Enterprise Software Absorbs AI Deployment Costs

On September 8, 2026, customer engagement software provider Braze (BRZE) and home-services software provider ServiceTitan (TTAN) both said demand for paid AI modules had exceeded implementation capacity. Vendors are absorbing customer rollout costs through services, free onboarding and deferred billing.[1][2]

---

AI implementation capacity has become a revenue constraint

These AI products act on customer data and change operating workflows, so deployment requires process redesign, governance and user training. Software vendors must now supply more of that implementation labor if they want customers to adopt premium AI modules quickly. Braze had previously delayed some Decisioning Studio start dates by more than four months, while ServiceTitan said demand exceeded its ability to implement.[2][3]

The companies absorb the cost differently. Braze includes forward-deployed engineers, deliverability support and customer-success services in recurring professional services, while certified agencies onboard most new customers. ServiceTitan waives onboarding fees for existing customers moving to MAX and typically does not bill the first quarter, so faster adoption can reduce near-term reported revenue.[1][2]

---

Services mix and deferred revenue reveal the cost

Braze's professional services share rose from 5% of revenue in fiscal 2026's third quarter to 7% in fiscal 2027's first quarter and 9% in the second quarter. The company expects 9%-10% going forward, with about 90% recurring.[1][3] Management named demand growth as one cause, but earlier said part of the increase reflected a revenue-classification change. The full increase therefore cannot be attributed to AI implementation.

ServiceTitan quantified the same issue as a fiscal-year headwind. MAX mix and revenue-recognition timing will reduce subscription revenue by $2-$3 million, while waived onboarding fees will reduce professional services revenue by about $2 million, for a combined near-term impact of $4-$5 million.[2] Management said:

“We have greater demand than we have the ability to implement right now and hence the focus on making the implementation much more efficient so we can scale more quickly.”

Demand is available; implementation throughput limits expansion.

---

Self-service products and partners determine scaling speed

Vendors are using lighter products and external partners to work around the labor bottleneck. Braze launched the nearly self-service Decisioning Studio Go and said more than 80% of new customers use certified agency partners to begin long-term services engagements. ServiceTitan plans a lighter MAX entry package that lets customers start with demand orchestration before transforming field operations.[1][2] These approaches increase reach but may trade lower prices or slower revenue recognition for deployment speed.

The main test is whether implementation time falls without larger customer-facing teams, followed by lower deferred revenue and services costs. The current economics remain small: Decisioning Studio represented about 2.9% of Braze's quarterly revenue, and ServiceTitan's $4-$5 million headwind is limited relative to annual revenue of roughly $1.2 billion.[1][2] Lighter modules may also use in-product agents to automate adoption, so this is a constraint on implementation-heavy AI products rather than the entire software industry.

---

Companies exposed to this shift

• Klaviyo (KVYO): Free credits and its self-service Composer lower activation friction, exposing it to the same adoption mechanism, although it has not disclosed Braze-like implementation-services pressure.[4]
• Alkami (ALKT): It has $61 million of annual recurring revenue in an implementation backlog across 37 signed clients, making rollout time a visible revenue-recognition issue.[5]
• Twilio (TWLO): More AI-generated multichannel campaigns could increase messaging traffic on its communications platform, but the company says AI support for messaging remains early.[6]

---

Sources

[1] Drillr · Braze · 2026-09-08 · Fiscal 2027 second-quarter earnings call

[2] Drillr · ServiceTitan · 2026-09-08 · Fiscal 2027 second-quarter earnings call

We have greater demand than we have the ability to implement right now and hence the focus on making the implementation much more efficient so we can scale more quickly.

[3] Drillr · Braze · 2026-05-27 · Fiscal 2027 first-quarter earnings call

[4] Drillr · Klaviyo · 2026-08-05 · Second-quarter 2026 earnings call

[5] Drillr · Alkami · 2026-07-29 · Second-quarter 2026 earnings call

[6] Drillr · Twilio · 2026-08-06 · Second-quarter 2026 earnings call

This material identifies potentially overlooked industry changes and companies. It is not a stock recommendation.