Not a maturity test. A quantified opportunity diagnostic, written and dated.
The free AI audit from Perello Consulting is an opportunity diagnostic, quantified in hours and euros. You answer a thirty-question questionnaire in ten minutes. We read your answers, cross-reference the hourly cost for your sector with the tasks reasonably automatable in your business, and deliver a written report within forty-eight hours. The document belongs to you. No commercial obligation is attached to it, no automatic follow-up will be triggered. If you decide to entrust one of the identified projects to us, that conversation happens at your initiative. If you prefer to hand it to another provider, the report is detailed enough for them to act on it without consulting us.
What it is — and what it is not
An opportunity diagnostic is a document that quantifies, for your business, how many hours per month are currently consumed on tasks that are reasonably automatable, and what that costs in euros based on the estimated hourly rate for your sector. It is an accounting lens on your operations — dated, signed, actionable the day after delivery. It is not a maturity assessment, and it is not a sales brochure.
The French market for free AI diagnostics sits at two extremes. On one side, institutional self-assessments that are rigorous but generic: France Num, run by the DGE and Bpifrance, offers an AI self-assessment online that points toward public schemes without any personalised financial quantification. On the other, marketing diagnostics that promise guaranteed ROI with no published methodology and no written report ever delivered. In between, maturity audits scored out of a hundred — useful for knowing where you stand, useless for knowing what you are losing.
Our position is elsewhere. To be clear, here is what this diagnostic is not.
- Not a disguised brochure. No section of the report links back to our service catalogue.
- Not a colour-coded score out of a hundred. We do not rank you. We quantify you.
- Not an automatic commercial follow-up. If you do not come back, we do not come back.
- Not a large-account audit at fifty thousand euros. That format exists, elsewhere, with other firms.
- Not a five-minute public self-assessment. France Num covers that need very well, and we refer you there when it is the right level of response.
What you receive instead is a written, quantified document, delivered within forty-eight hours, and yours from the moment you read it.
How it works, step by step
The process runs in four stages, for a total cycle of forty-eight hours between submitting your answers and receiving the report. Each stage is bounded in time and produces a concrete output. Nothing is stretched to create an impression of effort. The report is ready when it is ready — and it is ready quickly.
Stage 1. The questionnaire, ten minutes
You receive a link to a structured thirty-question questionnaire, spread across six sections: company context, operations and processes, data and tooling, team and skills, hours consumed by function, vision and priorities. Most answers use a Likert scale (one to five) for qualitative judgements, and numbers for volumes (headcount, revenue, hours by function). Nothing confidential in the regulatory sense. No financial documents required. No access to your systems.
Stage 2. Reading and analysis, twenty-four to thirty-six hours
Your answers are processed by our internal AI orchestrators, cross-referenced with INSEE statistics on value added per employee for your sector, and compared against France Stratégie research on cognitively automatable tasks. A reasonable automation rate is applied by function. Laurent Perello reads everything before drafting, arbitrates on fragile assumptions, and signs off on the framing.
Stage 3. Writing the report, approximately twelve hours
The report follows a fixed structure: map of automatable tasks by function, hours-per-month quantification, euros-per-month quantification, three projects prioritised by impact crossed with feasibility and team buy-in, budget ranges by project, thirty-sixty-ninety-day action plan. The document runs to approximately ten dense pages. Exported as PDF and as an editable Google Doc.
Stage 4. Delivery and optional debrief, thirty minutes
The report is sent by email, signed and dated. A thirty-minute slot is offered within the two weeks following delivery, at your sole initiative, to discuss it. No commercial pressure during that call: we answer your questions, clarify assumptions if needed, and only produce a quote if you explicitly ask for one.
This pace is sustained because seven AI orchestrators analyse in parallel, and one human arbitrates and signs. Without that architecture, neither the turnaround nor the zero cost would hold.
Get my quantified diagnostic →
What you receive, exactly
The deliverable is a document of approximately ten pages — structured, dated, signed. Exportable as PDF for archiving and as Google Doc for annotation by your teams. Designed to be read alone, shared with your management committee, or handed to a third-party provider without any need to contact us again.
Here is the exact list of what the report contains.
- A map of automatable tasks by function. Sales, marketing, administration, operations, support, management. Each function listed with its identified repetitive tasks.
- Hours-per-month quantification currently consumed on those tasks, derived from your questionnaire answers and INSEE ratios for your sector and size.
- Corresponding euros-per-month quantification, calculated using the published formula: hours per month multiplied by estimated hourly cost, multiplied by the reasonable automation rate. Each component of the formula is explained.
- Three prioritised projects ranked by economic impact, technical feasibility, and likely team buy-in. Not ten projects to choose from. Three, ranked, justified.
- Budget ranges by project, in euros excl. VAT, based on our dated internal portfolio and market benchmarks for AI integration in France.
- A thirty-sixty-ninety-day action plan describing, by project, what can be done in the first month, at the quarter, and at three months.
- The complete report as PDF and editable Google Doc, signed by Laurent Perello and dated on the day of delivery.
An anonymised sample extract, to convey the tone and density of the deliverable:
Sales. Thirty-eight hours per month on lead qualification and follow-up. Estimated cost one thousand five hundred and twenty euros per month. Reasonable automation rate seventy per cent. Target saving twenty-six hours and one thousand and sixty-four euros. Priority project: automated qualification and follow-up workflow. Estimated budget six thousand five hundred to nine thousand euros excl. VAT. Estimated payback seven to nine months.
No superficial slide deck. No decorative charts. A dense, actionable document worth reading even if you never work with us.
Why it is free, and what it actually costs
The audit is free because it is the most honest way to demonstrate what we know how to do. Reading a business, quantifying it, proposing three reasonable projects: that is the first hour of a consultant who knows how to read an operation. Making it free imposes a discipline. We cannot sell empty air, because nothing has been sold yet. We can only be useful — quickly and in writing.
What this costs the practice comes down to an architecture we built for exactly this purpose. A forty-eight-hour orchestrated cycle is not forty-eight human hours. Seven specialised AI orchestrators carry the load of reading, calculating, and formatting; one human arbitrates and signs. The effective human time per audit remains between thirty minutes and two hours, depending on the complexity of the case. This architecture is described in detail on the methodology page, and the team that operates it is presented on the team page.
The marginal cost is low for three reasons. The internal infrastructure already exists and runs for our own needs: VantagePeers, published in open source on npm, coordinates the orchestrators without human intermediary. The three-tier architecture — human, then orchestrators, then agents — eliminates the rebooking staffing model of a traditional firm. The analysis grid is stable: each new audit reuses an already-documented foundation.
A fair comparison with the main French public alternatives. The Bpifrance Diag Data IA deploys an accredited expert for ten days over a three-month cycle, with an estimated residual cost around seven thousand five hundred euros excl. VAT depending on the scheme in force (check current subsidy rates with Bpifrance directly). Large-account audits from firms such as Artefact, Onepoint, or Capgemini run over several weeks and bill between twenty-five thousand and one hundred thousand euros excl. VAT. Our format is not a better version of those: it is a different format, suited to a different type of decision — faster and less committing.
What this proves is simple. You see our method working before you buy anything. If the report is sufficient, you leave with it. If you decide to move to execution, you do so knowing exactly what we produce.
The methodology behind the numbers
Every figure in the report is traceable to a public rule. No black box, no unverifiable proprietary benchmark. The grid is documented, the weightings are named, the hours-to-euros conversion formula is explicit. That is the condition for the report to hold without us.
The reading grid
The thirty diagnostic questions cover six weighted dimensions. Operations and processes: twenty-five per cent, because that is where automation produces the fastest gain. Data and tooling: twenty per cent, because a system can only be automated if its data is readable. Team and skills: twenty per cent, because buy-in conditions execution. Hours consumed by function: twenty per cent, the factual base of the quantification. Vision and priorities: fifteen per cent, because a project must align with your strategy. Each dimension is scored on a one-to-five scale, then weighted.
The quantification formula
It holds in one line: monthly gain in euros equals hours-per-month identified, multiplied by the estimated hourly cost of the role, multiplied by the reasonable automation rate. The hourly cost is derived from value-added-per-employee statistics published by INSEE for your sector and size. The automation rate is calibrated against France Stratégie research on AI's impact on work: by default between forty and seventy per cent depending on the nature of the task.
The compliance framework
Your data is processed in accordance with the CNIL framework on AI and European Regulation 2024/1689, the AI Act. ANSSI recommendations for AI systems are applied across our infrastructure. The CNIL AI sandbox frames our innovation approach and we reference it accordingly. None of your answers feeds a public benchmark; none is used to train a model.
The full architecture underpinning this methodology is described on the methodology page. The French institutional digital framework is set out by France Num and the DGE on entreprises.gouv.fr.
An anonymised example
A typical case, reconstructed from the most frequently observed parameters in our grid. No client named, no precise figure drawn from a real file. The example illustrates the tone and density of the deliverable — not a guaranteed outcome.
Profile. B2B operational consulting services firm, fifteen employees, annual revenue of approximately one million eight hundred thousand euros, regional presence, internal functions covered: sales, marketing, administration, delivery.
Audit result. The map identifies one hundred and forty-two hours per month of reasonably automatable tasks, distributed across sales (fifty-two hours of qualification and follow-up), marketing (thirty-eight hours of reporting and publication), and administration (fifty-two hours of invoicing, chasing, and tracking). Estimated hourly cost of thirty-four euros based on INSEE ratios for this profile. Estimated theoretical monthly saving of approximately four thousand eight hundred euros of redeployable payroll.
Three prioritised projects. Project one: automation of prospecting and email follow-up, budget eight to twelve thousand euros excl. VAT. Project two: automated weekly marketing reporting, budget three to five thousand euros excl. VAT. Project three: invoicing and client follow-up workflow, budget five to seven thousand euros excl. VAT. Total estimated budget between sixteen and twenty-four thousand euros excl. VAT over three to six months.
Estimated first-year return on investment. Projected net gain of approximately forty-five thousand euros after deducting integration and licence costs. Estimated payback between five and eight months.
All these figures are plausible for this profile, not precise for your case. Your own report will settle that on your actual data.
Who this audit is for — and who it is not for
The audit is designed for a specific profile: directors and owners of businesses with five to fifty employees, whose operations, marketing, sales, and back-office functions are visibly consuming time without producing the expected value. The sectors where the format fits best: B2B services, consulting, coaching, communications, agencies, independent retail, training, structured liberal professions.
The non-fit is equally explicit — and we prefer to state it upfront rather than at the end.
- Large accounts with an established in-house data team. The format is too light for that context; audits from Artefact, Capgemini, or Onepoint are built for that need.
- Pure AI research and development projects. We deploy existing systems in a business context; we do not do algorithmic R&D.
- Pre-product startups with no stable processes. There are no repetitive tasks to map yet: the diagnostic would be empty.
- Companies looking for a catch-all audit with no intention to act. The report is dense and actionable; it is less useful to anyone who plans no follow-through.
Honesty about the non-fit is economical — for you and for us. You do not waste ten minutes on a questionnaire that will produce nothing actionable. We do not produce a report that will sit in a drawer. If you are on the edge of the target profile, write to us before filling in the questionnaire: we will tell you frankly whether the exercise will be of use.
Frequently asked questions
Why is the audit free?
Because a precise, dated diagnostic is produced in the first hour of work of a consultant who knows how to read a business. Making it free imposes a discipline: be useful immediately, without commercial detour. If the report is sufficient, you leave with it. If you want us to execute one of the identified projects, that conversation happens at your initiative, not ours. No automatic follow-up, no conditional obligation, no brochure attached to the report.
How long does the audit take?
About ten minutes to complete the questionnaire, then forty-eight business hours for analysis and drafting. An optional thirty-minute debrief is offered afterwards, at your convenience, to discuss the report. This pace is sustained by the practice's architecture: seven AI orchestrators analyse in parallel, one human arbitrates and signs. We do not extend the timeline to simulate effort. The report is ready when it is ready.
What information do I need to provide?
Around thirty answers about your business: sector, headcount, revenue, which functions consume the most time, current tooling level, main perceived barrier to automation, twelve-month priorities. Nothing confidential in the regulatory sense. No detailed accounting documents. No access to your information systems. Your answers are sufficient to produce a defensible estimate. If your case warrants more precision, the report will signal that explicitly and you will decide what to do next.
Who sees my data?
Your answers are processed by our internal orchestrators and read by Laurent Perello before the report is sent. They are not shared with any third party, do not feed a public benchmark, and are not used to train an AI model. They are retained in our infrastructure for the duration of the relationship, or until you request deletion. The processing framework follows the GDPR, CNIL recommendations, and the requirements of Regulation 2024/1689 on AI.
What does the report contain?
A document of approximately ten pages structured as follows: map of automatable tasks by function, hours-per-month quantification, corresponding euros-per-month quantification, three projects prioritised by impact and feasibility, budget ranges, thirty-sixty-ninety-day plan. Exported as PDF and editable Google Doc, signed and dated by Laurent Perello. It is a standalone deliverable: you can read it alone, share it with your management committee, or hand it to another provider without consulting us.
Am I obliged to buy something afterwards?
No. The report belongs to you. No clause obligates you to continue. No automatic commercial follow-up will be triggered. If you want us to execute one of the identified projects, you simply ask. If you prefer to hand that project to another provider, the report is detailed enough for them to take it on. If you decide to do nothing, you still have a reading of your business that you did not have before.
How are your figures calculated?
The recoverable hours come from your answers, interpreted through a public grid with named weightings (operations twenty-five, data twenty, team twenty, hours twenty, vision fifteen). We convert them to euros using the formula hours multiplied by hourly cost multiplied by automation rate, where the hourly cost is derived from INSEE value-added-per-employee statistics and the rate is calibrated against France Stratégie research. Budget ranges come from our dated internal portfolio, consultable at perfectaiagent.xyz.
What happens after the report is delivered?
Nothing automatic. If you wish to discuss it, a thirty-minute slot is available for the two weeks following delivery. If you do not request it, we do not press. If you decide to execute a project with us, a quote is produced from the elements already framed in the report. If you decide to execute with another provider, the report is designed for that. If you do nothing, you keep the document.
Request your audit
You complete the questionnaire in ten minutes. We deliver a written, quantified report within forty-eight business hours. The document belongs to you, signed and dated. A debrief slot is offered within two weeks, at your sole initiative. No price to negotiate, no prior discovery call, no brochure to read. The full methodology underpinning this diagnostic is published on the methodology page. The team that operates it is presented on the team page. We commit to responding within twenty-four business hours of receiving your answers to confirm the delivery slot.
Secondary links: Read the methodology · Meet the team · Public journal
Orchestrator: Alpha — Perello Consulting | 2026-04-16