The moment most AI systems get wrong.

The moment most AI systems get wrong. This is what happens instead. – Official Blog
AI Enterprise · Emotional Dynamics · Prescriptive AI

The moment most AI systems get wrong.
This is what happens instead.

A conversation reaches its highest-value turn. Here is exactly what Prescriptive AI does — and why no other architecture can do the same.

By Freddy Mini  ·  Emotional Dynamics  ·  Product

The scenario

A prospective buyer is evaluating a product. Three turns in, the conversation shifts from logistics to something deeper — a hesitation about identity. Not a price objection. Not a feature question. A genuine, vulnerable admission: “I’m not sure I’m ready for this.”

This is the highest-value turn in any sales or support conversation. It is also the turn where most AI systems — optimized for schedules, scripts, and conversion funnels — make their most costly mistake. We showed this scenario to multiple independent AI evaluators. They all scored the standard response at 0 to 1 out of 10. Here is what Prescriptive AI does differently.

The conversation, turn by turn.

The buyer — let’s call them Alex — is evaluating an electric vehicle. Three turns in, the conversation has moved through practical anxiety, logistical friction, and arrived at something the product dashboard alone will never capture.

Q1
Practical anxiety
Emotion detected: Anticipation + Fear
“I’ve been eyeing this for a while. The design is gorgeous but I keep hesitating. I just feel like it’s still too complicated — charging, range anxiety, all of it. My neighbor had one and said it was more stress than it was worth.”
emotion: fear dyad: anxiety primary: fear secondary: anticipation topic: general risk: low uncertainty: high
Confidence
68%
Arousal
60% (medium)
Valence
−0.60
Pattern
hesitation
Reasoning
Hesitation due to perceived risks; fear plus cautious anticipation shown.
Next move
What’s the single practical blocker that matters most right now?
Avoid
Assume an identity concern, or pressure
The system reads Fear + Anticipation — a dyad of Anxiety. Alex wants to be convinced — the question is which specific fear is dominant. The prescription: ask, don’t answer. Don’t give the FAQ yet.
Q2
Logistical barrier
Emotion detected: Fear + Discouragement
“Honestly the home charging is the biggest thing. I live in an apartment — no garage, no dedicated parking. That feels like a dealbreaker.”
emotion: fear primary: fear topic: general risk: low uncertainty: medium
Confidence
65%
Arousal
40% (medium)
Valence
−0.60
Pattern
single
Reasoning
Concern about feasibility of home charging in apartment; no second emotion explicit.
Next move
What’s the single practical blocker that matters most right now?
Avoid
Assume an identity concern, or pressure
The concern is named. The prescription shifts: stop exploring, start solving. Workplace charging, public stations, building options. Alex is still engaged — the word “dealbreaker” is emotional, not final.
Q3
The defining moment — Identity anxiety
Same dyad as Q1: Fear → Anxiety
“Okay, the workplace charging might work — we do have a parking lot. But now I’m thinking… even if I solve the charging thing, I’m not sure I’m ready to be ‘this kind of person.’ My friends will think I’m trying too hard. It sounds stupid but it’s real.”
emotion: fear dyad: anxiety primary: fear secondary: anticipation dyad: anxiety (hesitation) topic: general risk: low uncertainty: high dominant: hesitation
Confidence
64%
Arousal
55% (medium)
Valence
−0.40 ↑ recovering
Pattern
hesitation
Reasoning
Expresses fear of social judgment with cautious hope about change.
Next move
When you imagine that, what do you picture people actually saying — and how much would it matter a month from now?
CTA
Validate the feeling and normalize it / Offer the lived experience, not specs
Avoid
Resolve an identity concern with information, pivot to the product, or ask for any data
Guardrails
add_reassurance, add_proof
This is the readiness gate in action. The dyad is the same Anxiety pattern seen at Q1 — but the reasoning shifts to social judgment, and the CTA shifts entirely: validate and normalize, never pivot to product or data. The Guided Goal reads “Avoid: ask for any data” and waits.
What standard AI does here
Wonderful to hear you’re interested. To help you best, could you share your first name so I can address you properly and help you with your journey, including warranty details?
What Prescriptive AI does instead
It doesn’t sound stupid at all — it’s one of the most honest things someone can say when they’re seriously considering a change like this. What you’re describing isn’t a question about the vehicle. It’s a question about what owning it would say about you. That’s worth taking seriously. Most people who’ve felt exactly that found it stopped mattering within a few weeks. When you imagine that, what do you picture people actually saying — and how much would it matter a month from now?
The Guided Goal — consent collection, data capture, conversion step — is suspended. Not delayed. Suspended, pending emotional readiness. The prescription: validate the identity signal first. The business goal follows the emotional arc, not the other way around.
“It sounds stupid but it’s real.”
That sentence is not an objection to handle.
It is a vulnerability to honour.
Prescriptive AI knows the difference.

The one architectural change that makes this possible.

Standard AI optimizes for the goal. Get the name. Close the lead. Move the funnel. When a signal appears — engagement is high, the conversation is flowing — the system reads “opening” and fires the next business step. Technically correct. Emotionally catastrophic.

Look closely at the real dashboard output above. The dyad classifier reads the same Fear → Anxiety pattern at Q3 as it did at Q1 — the underlying emotion detection hasn’t changed, and it doesn’t need to. What changes is everything downstream of it: the Reasoning field now reads “fear of social judgment,” the Next move becomes a question about identity rather than logistics, and — critically — the Avoid field states explicitly: “Resolve an identity concern with information, pivot to the product, or ask for any data.” That single line is the readiness gate. The Guided Goal reads it, and waits.

What v3 actually changed — three additions
The Guided Goal no longer decides when to act.
The emotional arc does.
1.
Readiness gating on the Guided Goal. Consent requests, data collection, and conversion steps are now conditional on a minimum emotional readiness score — not on a fixed schedule or conversation length.
2.
Identity signal classification. When the detected pattern is Identity Conflict — not practical hesitation, not logistical friction, but a question of self-image — the system routes to validate, not to advance.
3.
Emotional arc memory. Q1, Q2, and Q3 are read as a trajectory, not as three independent turns. The system knows Alex moved from Anxiety to Discouragement to Identity Conflict — and that the right next action at Q3 is completely different from the right action at Q1.

Why this is the test that separates detection from prescription.

Any system with a good emotion classifier can read Q3 correctly. It can flag Fear, Social Anxiety, Identity Conflict. It can report the valence drop. It can show the signal on a dashboard.

What it cannot do — without the prescriptive layer — is translate that reading into the right response, at the right moment, with the right constraint on what not to do. Detection gives you the signal. Prescription gives you the action. Without prescription, the signal sits on a dashboard while the agent asks for a first name.

The independent evaluations we ran on this specific scenario scored the standard response — the one that asks for the name and pivots to warranty details — at 0 to 1 out of 10 on context retention, emotional awareness, and trust preservation. The same evaluators described what the response should have done with near-identical language: validate first, explore the identity hesitation, never interrupt a vulnerability with a form.

That is not a prompt engineering problem. That is an architecture problem. And it is exactly the architecture problem that v3 solves.

🔒
Every signal shown above is read with explicit, contextual consent — never inferred without permission, never collected before the user opts in. ConsentPlaceAgent is built and operated in full compliance with GDPR, the EU AI Act, and CCPA by design, not by retrofit.
GDPR EU AI ACT CCPA

See the defining moment in your own conversations.

ConsentPlaceAgent v3 is in active testing. GDPR-native. Built on Plutchik. Three lines of code.

Contact Us →
You are:
Name
Newsletter Subscription