Welcome to the ninth episode of Oh, Shift!: Agentic Software Leader Series, where we cut through the hype and have candid conversations with software leaders navigating the seismic shift of the AI era. This series is designed for founders and operators who are focused on building internal AI momentum without relying on silver-bullet thinking.

In this episode, host Rick Watson talks with:

Karthik Marudur, VP of Strategy at OneStock, a founder-led order management platform used by retailers to orchestrate complex omnichannel operations. Karthik’s a veteran of retail tech and supply chain—he helped build large-scale OMS programs and brings a pragmatic, ‘business-problem-first’ view to AI.

We go inside a founder-led Order Management Software company where a quiet internal bot has already rewired sales, support, and product—and a new ‘config agent’ aims to turn six months of set-up into minutes. If you’re leading teams through AI without rewriting everything, this one’s for you.

OneStock's AI initiatives include a configuration agent that compresses traditional setup times from months to minutes by dynamically configuring and simulating order workflows, improving operational efficiency and profitability. They build a data lake from sources like Confluence and Jira to power AI tools that assist sales, marketing, product, and development teams. Security and privacy are managed carefully, with legal teams involved in testing.

Marudur emphasizes the importance of measuring the adoption and ROI of AI tools, maintaining agility in technology choices, and evolving from internal experimentation to co-innovation with customers.

The discussion also covers the emerging role of MCP in providing contextual, conversational AI experiences beyond traditional APIs.

The episode highlights a pragmatic, business-focused approach to integrating AI in SaaS retail solutions, advocating for starting small, learning continuously, and aligning AI efforts with real business outcomes.

#Ohshift #ai #ordermanagement #agent

“Change isn’t painful—resistance is. Today, Yotpo co-founder Tomer Tagrin shares how his teams ship outcomes over code, why 70% of support tickets are now deflected by AI, and how to keep authenticity and trust when LLMs are everywhere.”

Welcome to the eighth episode of Oh, Shift!: Agentic Software Leader Series, where we cut through the hype and have candid conversations with software leaders navigating the seismic shift of the AI era. This series is designed for founders and operators who are focused on building internal AI momentum without relying on silver-bullet thinking.

In this episode, host Rick Watson talks with:

Tomer Tagrin, Co-Founder and CEO of Yotpo, explores Yotpo's comprehensive reset in culture, strategy, product development, and go-to-market approaches driven by AI and post-COVID realities.

Emphasizing authenticity as a superpower, Yotpo focuses on building trust with customers and agents in an AI-first world, particularly through enhancing reviews and loyalty programs.

Tagrin acknowledges execution challenges, and Yotpo divested some products to focus on core strengths.

AI is transforming engineering roles, accelerating coding, prototyping, and customer feedback analysis, while also reshaping support and sales functions.

Yotpo has invested in external expertise to improve product development and is experimenting with agentic commerce to adapt to new market dynamics.

The interview highlights the importance of culture change, transparency, and continuous learning within the company. It also discusses the future of commerce across multiple channels and interfaces, such as voice, and stresses the need for brands to maintain authenticity amid increasing AI-driven interactions.

Tagrin advises leaders to embrace change, remain humble, and focus on long-term growth despite short-term challenges.

Welcome to the seventh episode of Oh, Shift!: Agentic Software Leader Series, where we cut through the hype and have candid conversations with software leaders navigating the seismic shift of the AI era. This series is designed for founders and operators who are focused on building internal AI momentum without relying on silver-bullet thinking.

In this episode, host Rick Watson talks with:

Jason Cottrell, CEO of Orium and incoming president of the Mach Alliance, discusses the profound shift AI is causing in business culture, hiring mindsets, and workflows, comparing it to past pivots like Agile adoption. He stresses that managing change is harder than the technology itself due to initial fear and hesitancy. Orium has adopted a forward-leaning strategy to stay ahead, integrating AI into workflows and emphasizing the QA mindset crucial for building reliable AI agents.

Jason outlines a four-stage employee progression model for AI adoption—from resistance to transformative use—and highlights peer learning and safe environments as key to advancing adoption. He notes challenges such as enterprise AI bans and evolving governance, advocating for cross-functional teams with autonomy to execute AI strategies.

As Mach Alliance president, he aims to guide members through AI and agentic ecosystem development, anticipating changes in membership criteria to reflect AI's growing role.

Overall, the discussion emphasizes rethinking processes with AI at the core rather than simple tool insertion, and the importance of leadership in navigating this transformative era.

#ohshift #machalliance #culture

In this episode, host Rick Watson talks with:

Bryan House, CEO of Elastic Path, discusses AI's transformative impact on commerce. OpenAI's instant checkout exemplifies AI's shift toward directly monetizing transactions, leveraging its vast user base and memory capabilities for personalized shopping.

House discusses how ElasticPath embraces AI through Model Context Protocol (MCP) to streamline storefront development and enhance merchandiser workflows via natural language prompts. How AI accelerates product development and prototyping, improving communication between product managers and engineers.

He discusses potential AI challenges, including managing probabilistic AI outputs, ensuring agent authentication to prevent fraud, and balancing speed with accuracy in user experiences. The company fosters AI adoption by providing 'paved roads'—standardized tools—while allowing innovation through 'unpaved roads.' Tools like Replit and Lovable help users at different skill levels with AI integration.

In this episode, host Rick Watson talks with:

Vinod Kumar, CEO of Syntheum AI and formerly Head of Global Business Insights & Growth at Demandware/Salesforce CommerceCloud. He shares insights on leveraging AI in e-commerce and startups amid rapid technological change. He identifies trust as the top concern, emphasizing the need to earn and maintain it quickly to enable users to offload tasks to AI systems. Technological commoditization shortens the runway for startups, making efficiency metrics like ARR per employee critical. Synthium AI develops agentic merchandising systems that integrate AI into existing workflows to simplify operations and improve margins. Hiring now involves evaluating candidates' proficiency with AI tools.

Within large organizations, balancing entrepreneurial innovators, strategic leaders, and cautious employees is key to adopting AI effectively. Customer demand drives prioritization of AI features. Vinod advises startups to prioritize distribution and trust by building communities and thought leadership before product creation.

In this episode, host Rick Watson talks with:

Alex Sayyah, CEO of Aleran Software, a B2B ecommerce platform tailored for industrial manufacturers, discusses the company's pivot from serving sales agencies to focusing on industrial manufacturers with complex sales needs. Aleran integrates AI tools, such as CPQ, PIM, and order management, atop ERP systems to streamline sales processes.

Alex emphasizes AI's role as an accelerant that enhances efficiency without replacing people, highlighting internal adoption across marketing, sales, and product teams. He notes the importance of triangulating AI-generated information with other sources to ensure accuracy.

The company adopts a pragmatic, empathetic approach to AI, embedding it in customer-facing features like AI shopping assistants and recommendation engines to improve sales effectiveness.

Alex advises a gradual adoption of AI, focusing on practical applications that reduce administrative burdens and enhance customer relationships, while acknowledging the initial challenges and the need for 3ongoing learning.

The discussion emphasizes the importance of embracing digital transformation to meet evolving customer expectations and maintain competitiveness in industrial manufacturing.

In this episode, host Rick Watson talks with:

Vikram Saxena, the Founder of BetterCommerce, a modular, composable B2B e-commerce platform designed for small to medium-sized businesses. Saxena has also worked at Vsworx Information Systems, HCL Technologies, and Axtrum Technologies (formerly BlueThread Solutions).

This episode can be seen as an example of a developer/practitioner using AI to experiment and then build solutions to challenges.

The initial use of ChatGPT to create analytics on product performance and spending and a small proof of concept with a customer.

Vikram discusses how Lovable helps software developers further peel the onion in terms of next steps and the outcome required in building software.

How BetterCommerce stopped using several SaaS solutions for their software development business and replaced them with a custom-built solution, SPDM.

Their AI system categorizes customer feedback into bugs, change requests, or new ideas, researches the value and industry applicability of these ideas, and assists in building internal business cases for product development.

Moving from siloed data found in email, tickets, and customer feedback across multiple solutions to consolidating it all in a single place, enabling greater efficiency.

Their testing of AI use inside their go-to-market strategy, and Vikram's skepticism about using AI in business development and sales. Understanding the gap between AI-adoption for sales and business development staff and moving it to AI-based platforms.

He stresses that AI should not replace the personal touch in sales emails and outreach, as unique personal communication is critical, and AI-generated mass emails often become noise.

He also discusses his enjoyment of experimenting with AI development and building software to optimize cloud costs and subscriptions, which he finds increasingly expensive and difficult to control.

In this episode, host Rick Watson talks with:

Shailesh Kumar, SVP of Engineering at Salesforce, is the Head of Engineering for Salesforce Commerce & Retail Cloud. He previously served as SVP and Head of Engineering at ClickUp and VP of Engineering at MuleSoft, both of which were acquired by Salesforce.

This episode can be seen as an AI primer that will benefit those who listen to this wide-ranging discussion.

Shailesh discusses the opportunity that the model context protocol (MCP) provides developers and, in the future, non-developers. Salesforce currently has a pilot within Commerce Cloud to determine how customers use agents and what agents consumers want to build, alongside vendors.

Shailesh also discusses how software development will evolve as more tools and agents democratize the creation of software. The difference between public large language models and custom-built large language models.

Shailesh distinguishes between deterministic output and non-deterministic flow, which has existed in software engineering and the risks it poses to businesses. He also describes AgentForce, which Salesforce offers partners to develop agents on, and provides pre-built agents to Salesforce customers.

Agentic commerce is unknown, and Shailesh believes that giving partners and merchants the tools will show the future of agentic commerce. He also discusses how agents and platforms can interact with stitching and have guardrails around them to enable merchants to deliver joy to consumers.

The end of silos as agents interact with many different data sources inside businesses, and the speed of development and innovation.