Most guarantees in consulting are theater. They sound bold in the pitch, then quietly dissolve into fine print about "effort," "scope," and "factors outside our control." The recovery guarantee is the opposite of that. It is a number, agreed before work starts, that we are on the hook to hit.
What is the recovery guarantee?
The recovery guarantee is a promise that the specific revenue leak we identify in your free audit will close, or kratt keeps working at no extra cost until it does. The audit sets a measurable dollar target first. We build and run the system. If the agreed leak does not close, we work free. There is no lock-in.
That definition matters because every word carries weight. "Specific" means one named leak, not a vague vibe about "growth." "Measurable dollar target" means a number both sides sign off on. "Work free" means our time, not yours. And "no lock-in" means you can leave whenever you want. Read on for the mechanics.
Why can a two-person consultancy promise this when agencies will not?
Because we built the model around it. kratt is an audit-first AI consultancy, not an agency. Oskar builds the systems. Karl-Kristjan runs the audit and owns the outcome. The free audit gives us a defensible number before we touch your stack, and our done-for-you delivery means we control the variables that move that number.
An agency cannot do this for a structural reason. The retainer model pays the agency whether or not the work moves revenue. Industry analysts note that retainers "lack strong performance incentives, since payment is not directly tied to specific outcomes." When the invoice clears on the first of the month regardless of results, why would anyone bet their fee on an outcome they do not control? They would not, and most do not.
How does the audit set the target number?
The guarantee is only as honest as the number behind it. So the audit comes first, and it is free. We map where your business loses money between a lead arriving and cash landing. We call these open loops. A missed call at a service business. A quote that never gets followed up. A booking form that breaks on mobile. Each one has a dollar cost we can estimate from your own data.
Then we rank every leak by annual dollar impact. The largest, most fixable leak becomes the target. If your front desk misses 40 calls a week and each booked job is worth 280 dollars, the math writes itself. That ranked number is the figure the recovery guarantee attaches to. You can run a rough version yourself with our open loop tax calculator before you ever talk to us.
What counts as a measurable leak?
A leak counts when three things are true. It has a clear before-state we can measure today. It has a clear cause we can fix with a system. And it has a dollar value both sides agree on. If we cannot measure it, we will not guarantee it. We explain the ranking method in our closed-loop score framework and in the deeper piece on the open loop tax.
Why do we rank leaks instead of fixing everything?
Because attention is finite and so is your patience. Fixing five small leaks at once usually means fixing none of them well. We pick the largest fixable leak, close it, prove the number moved, then move to the next one. That sequence is what makes the guarantee testable. One target, one outcome, one verdict. If we tried to guarantee a dozen things at once, the promise would blur into the same fog every agency hides behind.
What happens after we agree the number?
We build the fix. Done-for-you means we design, host, and run the system ourselves. You do not hire a team. You do not learn a tool. For a missed-call leak that usually means a voice agent that answers every call and books the job, which is the kind of work covered by our voice agents service. For a broken follow-up loop it means an automation pipeline, covered under automation.
We control delivery on purpose. When you own the stack and we only advise, the outcome depends on whether your team executes. We cannot guarantee that. When we build, host, and operate the system, the outcome depends on us. That is the only honest condition under which a results guarantee can exist. We wrote about why advice-only work fails to move numbers in most AI implementation is theater.
How long does the build take before the leak closes?
Most single-leak systems go live within 2 to 4 weeks. A missed-call voice agent can be answering calls inside a week. A follow-up automation usually takes a little longer because it touches your CRM and your data. The point is not speed for its own sake. The point is that the clock on the guarantee does not start until the system is live and handling real volume, so you are never paying for a half-built thing that has not had a fair chance to work.
What does the recovery guarantee actually cost kratt?
It costs us real money and real risk, and we want to be specific about that rather than wave at "skin in the game." If the system underperforms, we eat the hours. We keep building, fixing, and iterating with no new invoice until the leak closes. Those are billable hours we will never bill. For a two-person shop, that is the single most expensive promise we make.
It also costs us deals. We turn down prospects whose leaks we cannot measure, because guaranteeing an unmeasurable target is just a different kind of theater. A bigger firm would take that revenue. We pass on it. The guarantee disciplines who we work with as much as how hard we work.
And it caps our downside in a way most firms refuse to accept. We are betting our delivery time against your result. If we are wrong about the fix, that is our problem to solve, not a change order we send you. We compared this honestly against the usual billing models in the hidden cost of agency retainers.
There is a quieter cost too, and it is the one that keeps us honest. The guarantee forces us to say no to our own optimism. It is easy to look at a business and assume a fix will work. It is harder to commit free labor against that assumption. So before we ever attach the recovery guarantee to a number, we pressure-test our own plan as if we were the skeptic in the room. That extra rigor slows our sales cycle. We think it is worth it, because a guarantee you have to fight to honor was never really a guarantee.
How does this change the kind of client we want?
It narrows it, on purpose. The model only works with operators who have measurable volume and a real leak we can see in the data. A pre-revenue startup with no traffic has nothing for us to measure, so there is no honest number to guarantee. A 2 to 30 million dollar business with a busy phone line and a leaky follow-up process is exactly where the math lands. We would rather serve ten of those well than chase a hundred we cannot stand behind.
How does this compare to a normal retainer or a money-back guarantee?
The difference is who carries the risk and what triggers a payout. A retainer carries no result risk for the vendor. A project fee ends when the deliverable ships, working or not. A money-back guarantee returns your cash but leaves your problem unsolved. The recovery guarantee is the only one of the four where the vendor keeps working until the outcome lands.
| Model | What you pay for | Who carries result risk | If it fails | Lock-in |
|---|---|---|---|---|
| Recovery guarantee (kratt) | A closed leak | kratt | We work free until it closes | None |
| Typical retainer | Time and access | You | You keep paying | Usually 6 to 12 months |
| Project fee | A deliverable | You | You paid anyway | None, but no recourse |
| Money-back guarantee | A product or service | Split | You get a refund, not a fix | None |
Notice the last column. Even guarantees that return your money still leave the leak open. You are back where you started, minus the time you spent. The point of the recovery guarantee is that you wanted the leak fixed, not your money back, so we fix the leak. See the broader contrast in AI consultancy vs agency and in closed-loop systems vs a fragmented agency tool stack.
Why is outcome-based pricing so rare in consulting?
Because it is hard, and because the incentives fight it. For decades the industry billed by the hour or the head. Outcome-based pricing is a model the consulting industry has flirted with for years, and only recently has it scaled at the largest firms. McKinsey's operations research has long argued that fees should track measurable client value, yet roughly only a quarter of its own fees now come from outcome-based work. That tells you how stubborn the old default is.
The barrier is measurement. Tying a fee to a result only works when the result is measurable and the vendor controls delivery. Deloitte's analysis of outcome-based models stresses that both sides must agree on a clear, measurable goal up front. That is exactly what the audit produces. No audit, no number. No number, no honest guarantee.
Does the AI angle make guarantees riskier?
It would, if we sold the hype. The failure data is brutal. The Gartner forecast that 30 percent of generative AI projects would be abandoned after proof of concept by the end of 2025 turned out to be conservative. Roughly 80 percent of enterprise AI projects fail to deliver promised value, and most see zero measurable return.
That is precisely why we guarantee the business outcome, not the technology. MIT Sloan research on the gap between AI adoption and AI return points at the same root cause: most projects optimize for the tool, not the business result. Nobody at your company cares whether we used a model or a macro. They care whether the missed calls stopped. By anchoring to a revenue number instead of a tool, we sidestep the trap that sinks most AI work. We are not selling AI. We are selling a closed loop.
What are the limits of the guarantee?
There are honest ones. We guarantee the specific leak we audited, not your entire P&L. We do not guarantee outcomes that depend on your team executing work we did not build. And we do not invent numbers. This page is not a claim that the guarantee has paid out a certain number of times. It is a description of how the mechanism works, which is the part you can verify before signing anything.
Why risk reversal works, and why we are not afraid of it
Shifting risk from buyer to seller is one of the oldest moves in sales for a reason. It removes the single largest barrier to a yes, which is the fear of paying for nothing. Work published in Harvard Business Review on pricing strategy has long held that buyers pay more, and decide faster, when perceived risk drops. Research on B2B selling backs this up, with one analysis citing a 32 percent lift in win rates from risk-reversal language. We are not using it as a trick. We are using it because we can afford to.
An honest guarantee is also a filter on us. It forces us to only promise what we can measure and control, which keeps our work focused on real leaks rather than busywork. If you want to understand why we give the audit away in the first place, read why kratt gives the audit away. You can learn more about how we operate on the about page or see the work on our case studies.
You can also just start. Take the 2-minute leak quiz and we will tell you roughly where your money is going. That same quiz is the front door to the free audit that sets your guaranteed number.
The bottom line for operators
If you run a business doing 2 to 30 million dollars a year, you do not need another vendor who gets paid to be busy. You need the leak closed. The recovery guarantee exists so that our incentive and yours point the same direction. We win when your revenue stops leaking. We lose when it does not. That is the entire deal, and we wrote it that way on purpose.
Recovery Guarantee: your revenue stops leaking, or we work free until it does. No lock-in. Start with the free audit by taking the 2-minute leak quiz.