AI product GTM
Turn an AI product, buyer concern, pricing problem, or enterprise sales blocker into positioning, qualification, demo, trust, and pricing recommendations for AI go-to-market and to-market execution.
When to invoke
- "How do we position this AI product?"
- "Buyers say they're worried about AI breaking production."
- "Should we call it autonomous, agent, copilot, or teammate?"
- "How do we price AI when usage varies 10x by customer?"
- "Enterprise security passed but ops rejected us; why?"
Positioning decisions
Use wording that matches buyer risk tolerance rather than internal architecture pride.
Does your AI act autonomously with no approval per action?
├─ Yes → Who are you selling to?
│ ├─ Developers → "Agent" framing
│ └─ Enterprises → "Teammate" framing
└─ No → "copilot" framing
| Framing | Use when | Avoid when |
|---|---|---|
| copilot | The AI suggests and a human approves each action. | The product actually takes unsupervised action. |
| Agent | Technical developer buyers expect automation and can inspect failure modes. | Enterprise buyers hear "autonomous" as unmanaged risk. |
| Teammate | Enterprise buyers need accountability, escalation, and control language. | The product is only a passive recommendation widget. |
Prefer words like teammate, augments, accelerates, and you stay in control. Avoid leading with autonomous, replaces, fully automated, or AI-first when the buyer has not accepted the operating model.
Buyer readiness and trust ladder
Qualify AI-agent buyers by operational maturity, not by enthusiasm.
Do they have incident response processes for tool failures?
├─ Yes → Continue
│ └─ Do they have on-call rotations for production systems?
│ ├─ Yes → Qualified buyer
│ └─ No → Help them build it first
└─ No → Not ready; come back in 6 months
The real objection behind "will it break production?" is "who is responsible when it does?" Map the answer to the buyer's operating model.
| Enterprise objection | Map it to |
|---|---|
| "Who gets paged when AI breaks production?" | Their on-call rotation. |
| "Who debugs AI failures?" | Their incident response process. |
| "Who owns customer communication?" | Their escalation path. |
Trust sequence matters: Transparency → Control → Performance → Scale. Provide model cards, security notes, and explainability before the demo; then show approval workflows, kill switches, and confidence scores; then benchmarks, case studies, and live demo; then enterprise deployments, compliance, and SLAs.
Pricing and demo patterns
Can you measure customer outcomes reliably?
├─ Yes → Outcome-based, or hybrid with outcome component
└─ No → Does usage vary 5x+ by customer?
├─ Yes → Hybrid: base + usage
└─ No → Seat-based
Pricing hybrid formula:
Base: $X/month (covers fixed costs)
Variable: $Y per unit (20-30% of customer's alternative cost)
Demo structure:
- Problem with quantified cost (30s).
- AI attempt including failure or uncertainty (60s).
- Human review and override (30s).
- Outcome with ROI (30s).
Show mistakes plus recovery. That builds more trust than a perfect AI demo that looks staged because buyers know real-world data is messy. Name failure/uncertainty explicitly.
Common mistakes
| Mistake | Why it loses deals | Correction |
|---|---|---|
| Using "autonomous" because it sounds impressive | It scares enterprises and slows deals. | Use "teammate" once the product takes action. |
| Hiding AI failure modes | Buyers assume missing failure examples are being concealed. | Show failures, recovery, and ownership. |
| Treating "will it break production?" as the objection | The buyer is asking about responsibility. | Map failure ownership to incident response and on-call. |
| Pricing usage-based AI like OpenAI | Your cost structure and customer value are different. | Price for 20-30% of the customer's alternative cost. |
| Skipping transparency docs before demo | Buyers need proof before performance claims. | Sequence Transparency → Control → Performance → Scale. |
| Demoing perfect AI | Fake perfection reduces trust. | Include uncertainty, review, and override. |
| Selling to buyers demanding 100% accuracy | They are not ready for agentic AI. | Filter for mature buyers with incident response. |
Also preserve the ceiling-moment qualification insight: when a buyer cannot define responsibility for failures, the deal has hit an operating-model ceiling rather than a feature objection.
Progressive disclosure and bundled resources
references/core-frameworks.md: detailed AI go-to-market frameworks and examples for full GTM strategy work.
Related primitives
| Name | Type | Use it when |
|---|---|---|
positioning-strategy |
skill | You need general positioning frameworks outside AI-agent GTM. |
technical-product-pricing |
skill | You need broader pricing model work beyond AI variable-cost patterns. |
enterprise-account-planning |
skill | You need account strategy for enterprise AI deal management. |
Output template
## AI GTM recommendation — <product or deal>
**Positioning:** copilot | agent | teammate
**Buyer readiness:** qualified | needs operating model | not ready
**Pricing model:** seat-based | hybrid base + usage | outcome-based
| Decision | Recommendation | Rationale | Evidence needed |
| --- | --- | --- | --- |
| Framing | <copilot/agent/teammate wording> | <why it fits buyer risk> | <proof to gather> |
| Trust sequence | <next trust asset> | <why now> | <doc, demo, benchmark, SLA> |
| Pricing | <model and unit> | <cost/value logic> | <usage, alternative cost, outcome metric> |
### Demo plan
1. <30s problem and quantified cost>
2. <60s AI attempt with failure or uncertainty>
3. <30s human review or override>
4. <30s ROI outcome>
Quality gate
- The recommendation chooses copilot, agent, or teammate framing with buyer-specific rationale.
- Production-responsibility objections are mapped to on-call, incident response, and escalation ownership.
- Trust assets follow Transparency → Control → Performance → Scale.
- Pricing accounts for 5x+ or 10x usage variance and uses 20-30% of alternative cost when relevant.
- The demo plan includes failure or uncertainty plus recovery.
- Buyer readiness is qualified before recommending autonomous or teammate messaging.