The 90-Day Illusion: When AI Strategy Is Just a Sales Funnel With Better Design
A well-designed PDF landed on my desk recently. Fourteen pages. Clean layout. Professional graphics. Bold statistics in large fonts. The title promised a 90-day plan to integrate artificial intelligence into a business and "transform processes into results."
I read it carefully. Twice. And then I read it a third time, because I wanted to be sure I wasn't being unfair.
I was not being unfair. The document is a sales funnel disguised as a strategy guide. And it's far from the only one circulating.
The Pattern
Over the past few months, a specific type of document has proliferated across the European SME landscape. The format is remarkably consistent regardless of who publishes it:
A telecommunications company. A technology reseller. A consulting firm. A cloud provider. The branding changes. The structure doesn't.
Page one establishes urgency. A large, alarming statistic — typically about wasted time or lost productivity — creates the emotional hook. The source is usually a study commissioned or conducted by the same vendor whose product appears on page four.
Pages two and three present a structured plan. Three phases, three cycles, three steps — the number varies but the rhythm is always the same. The first phase is invariably called something like "diagnose" or "assess." This creates the impression that the process begins with understanding the problem.
Page four reveals the solution. It is always the vendor's product. The same vendor whose logo appears on the cover. The same vendor who published the alarming statistic on page one. The same vendor whose "case studies" appear on pages eight through twelve.
The remaining pages provide statistics about the solution's effectiveness. The statistics are sourced from the vendor, from studies commissioned by the vendor, or from analyst reports sponsored by the vendor.
The document ends with a call to action. Contact us. Start your transformation. The future is now.
This pattern is worth examining — not because any individual document is harmful, but because the cumulative effect on how organisations make technology decisions is significant.
The Diagnosis That Isn't
The most revealing element is always the "diagnostic" phase. In the document I reviewed, the diagnostic cycle was allocated 30 days. Within those 30 days, the organisation was expected to conduct a productivity audit across five departments in one week, appoint internal champions for the pre-selected tool, activate the tool's basic features, and schedule training sessions on how to use the tool.
Read that sequence again. The "diagnosis" lasts one week. The remaining three weeks of the "diagnostic phase" are spent implementing the solution that was supposedly yet to be determined.
This is not diagnosis. This is a purchasing decision with a diagnostic preface. The conclusion was written before the assessment began. The "plan" doesn't help an organisation understand what it needs — it helps the vendor's product appear to be what the organisation needs.
In medical terms, this is the equivalent of a pharmaceutical company publishing a "health assessment guide" where step one is "identify your symptoms" and step two is "take our medication." The diagnosis exists to legitimise the prescription, not to inform it.
The Statistics That Prove Themselves
The documents invariably contain impressive numbers. Productivity gains of 40%. Time savings of 30 minutes per day per employee. Return on investment exceeding 100%. Payback periods of less than a year.
These numbers deserve scrutiny — not because they're necessarily false, but because of where they come from.
In the document I reviewed, every statistic was sourced from one of three places: the vendor's own internal studies, research firms whose studies were commissioned and funded by the vendor, or surveys of the vendor's own customers conducted by the vendor.
This creates a closed evidentiary loop. The vendor funds the research. The research validates the vendor's product. The vendor publishes the research as independent proof. The customer reads the "independent proof" and buys the product. The vendor surveys the customer and produces more "proof."
At no point does an independent voice enter the process to ask: compared to what? Compared to doing nothing? Compared to a different tool? Compared to fixing the underlying process problems that no tool can solve?
The absence of a control group is telling. When a document claims that a tool saved 30 minutes per day per employee, the implicit comparison is "versus not having the tool." But the relevant comparison for a decision-maker is "versus all other ways we could invest the same money to achieve the same outcome." That comparison never appears — because it might not favour the product being sold.
The Case Study With No Case
Perhaps the most instructive element is what passes for evidence.
The document I reviewed contained a "case study" described as follows: a large multinational in the beverage sector confirmed that teams work faster, with higher quality, and less effort. The company had over 5,000 employees and reported daily time savings of 16 to 30 minutes per employee.
That's the entire case study. No company name. No specific process described. No before-and-after metrics. No implementation challenges. No cost figures. No timeline. No mention of what didn't work. No independent verification.
This isn't a case study. It's a testimonial without attribution. A real case study would include: what specific problem the organisation was trying to solve, what they measured before and after, what the total cost of implementation was (including training, adoption time, and productivity dip during transition), what problems they encountered, and what the verified net result was after a meaningful period.
The reason these details are absent is not carelessness. It's strategy. A vague success story is more useful to a vendor than a detailed one, because details invite scrutiny and scrutiny reveals complexity. "Teams work faster" is unfalsifiable. "We reduced invoice processing time from 4.2 hours to 2.8 hours by automating three specific steps, at a cost of €47,000 over six months, with a four-week productivity dip during adoption" is falsifiable — and therefore risky to publish.
The Cost That Doesn't Appear
Every document in this genre shares a conspicuous absence: the actual cost of what's being proposed.
The document I reviewed mentioned gains, savings, ROI percentages, and time recovered. It did not mention — anywhere, on any of its fourteen pages — what the solution costs. Not per user. Not per month. Not as a total investment. Not as a comparison to alternatives.
This isn't an oversight. Enterprise AI tools typically cost €25-35 per user per month. For a company with 100 employees, that's €30,000-42,000 per year in licensing alone — before training, before integration, before the consulting engagement to "help you get started," and before the productivity dip that occurs during any technology transition.
The guide also doesn't mention that the AI tool in question requires a specific tier of the vendor's productivity suite as a prerequisite — an additional cost that many SMEs may not currently be paying.
And it doesn't mention alternatives. Not because alternatives don't exist, but because the document's purpose is not to help the organisation make the best decision. It's to help the organisation make a specific decision — the one that generates revenue for the company whose logo is on the cover.
Why This Matters
One might argue that vendor marketing has always worked this way. Brochures have always been sales tools. This isn't new.
What's different is the packaging. These documents don't present themselves as brochures. They present themselves as strategic guides. They use the language of consulting — "diagnose," "assess," "transform," "scale." They mimic the structure of independent analysis. They create the impression that the reader is being helped to think through a decision, when in reality the decision has been made for them.
For large enterprises with procurement departments, technical evaluation teams, and competitive bidding processes, this type of content is relatively harmless — it enters a pipeline where it will be compared, challenged, and negotiated.
For SMEs — which is the explicit target audience — the impact is different. A small company with 50 employees, no CTO, and limited technology expertise receives a professionally designed "strategic guide" from a trusted brand. The guide appears to offer a structured approach. The statistics appear credible. The case study appears to validate the approach. The cost is invisible. The alternatives are absent.
The decision that follows is not an informed technology choice. It's a marketing conversion.
What Would Genuine AI Strategy Look Like?
If these documents are what AI strategy looks like when sold by the vendor, what does it look like when it's independent?
Start with the problem, not the tool.
Before any technology decision, identify the specific operational problems that need solving. Not "we need AI" — that's a solution looking for a problem. "Our invoice processing takes 4 hours because three people handle it manually and the data comes from four different formats" — that's a problem that might or might not benefit from AI. The diagnosis determines the prescription, not the other way around.
Demand independent evidence.
When a vendor presents ROI figures, ask who funded the study. When a "case study" has no company name and no specific metrics, treat it as what it is: marketing copy. Look for independent analyst reports that compare multiple solutions. Look for case studies with named companies, specific numbers, and acknowledged challenges. If the evidence only comes from the vendor, it's not evidence — it's advertising.
Calculate the real cost.
Licensing per user per month. Prerequisite subscriptions. Training time. Productivity dip during adoption. Internal IT support. Integration with existing systems. Ongoing administration. And the opportunity cost: what else could you do with the same budget? A €40,000 annual investment in AI licensing might generate less value than hiring one additional person, redesigning a broken process, or investing in training for existing tools that are underutilised.
Consider doing nothing.
The most underrated option in any technology decision is the status quo. Not because change is bad, but because the cost of the wrong change often exceeds the cost of no change. If the current process works — imperfectly but reliably — the burden of proof is on the new solution to demonstrate that it will be better, not just different. "Everyone else is doing it" is not a business case. It's peer pressure with a budget.
Separate the diagnostic from the prescription.
The person or company that assesses your needs should not be the same person or company that sells you the solution. This is a basic principle of independent advice that applies to medicine, law, engineering, and finance. It should also apply to technology. If the "consultant" recommending AI is the same entity that earns revenue from AI licenses, the recommendation is structurally compromised — regardless of the consultant's intentions.
The Bottom Line
The 90-day AI integration guide is not a plan. It's a purchase order dressed in strategic language.
This doesn't mean AI tools aren't valuable. Many of them are genuinely useful for specific, well-defined tasks. It doesn't mean vendors are dishonest — they're doing what vendors do: selling products.
What it means is that the decision to adopt AI in an organisation deserves better than a 14-page brochure where the conclusion was written before page one. It deserves genuine diagnosis, independent evidence, honest cost analysis, and the freedom to conclude that the best decision might be a different tool, a different approach, or no tool at all.
The organisations that will benefit most from AI are not the ones that adopted fastest. They're the ones that understood what they needed before they bought anything.
A 90-day plan sounds decisive. But a decision made in 90 days based on a vendor's brochure is not decisiveness. It's velocity without direction.
And in business, as in medicine, the most expensive treatment is the one prescribed without a proper diagnosis.