Understanding the Echo Mode in AI Engineering
An Echo is a domain rebel: someone who came from the domain, understands it deeply, and has concluded that the current way of doing things is inadequate.
Why the Most Experienced Expert Can Mislead
When a small team starts building an AI product for an unfamiliar industry, a natural instinct is to find the most experienced person in that industry and ask what they think. The reasoning seems sound: the person who has done the work longest should be able to say whether a new system is needed. This instinct rests on a hidden assumption, namely that a practitioner who knows a field well will also see its problems clearly. In practice, the most experienced person is often the least likely to say that the way things are done is broken, because they have built their working life around that way of doing things.
What do you think happens?
Two people have years of experience in the same domain. One is content with how the domain currently operates. The other understands the domain just as deeply but has concluded that the current approach is inadequate. Which of them would be more useful when you are testing whether your new AI system solves a real problem?
Reveal answer
The one who judges the current approach inadequate. The source calls this person an Echo.
Both people know the field, so knowledge alone does not separate them. What separates them is how each relates to the current state of the domain. The contented expert has a stake in the status quo, while the Echo can say precisely what is broken.
An Echo is a domain rebel: someone who came from the domain, understands it deeply, and has concluded that the current way of doing things is inadequate.
Two Profiles Behind the Same Expertise
The source draws a sharp line between the Echo and what it calls a satisfied practitioner. Both can have deep knowledge of the domain. The difference lies in the second trait: whether the person has concluded that the current approach is inadequate. A satisfied practitioner regards the current approach as fine. Their skills, habits and sense of competence are tied to how things are done today, so when they see a new system, the most likely response is that the existing approach already works and the new system is unnecessary. For that reason, the source treats the practitioner's experience alone as a weak signal that a new system is needed.
Notice what the diagram shows. The two profiles do not differ in how much they know. They differ in where they stand relative to the domain's current methods. The Echo's judgment is the trait that changes what their knowledge is used for.
From Deep Understanding to Usable Feedback
Holding both traits at once is what makes an Echo more useful than a contented expert. Deep understanding gives the Echo the vocabulary and detail to locate a problem. Rejection of the status quo gives them a reason to look for the problem rather than explain it away. Together they produce feedback that is specific enough to act on. The source says an Echo can tell you three things: where the failure is, why previous efforts to fix it did not work, and what would need to change before a new attempt could succeed.
A Customer-Side Rebel in a Small Startup
The source moves from the individual profile to the startup that needs one. A small AI startup without an Echo team should have one of its early hires be a customer-side rebel. This person needs two things. The first is domain credibility, so that people on the customer side take them seriously. The second is institutional skepticism, so that they are willing to question how the customer's work is being done. Their role is to translate between the startup's engineering team and the customer's reality, so that engineers build against what the customer actually faces rather than an idealized version of it.
A startup building for an unfamiliar industry
A three-person startup is building an AI tool for a industry its founders have never worked in. Their engineers have a design that looks reasonable on paper. How should the team use a customer-side rebel?
- 1Identify the risk
The engineers are working from their own picture of the customer's work. Without someone who knows that work from the inside, the design may solve an idealized problem.
- 2Look for the right profile
The team needs someone with domain credibility who is also willing to question current practice. An experienced person who is satisfied with current methods would not provide the critical feedback the team needs.
- 3Use the person as a translator
The rebel takes the engineers' design to the customer and brings back what the customer actually faces. The engineers then revise the design against that reality.
- 4Test the feedback
The rebel's answers about where the failure is, why earlier fixes failed, and what must change become the requirements the team builds against.
The startup's engineers build against the customer's real situation, guided by a person who can say exactly what is broken and what conditions a new system must meet.
Arrangement Versus Presence
The source is explicit that the customer-side role can be filled in more than one way. The rebel may be paid by the startup, may serve as an advisor, or may already work inside the customer organization and have been thinking about the problem for years. The source adds that the structure of the arrangement does not matter. What matters is that the rebel is present. A startup should therefore focus its effort on finding a rebel who will actually be part of the work, and be flexible about the contract that brings them in.
The contract is how the Echo is engaged. The Echo's presence is what gives the team the diagnosis it needs. Choosing an arrangement is secondary to making sure the right person is actually in the room.
Mistakes Teams Make When Choosing Advisors
Common mistake Treating the most experienced person in the industry as the best source of guidance
What it looks likeA team asks the longest-serving practitioner whether a new system is needed and accepts a no as the answer.
Why it failsExperience alone is a weak signal of need. A satisfied practitioner has built their working life around the current approach and is likely to see the new system as unnecessary.
Do this insteadAsk instead how the person judges the current approach, and look for someone who can name specifically what is broken.
Common mistake Assuming deep knowledge is enough on its own
What it looks likeA team hires an industry insider who knows the field thoroughly but never questions its methods.
Why it failsThe source describes the Echo through a combination of origin, depth and judgment. Any one trait on its own does not make an Echo.
Do this insteadCheck for all the relevant traits, especially the judgment that the current way of doing things is inadequate.
Common mistake Insisting on one particular contract structure
What it looks likeA startup refuses to engage a rebel unless they are on the payroll as a full employee.
Why it failsThe source states that the arrangement can take several forms and that the structure does not matter as much as the rebel's presence.
Do this insteadStay open to paid, advisory, or customer-embedded arrangements and prioritize the person's involvement.
Common mistake Confusing feedback that flatters the design with feedback that exposes its failures
Checking Your Understanding of the Echo
A colleague says that a veteran in your target industry is the ideal reviewer for your new AI system, because they have decades of experience and say the current process works well. Using the two traits that separate an Echo from a satisfied practitioner, explain why this reviewer may not be the right source of guidance, and describe what kind of feedback you should ask for instead.
Hints
- Ask whether the veteran has concluded that the current approach is inadequate, or whether they regard it as fine.
- Consider what the source says a satisfied practitioner is likely to say when shown a new system.
- Think about the three things an Echo can tell you: where the failure is, why earlier fixes failed, and what must change.
Key Takeaways
- An Echo is a domain rebel: someone who came from the domain, understands it deeply, and has concluded that the current way of doing things is inadequate.
- A satisfied practitioner regards the current approach as fine, so their experience alone is a weak signal that a new system is needed.
- Deep domain understanding combined with rejection of the status quo produces specific feedback: where the failure is, why earlier attempts failed, and what must change.
- A small AI startup without an Echo team should bring in a customer-side rebel early to translate between engineering and the customer's reality.
- The arrangement can take many forms, and what matters most is that the rebel is present.