurements were made in Odoo 19 and Notion with automatic timestamps.
Before / After, department by department
What the week looked like before Autom8, what it looks like now, and how many hours it frees up.
Source: Agile Minds internal case study, March 2026 — measurements consolidated over the last 4 months of the period (December 2025 to March 2026), when all 8 modules were running at cruising speed.
"I deployed Autom8 on my own organisation before offering it to clients. In 7.5 months, on my own, I produced the equivalent of what a team of 5 people would have taken 2 years to build. Not because I work faster — because 93% of tasks are automated. I take the decisions. The agents execute."
Methodology — and its limitations
Measurements were made in Odoo 19 and Notion with automatic timestamps on every deliverable creation. The "human" times correspond either to tasks actually carried out manually before the switch, or to an external consultant quote obtained for the same task (lower end of the Belgian market in 2025). The "Autom8" times correspond to the real time spent by Patrick to brief the agent, review the deliverable and validate it — we count everything, not just the agent's execution.
The limitations to keep in mind. First, Agile Minds is a tech SME, with a founder familiar with AI tools and used to writing prompts. Second, we designed the agents ourselves — which reduces the learning curve a standard client will go through. Third, some measurements (notably the external consultant comparisons) are market estimates, not actual invoices.
In other words: your context will be different. Your time savings will probably be lower in the first few weeks, while the agents learn your business. The order of magnitude, however, is being confirmed progressively with our first clients. If you want an honest estimate for your case, thirty minutes of conversation are worth more than a projection drawn from ours.
And for your SME, what does it mean?
A few matches between the situations we see most often with our prospects, and what Autom8 actually changes.
Which task of your week would you like to start with?
It is often the best entry point: we pick one irritant, we see whether an Autom8 module handles it, and we tell you honestly if it is worth it — or not.