In this episode of Terminal Value, Nik Singh sits down with Dhaval Shreyas, co-founder of Pie, to discuss how QA is evolving from manual scripts and brittle automated tests into product intelligence.

As AI accelerates software development, engineering teams are shipping faster than traditional QA teams can keep up. Dhaval explains why the future of quality is not just “agents writing tests,” but systems that understand the product, the user experience, and the business context behind each flow.

We cover how Pie discovers a product from a staging URL or app build, how it builds and maintains coverage, why product context matters more than scripts, and why QA may increasingly merge with product, engineering, and product management.

We cover:

Why QA is underinvested in and often becomes a bottleneck
How AI-driven development is increasing pressure on quality teams
Why QA is moving from scripts to product understanding
How Pie discovers product flows and builds coverage
Why product context is hard for coding agents like Claude Code to replace
How Pie compares to Selenium, QA Wolf, and agentic coding tools
The role of PQEs and human-in-the-loop deployment
How the QA role may evolve as software teams become more AI-native
Why product context could become valuable beyond QA, including documentation and support

Terminal Value explores where value accrues as AI, software, markets, and infrastructure change.

Subscribe for more conversations with founders, operators, and investors building the next generation of software.

Chapters

00:00 Cold Open: QA Is Falling Behind
00:55 Intro: Pie and the AI-Native QA Layer
01:50 What QA Looks Like Today
04:02 Why Traditional QA Breaks Down
04:53 Agentic QA and Product Intelligence
07:11 Pii’s Deployment Journey
09:04 How Pie Builds Product Context
10:56 Ingesting PRDs, Help Docs, and Test Cases
12:52 Who Buys AI-Native QA?
14:43 Human-in-the-Loop Deployment and PQEs
15:58 Pricing the PQE Model
16:36 Pie vs. Selenium, QA Wolf, and Testing Agents
18:33 Why Claude Code Alone Is Not Enough
19:49 How the QA Role Changes
21:49 Will QA Merge Into Product and Engineering?
22:48 Pie’s Bigger Vision Beyond QA
25:37 Closing Takeaways

#AI #Software #QualityAssurance #QA #ProductIntelligence #EnterpriseSoftware #DeveloperTools #AgenticAI #TerminalValue

Finance software is starting to deploy itself.

In this episode of Terminal Value, Nik Singh sits down with Ahikam Kaufman, Co-Founder and CEO of SafeBooks, to discuss how AI agents are changing finance operations, revenue integrity, and the modern CFO stack.

SafeBooks is building an agentic revenue integrity platform that connects systems like CPQ, CRM, contracts, billing, ERP, and revenue recognition — then gives finance teams a way to validate transactions, catch errors, automate workpapers, and ask questions across their financial data.

We discuss why finance teams still spend so much time manually checking data, how AI can create a financial data graph across systems, why forward-deployed engineering may become less important over time, and why many finance AI workflows may not require frontier models.

We also cover the future of accounting work, the difference between SafeBooks and legacy close-management platforms, and what happens when finance operations become real-time, agent-driven, and self-serve.

Chapters:
00:00 Cold open: AI agents for finance operations
00:45 Introduction: SafeBooks and the revenue integrity problem
01:44 What revenue integrity means in practice
03:20 The finance stack behind revenue integrity
04:31 What an agentic revenue integrity workflow does
06:09 The financial data graph behind SafeBooks
08:54 Deploying across CRM, billing, ERP, and accounting systems
10:27 Are forward-deployed engineers going away?
11:16 SafeBooks’ target customer and mid-market use case
13:03 Pricing AI software with SaaS + usage models
16:20 Why finance AI may not need frontier models
18:39 SafeBooks vs. BlackLine, FloQast, and close-management tools
21:09 Why deployment, context, and real-time controls matter
23:40 How AI changes finance employment
25:16 The future of automated finance operations
27:51 Closing and final takeaways

Subscribe for more conversations on AI, software, markets, and where value accrues.

AI agents are starting to change how marketing teams turn digital traffic into revenue.

In this episode of Terminal Value, Nik Singh sits down with AJ Goyal, co-founder and CEO of Fibr AI, to unpack why web conversion is still so manual — and how agentic AI could reshape the workflow between paid traffic, websites, experiments, and revenue.

For years, marketing teams have gotten increasingly sophisticated at targeting ads, segmenting audiences, and optimizing acquisition. But once that traffic reaches the website, the experience often becomes static, generic, and dependent on analysts, agencies, marketing operations, and engineering tickets.

Fibr sits in the middle of that workflow. The company helps teams take the intent already captured in ads, campaigns, and analytics and turn it into personalized web experiences that can be tested and improved continuously.

We cover:

* Why web conversion is still such a manual workflow
* How marketers depend on analysts, agencies, marketing ops, and engineering
* Why personalization often breaks down after the ad click
* How Fibr uses AI agents to create and improve web experiences
* Why reducing CAC is the wedge for enterprise buyers
* How agentic software changes pricing, deployment, and GTM
* Why incumbents may struggle when AI threatens their agency ecosystems
* What happens when AI agents become users of the web

Chapters:

00:00 – Hook
00:30 – Introducing AJ Goyal and Fibr AI
01:39 – How web conversion works today
04:10 – The modern marketing and web stack
06:49 – Where the workflow breaks down
09:29 – What Fibr does
11:39 – How Fibr changes the marketer’s workflow
14:16 – The first wedge: reducing CAC
15:00 – Customer example: 48% CAC reduction
16:35 – Where humans stay in the loop
18:12 – Why more data improves the agent loop
19:09 – Fibr’s enterprise ICP
21:12 – Go-to-market: events, conferences, and LinkedIn
23:04 – Pricing agentic software
25:12 – Enterprise pricing pushback
27:20 – Proof-of-concept motion and ROI
28:59 – Deployment and implementation
29:47 – Competitive landscape: Optimizely, Adobe, and agencies
31:38 – Incumbents, agency channels, and agentic conflict
33:31 – Product utilization and always-on experimentation
35:10 – Could agencies become a channel?
36:17 – How AI changes marketing orgs and agencies
38:59 – Will marketing teams get leaner?
39:28 – The future of agentic web experiences
41:29 – Final takeaways: workflow collapse, GTM conflict, and agents as users

Subscribe to Terminal Value for conversations with founders, operators, and investors on where value accrues across markets, software, AI, and infrastructure.

#AI #MarketingAI #AgenticAI #EnterpriseSoftware #SaaS #ConversionOptimization #DigitalMarketing #TerminalValue