Economics And Pricing: Outcomes-Based Contracts And Demand Expansion Effects
Sources: 1 • Confidence: Medium • Updated: 2026-04-11 19:44
Key takeaways
- As AI absorbs simpler support tickets, the remaining human-handled cases become more complex, which can increase average handle time while improving agent job satisfaction.
- AI creates uncertainty for software valuations by potentially reducing the durability of SaaS annuity revenue and altering discounted cash flow assumptions.
- Sierra has seen AI agents from different healthcare entities place and receive calls from each other using English over the public switched telephone network.
- Well-known brands can be harder to ground because models may rely on public internet knowledge and go off-script, whereas obscure domains can be easier to constrain.
- The atomic unit of AI-driven productivity is an end-to-end process rather than an individual role because AI excels in digital steps but not in physical-world tasks.
Sections
Economics And Pricing: Outcomes-Based Contracts And Demand Expansion Effects
- As AI absorbs simpler support tickets, the remaining human-handled cases become more complex, which can increase average handle time while improving agent job satisfaction.
- Outcomes-based pricing is different from usage-based pricing because tokens/utilization do not reliably correlate with business value, and outcomes pricing shifts token-efficiency incentives onto the vendor.
- Outcomes-based pricing increases a software vendor's accountability for customer success by tying revenue to realized results rather than product usage or implementation milestones.
- AI can reduce the marginal cost of a service interaction by orders of magnitude (from roughly $10–$20 to cents and potentially fractions of a cent), enabling more and better support that changes churn and lifetime value dynamics.
- Sierra reports that some customers automate a very high share of support cases, with examples ranging from about 70% to 90%, and Ramp cited at 90%.
- In at least one retail deployment, AI improved the experience enough that total customer conversation volume increased 2–3x, largely offsetting cost savings while increasing perceived business value.
Market Structure And Governance Watch Items (Commoditization, Valuation Uncertainty, Openai Mission Duty)
- AI creates uncertainty for software valuations by potentially reducing the durability of SaaS annuity revenue and altering discounted cash flow assumptions.
- As agents perform valuable labor, value may shift away from traditional systems-of-record toward agent layers, especially in systems-of-engagement where agents orchestrate actions without humans logging into the core app.
- Bret Taylor argues that AI will not concentrate into only a couple of dominant companies and that the absence of many applied AI vendors is slowing adoption.
- Applied AI product teams should expect to throw away significant bespoke work as foundation model capabilities commoditize previously differentiating features over a few years.
- The applied AI market will be large and durable because most enterprises want solutions to specific business problems rather than models or generic AI software.
- OpenAI's nonprofit board fiduciary duty is to the mission of ensuring AGI benefits humanity rather than to shareholder value.
Customer Experience Agents On Legacy Rails (Telephony) And Omnichannel Deployment Pattern
- Sierra has seen AI agents from different healthcare entities place and receive calls from each other using English over the public switched telephone network.
- Sierra builds customer experience AI agents that can answer questions and take actions across phone and digital channels, including replacing IVR systems by answering calls directly.
- Most Sierra customers begin deployment on a single channel with a few use cases and then expand toward both phone and digital coverage.
- Sierra enables a single branded agent to operate consistently across channels (phone, website chat, WhatsApp, mobile app), potentially unifying previously separate digital and call-center teams.
- Some Sierra deployments expand beyond support into product usage workflows, including AI-assisted home search at Redfin and mortgage origination/servicing at Rocket properties.
- A company's primary AI agent will become the majority of its digital interactions, with digital increasingly including the telephone.
Reliability, Grounding, And Supervision Architecture
- Well-known brands can be harder to ground because models may rely on public internet knowledge and go off-script, whereas obscure domains can be easier to constrain.
- Chaining a ~90%-accurate reasoner with a ~90%-accurate supervisor can yield roughly 99% effective behavior through layered checks.
- Basic reasoning capability in modern LLMs is the key breakthrough enabling effective customer experience agents compared to pre-LLM chatbots.
- Sierra uses a constellation approach in which supervisor models inspect and can reject an agent's reasoning when it violates guardrails or fails to consult required policies.
- LLMs' broad pretrained knowledge can materially improve support agent effectiveness even before adding company-specific manuals and telemetry.
Enterprise Adoption Bottlenecks: Procurement Mismatch And Process-Level Operating Model
- The atomic unit of AI-driven productivity is an end-to-end process rather than an individual role because AI excels in digital steps but not in physical-world tasks.
- AI productivity gains are slower in many companies because organizations are structured around departments rather than clear ownership and KPIs for cross-functional processes that AI could optimize.
- Narrowing AI to a specific end-to-end business process can turn an open-ended science problem into an engineering problem that can be automated with scaffolding and rigid rules.
- Enterprise AI adoption is hindered because frontier models are often sold to different internal buyers than departmental executives who purchase workflow solutions.
- To extract substantial AI uplift, companies should prioritize highly digital workflows instead of broadly issuing copilots to every employee.
Watchlist
- Second- and third-order effects of widespread AI adoption may reshape industry equilibrium beyond straightforward automation effects.
- AI creates uncertainty for software valuations by potentially reducing the durability of SaaS annuity revenue and altering discounted cash flow assumptions.
- As agents perform valuable labor, value may shift away from traditional systems-of-record toward agent layers, especially in systems-of-engagement where agents orchestrate actions without humans logging into the core app.
Unknowns
- What definitions and measurement methods underpin Sierra's reported ARR milestones, automation percentages, and 'resolved case' outcomes pricing unit?
- How do Sierra deployments perform on reliability and compliance metrics over time (error rates, escalation rates, policy adherence), especially for high-risk or regulated workflows?
- How much of the claimed marginal cost decline per interaction is attributable to model inference costs versus integration, supervision, monitoring, and exception-handling costs?
- Do harness-based integrations measurably outperform UI-driving agents on the same enterprise tasks in real deployments (latency, success rate, cost, maintainability)?
- Will competitive dynamics actually force AI savings into consumer surplus (price reductions/experience escalation), or will some sectors retain margins due to market power or switching costs?