Tech Transformed

Think of a robotic prototype that functioned well in March. Its robotic arm sorts through parcels flawlessly on the lab bench. It runs on a power-demanding compute module with a fan attached to the side. Finally, the production specifications are announced: battery power, a sealed case, a budget of 10 watts at most, and no connection to the cloud. Porting the application to hardware that fits those limits is the part of the roadmap nobody put on the slide.

This is where many physical AI projects stall between demo and shipment.

In the recent episode of the Tech Transformed podcast, host John Santaferraro, CEO and Head Research Analyst at Ferraro Consulting, sat down with Krishna Rangasayee, Founder and CEO of SiMa.ai, and Manuel López Roldán, GenAI Product Manager at SiMa.ai.

They break down everything you need to know about physical AI, why it’s taking machine intelligence out of the data centre and into the device itself, and explain the changes that SiMa.ai has made to its software so that AI agents and human engineers can build complex AI applications for that hardware in days and hours, making imagination the limit.

"You could run very high accuracy models entirely on device. You don't need the cloud at all,” the CEO of SiMa.ai states, setting the agenda for this episode.

Rangasayee says that we’re living through what is likely to be the largest upheaval of experiences. “AI is completely overhauling everything we know. We've gone through the social media revolution; now we're going through the cloud AI revolution.”

He believes that everything will be AI-enabled in some form or another, from the infrastructure enabling it in manufacturing, or in robotics, or in automotive, to everything we experience in our home, in our work, in medicine.

Takeaways

• Physical AI transfers reasoning from the cloud to the device itself.
• Rule-based control is replaced by systems aware of context & memory.
• Majority of physical AI customer applications require a sub-10 watt power envelope.
• Smaller & more optimised models operate on-chip with high accuracy.
• SiMa.ai carried out a complete rebuild of its industry-first agentic Palette Software Suite for use by agents.
• Palette lets developers explain and build systems in plain English.
• Current builds now take days or even hours, not "six to nine months".
• Humans retain the ideas & the decisions while the agents execute.
• Current use cases involve inspection, warehouse robots, drones & cars.

Chapters

00:00 Introduction to the podcast and guests

00:28 The move beyond screen-based AI to physical AI

00:58 Defining physical AI and its significance

01:26 Why physical AI will be larger than cloud AI

02:12 The large upheaval AI is causing in various sectors

03:23 The shift from rule-based to reasoning machines

04:22 How models are getting smaller and more efficient

05:54 Enabling new capabilities through AI in physical devices

06:12 Current innovations in drones, robots, and autonomous vehicles

07:35 The role of agentic software and agent engineers

08:59 Democratisation and ease of developing physical AI systems

10:03 Unlocking new business models with physical AI

11:56 The future of AI development and agentic engineering

12:50 The impact of physical AI on industries like manufacturing and healthcare

21:38 How to lead and join the physical AI revolution

23:16 Closing remarks and next steps for listeners

Read the full article: https://em360tech.com/podcasts/what-physical-ai-outgrow-cloud-ai

Watch episode on YouTube: https://youtu.be/GE4sTiBk95A

SiMa.ai: https://sima.ai

EM360Tech: https://em360tech.com

Watch more from the Tech Transformed Podcast: https://www.youtube.com/playlist?list=PLXoSlBGE35NqhHbvDxnrEKw2bVoNBQ8U1

​ ⁨@SiMaTechnologiesInc⁩ ⁨@enterprisemanagement360⁩

#PhysicalAI #AgenticAI #Robotics #TechTransformed #EM360Tech physical AI, SiMa.ai, agentic AI, AI agents, on-device AI, cloud AI, robotics, autonomous vehicles, drones, industrial automation, AI chips, technology, enterprise technology, Tech Transformed, EM360Tech

In the previous episode of the Tech Transformed podcast, EM360Tech's Shubhangi Dua sat down with David Hsu, founder and CEO of Retool, to tackle an uncomfortable truth for CIOs: vibe coding is already happening inside your organisation, whether you've sanctioned it or not.

From customer data turning up in AI-built apps on the public internet, to duplicate tools showing conflicting numbers, to AI bills climbing with no clear view of the return, David made the case that the answer isn't to ban AI-built software but to put guardrails around it. One central platform, with controlled access to trusted data, gives IT visibility, auditability, and control, because "with governance comes freedom." He finished with a bold prediction: by 2030, most of the work in your company, and the CIO's role, will become closer to a chief people officer for AI.

In this Big Data London special edition, host John Santaferraro picks up where that conversation left off with Ryan Wong, Head of Engineering at Retool. As AI moves from generating apps and content to agents that take real action, what does governing them actually take?

Takeaways

• The harness is the layer around the model that a business controls.
• AI harnesses context, identity, policy, action, and observability.
• Governance, not model quality, stops agents from emailing or touching sensitive data.
• Employees connecting Claude or ChatGPT to company data is hard to govern, so centralise access like an API gateway.
• Enforce data policy at the resource, not app by app, so employees can vibe code freely.
• Start agents with tight limits, then widen them. Log every action with the identity behind it.
• Vibe coding made building cheap. Deployment, access control, and observability are now the hard part.
• Retool says its platform has long handled those layers.
• Retool’s Assist tool struggled because it used a proprietary language. Building closer to what LLMs know works better.
• Models will change; the harness will last for years.

Chapters

00:00 Introduction to AI harness and its significance

01:08 Defining the components of an AI harness: context, identity, policy, action, observability

05:12 The role of centralization in AI harness architecture

06:44 Governance challenges in multi-agent AI environments

17:48 Implementing a five-step governance process

22:20 Why focusing on the harness is key for AI progress

27:17 The importance of governance as AI agents gain authority

28:24 Final thoughts on controlling autonomous AI agents

Watch the previous episode with David Hsu, CEO of Retool: https://youtu.be/rvdl4Fy2taw?si=3nMzO...

Learn more about Retool: https://retool.com

More from the Tech Transformed podcast: https://youtube.com/playlist?list=PLX...

#AIHarness #AIGovernance #AgenticAI #EnterpriseAI #Retool #TechTransformed #BigDataLDN #RetoolAI #VibeCoding #AIAgents #AIContext #AutonomousAIAgents #AIROI ‪

@Bigdataldn‬ ‪@Retool‬ ‪@enterprisemanagement360‬

Plenty of organisations have an AI agent pilot that worked well in a demo. Far fewer have hundreds or thousands of agents running reliably in the real world, and that gap is where the hard work sits. AI agents are moving beyond experimentation, but taking them from promising pilots into production at scale is proving to be much more difficult. At the same time, AI is reshaping technical teams, changing traditional career paths and forcing organisations to rethink how technology creates value.

In this episode of Tech Transformed, host Christina Stathopoulos, founder of Dare to Data, explores what that transition looks like in practice. She is joined by Italo Vietro, VP of Engineering at Parloa, and Joshua Burks, Global Managed Services Head at Thoughtworks. They cover platform foundations, workforce change and agent management at scale. Their shared message is that powerful models are only part of the answer.

From Task Execution to Orchestration

Burks opens with a clear observation he has seen, and that is AI is shifting technical teams from task execution towards architectural orchestration. For decades, he says, engineering and operations teams have spent around 75 per cent of their time and capacity clearing tickets, sifting through alerts and writing boilerplate syntax. Agentic AI and automated coding tools are making that kind of syntax generation and routine triage a commodity. So the bottleneck has moved downstream. Burks describes a new kind of team built from T-shaped pods of engineers who manage autonomous agents, validate complex system behaviour and focus on continuous modernisation. The aim is to stop merely keeping the lights on.

The Rise of the Product-Minded Architect

Vietro sees the same pattern at Parloa and has a name for the people at the centre of it, which, as he calls them, product-minded architects. Instead of fixating on the solution itself, they spend more time in the problem space.

On the positive side, he thinks this is good news for engineering. Engineers generally want to solve problems, and understanding the problem has become more important than ever. The practical challenge is how to architect many agents together so they deliver the outcomes you expect. For Vietro, this also means ownership of something. People need to feel they can deliver the solutions they want, with the right outcomes, rather than simply carrying out tasks.

What Value Arbitrage Really Means

Stathopoulos asks Burks to unpack a phrase he used early on, and his answer is one of the most useful parts of the conversation.

For roughly 40 years, he says, managed services has been a race to the bottom on a pricing model built entirely on labour arbitrage. Vendors bill for headcount, hours, utilisation and ticket throughput. The flaw is that the vendor profits from the complexity, manual effort and lingering technical debt in a client’s systems.

Technology arbitrage flips that. AI agents reduce the headcount needed to run operations or development. Burks believes the industry is now heading towards value arbitrage, where platform automation, predictable observability and agentic workflows take on the toil.

The question then changes from how many people or agents are on the account to what the organisation gets back. That means less technical debt, smaller ticket and development backlogs, and systems that are continuously modernised and ready for whatever demands come next, whether internal or external. Burks describes them as systems that never age.

Making The Jump From Pilot to Production

Stathopoulos points out the big difference between an agent working in a controlled playground and deploying hundreds or thousands of them in the real world. Vietro, whose company deploys agents at enormous scale, walks through what the transition involves. Let’s take a look at how he breaks it down:

Start with value, not just cost

Many organisations begin with the savings: fewer people, lower costs. Vietro encourages them to look at value instead. What better service will customers receive, and what problem is the agent solving?

Know how many agents you have

His second priority is observability, and he offers a simple test. Can you say how many agents are running in your organisation right now? If you can, you probably have a reasonable setup. If you can’t, there’s room for improvement. He adds that the question is far from trivial to answer.

Stathopoulos says she suspects many companies don’t realise how many agents are running in the background and can’t count them. Vietro agrees that this matters well beyond performance. Observability also supports security, because you need to know what agents are doing, spot when something goes wrong and steer them back on course.

Test until you trust

The third priority is evaluations and simulations. Parloa runs millions of them across its agents to check that they do their tasks with a given level of confidence. Vietro’s benchmark is demanding. You test until you can give customers the same quality of service an agent provides as they would get from a human.

Once that confidence exists, he says, deploying agents across enterprises becomes easier, as does replacing different work streams within an organisation.

The platform foundations agents need

If pilots work in a playground, production needs infrastructure. Burks says the current moment closely echoes the early days of cloud. Back then, accounts were spun up all over the place, there were no guardrails, and everything went everywhere. Today he sees the same thing in what he calls the “agentic brawl”. His view is that enterprises need a centralised control plane. Thoughtworks has one called AgentWorks. He frames the foundations in three parts:

1. Observability and telemetry. Agents are only as good as the real-time context they receive. Unified observability across logs, metrics and traces lets AI triage agents correlate incidents and carry out root cause analysis, including for other agents across the enterprise.
2. Governed skills. Rather than letting every team write its own isolated prompts, organisations should capture expert architectural knowledge in reusable, version-controlled skills applied across the business.
3. Governance, safety and security. A robust control plane manages model execution costs, token budgets and access controls. It also acts as an LLM-as-a-judge within evaluation loops. Burks says you can’t scale agents unless security and compliance teams have full auditability over every agentic action. Without it, budgets spiral, and security loopholes appear, just as they did in the early cloud years.

Lessons Cloud Already Taught Us

Stathopoulos says all of this gives her a feeling of déjà vu, and both guests agree the cloud era holds real lessons. Burks was in the middle of those migrations. His biggest warning is the lift-and-shift trap. Enterprises moved monolithic, brittle applications straight to the cloud without refactoring them, and ended up with the same technical debt at a different cost point.

The AI version of that mistake is easy to make. If you plug generative AI into fragmented legacy architecture, chaotic data, messy ticketing queues and poor observability, you simply produce bad outcomes faster. His advice is to modernise the underlying operational engine as you adopt the technology.

Vietro says the same thing in different words: work on your fundamentals. AI will magnify whatever it’s given. Put it into an unstructured environment, and it amplifies the bad. Build solid foundations, and it amplifies the good, helping teams deliver better software.

He also stretches the comparison. Cloud was big, but he believes today’s AI shift is closer to the arrival of the web in the 2000s, and may be even bigger. His advice for navigating it is to stay sceptical and curious in equal measure, as the industry did back then.

What getting it right looks like

After a lot of talk about what can go wrong, the conversation ends on the positive side. For Vietro, successful scaling starts with model routing, which Parloa calls its model mix. Each incoming task goes to the model with the best track record for it, based on earlier simulations and evaluations. Success is a good conversation, whether by voice or chat, handled reliably and at lower cost.

He gives a vivid example. A German-language agent serving Germany would use standard Hochdeutsch. In Switzerland, the same task would land better if the agent spoke Swiss German, with the right accent and customer warmth. Stathopoulos recognises the same effect from living in Spain, where accents and dialects vary from region to region.

Burks looks at success from the organisational side. He borrows a classic rule often credited to Jack Welch: make sure your internal rate of change matches or exceeds the rate of change outside. He says Parloa is doing this well. In managed services, he says, getting it right means clients break out of the legacy maintenance trap. The relationship moves from a vendor who reports ticket SLAs and counts to a strategic partner delivering value arbitrage. Success is measured in innovation capacity, continuous estate modernisation, lower cost to serve and guaranteed business outcomes. Thoughtworks calls this continuous operations, sustaining today’s IT while evolving it towards the transformation the client wants. To learn more, visit thoughtworks.com or follow Italo Vietro and Joshua Burks on LinkedIn.

Takeaways

• AI's impact on technical teams and workflows.
• Transitioning from pilot to production for AI agents.
• Platform foundations necessary for scaling AI.
• Security and governance in AI deployment.
• Lessons from cloud migration applicable to AI.

Chapters

00:00 Introduction to AI's impact on enterprise technology

02:09 How AI is shifting technical team roles

04:00 The importance of product-minded architecture

07:03 From pilot to production: key considerations

09:53 Platform foundations for scaling AI agents

15:13 Lessons from cloud migration for AI deployment

17:59 Risks and security in AI systems

21:14 What success looks like in AI scaling

Envision this: A ServiceNow ticket arrives at a big bank, and 50 customers can't log into their accounts. All the dashboards show a red indicator. About 60 to 100 people end up in a war-room meeting and remain there for about two hours. They discover the problem was a slow SQL query in a database ten hops down the stack, a system neither the login team nor their monitoring system had originally identified.

That is the situation which Anish Agarwal, CEO of Traversal, alludes to when explaining why incident response is becoming more difficult and why he believes the solution must be designed differently from the observability tools that most enterprises currently operate.

In the recent episode of the Tech Transformed podcast, Dan Twing, President and COO of Enterprise Management Associates (EMA), sat down with Agarwal, also a professor at Columbia University. They discuss how code laid out by artificial intelligence (AI) is affecting production incidents.

They also talk about what is needed to trace symptoms back to their causes within a portfolio of 5,000 applications, and how businesses decide when to switch from using AI suggestions to adopting autonomous remediation.

In terms of reliability, Agarwal says that the enterprise has to question whether the task was not just initiated but completed. However, it's also possible for a job to finish, yet the answer might not be the one you want.

Key Takeaways

• "Did it run?" no longer proves an AI-built system worked.
• Coding agents are trained on compile-and-run, so they can sideline the real goal.
• Same configuration, three runs, three different results.
• Engineers now write specs; agents write code, so code comprehension drops.
• Testing improves, but failures still happen at the seams between systems.
• Agents need to search far more data than dashboards were built for.
• Traversal's production world model maps entities, connections and live telemetry.
• Customers are using it to validate CMDBs, which was unplanned.
• Causal analysis separates upstream and downstream anomalies from spurious ones.
• Multiple paths converging on one cause builds confidence in a root cause.
• L4 cut a war room from 60–100 people to five to 10.
• Reported MTTR fell from two or three hours to 20–30 minutes.
• Coding agents now call Traversal's MCP to write production-aware code.
• Decide up front when to build in-house versus buy.

Chapters

• 00:00 Introduction to Hybrid Cloud Challenges
• 02:53 The Role of AI in Production Systems
• 06:34 Complexity of Troubleshooting in AI Systems
• 12:03 Causal Analysis in Production Environments
• 18:03 Agentic Systems and Human Collaboration
• 22:56 Framework for Agentic Readiness in Reliability
• 27:03 Future of Self-Driving Production
• 29:55 Practical Steps for Tech Leaders

Visit traversal.com.

agentic SRE, self-driving production, site reliability engineering, AI SRE, AI incident response, root cause analysis, causal AI, MTTR, mean time to resolution, production world model, CMDB, CMDB validation, ServiceNow, observability, AIOps, incident management, Traversal, Anish Agarwal, Dan Twing, Tech Transformed, EM360Tech, Enterprise Management 360, AI-generated code, coding agents, DevOps, enterprise IT

“Vibe coding is already present in your company,” according to David Hsu, CEO of Retool and governance is the only way to secure it.

Let's put this into context. Think of a Fortune 500 company’s CIO who has recently come across a vibe-coded application spotted on the public internet. This app was developed by an employee of the Fortune 500 company in question. The developer obtained a data dump from Salesforce pertaining to customers, then fed this data to an external AI tool not part of the company’s set of AI tools. The employee prompted the tool to host the resulting application on a public URL.

Since it was an external AI tool, a security review couldn’t be carried out, nor could the IT department view it. Without any internal visibility into a tool created by the employee, the actual customer data was live on the open web. It’s a scary situation, a breach of trust, so imagine the CIO’s wrath upon seeing it.

This is why, in the recent episode of the Tech Transformed podcast, host Shubhangi Dua, Podcast Producer and B2B Tech Journalist at EM360Tech, sat down for a conversation with David Hsu, CEO and Founder of Retool. They talk about the repercussions of vibe coding in addition to examining why most enterprises have no conception of how much of it is already taking place within their own enterprise.

The discussion then turns from the immediate issue of governance concerning uncontrolled AI app-building to Hsu's prediction that software engineering as a profession could vanish within two to three years.

“All the CIOs that I speak to are genuinely frightened about vibe coding, and even if you don't want to admit it, vibe coding is already present in your company,” he tells Dua. This is partly why Hsu founded Retool.

Retool secures vibe coding for enterprises. Hsu explains that Retool can be connected to a company’s Snowflake database, for instance or their Databricks data lake. “Users should be able to have access to particular tables only,” he says. Access controls are required at the group level.

“If you're in this particular Okta group, you can only access this table, and it's read-only. Whereas if you're in this other Okta group, you can actually write back to this other Salesforce over here." So one Okta group could be granted read-only access to a particular table, while another group might be given write access to a connected Salesforce instance. All applications that are built on top of that connection automatically inherit the same restrictions.

Once rules of such a nature are put in place, people could then be set free, explaining why he views governance as something that enables freedom rather than something that restricts.

Takeaways

• Vibe coding is already happening inside most enterprises, sanctioned or not
• A Fortune 500 CIO found a vibe-coded app exposed on the public internet
• One customer's AI coding spend hit 30x their original budget estimate
• Two employees independently built duplicate apps that returned different numbers
• Hsu forecasts fully automated coding arriving by late 2027 or early 2028
• By 2030, Hsu estimates 90% of company work will be done by AI employees
• The CIO's future role: a "chief people officer" governing AI agents, not humans

Chapters

00:00 The Rise of No-Code and Low-Code Development

04:52 Governance in the Age of Rapid Development

12:40 Centralisation and Control in Software Development

20:04 The Future of Software Engineering and AI

25:13 The Role of Humans in an AI-Driven World

vibe coding, AI governance, enterprise AI, CIO, AI agents, software engineering, Retool, David Hsu, enterprise software, AI security

AI is reshaping how businesses interact with customers, from handling routine enquiries to delivering more personalised and responsive experiences. As organisations automate more of the customer journey, however, customer support is increasingly being reconsidered not simply as a cost to reduce, but as an important part of how businesses build lasting relationships with their customers.

On this episode of Tech Transformed, host Christina Stathopoulos welcomed back Pranay Jain, CEO and co-founder of Enterprise Bot, to discuss what this shift means in practice. The conversation moves beyond the usual AI hype to consider how businesses can scale automation without losing meaningful human connection, while navigating growing regulatory expectations across Europe and the UK.

Jain has been building EnterpriseBot for almost a decade, since, as he puts it, "artificial intelligence was more artificial than intelligent." This vantage point shapes a conversation that's less about what AI can do and more about how enterprises should think about doing it.

Stop Treating AI as a Cost-Cutting Tool

Jain sees AI as an opportunity to rethink how customer service teams work, using automation to handle routine tasks while giving people more time to focus on interactions where human judgement and empathy matter most. Password resets, invoice lookups and basic status checks can be handled instantly by AI, allowing human agents to spend more time with customers during moments that require greater care, such as a major life event or an insurance claim following an accident.

He points to IKEA as a real-world example of this working. Rather than laying off staff as AI took over routine tasks, the company retrained employees to help customers plan their homes in a more personal, consultative way and saw sales increase as a result. It's a useful reframe that automation isn't about removing people from the equation; it's about repositioning them where their judgment and empathy are still needed.

Jain also flags a metric that trips up a lot of contact centres, which is average handling time. Teams are often rewarded for keeping calls short, but a one-minute call with a 40 per cent resolution rate is worse than a three-minute call that actually solves the problem. His recommendation is to automate the parts of a call that eat time without adding value - identity verification alone can consume up to a quarter of a call, so agents can spend the time they save on getting things resolved properly, not just quickly.

Why AI Needs Guardrails

Given EnterpriseBot’s European roots, regulation naturally comes into the conversation. Jain explains how the EU AI Act, Spain’s new customer service rules and the UK’s FCA expectations are starting to shape what “autonomous” really means in practice. For example, businesses can track sentiment at a general level, but analysing an agent’s tone of voice or emotions crosses into much more sensitive territory. The same applies to financial transactions. An AI agent can’t simply be left to make those decisions on its own. There needs to be a clear workflow and defined controls around what it can and cannot do.

His concern is that a lot of the AI tooling on the market is built with a US-first lens and simply wasn't designed with these boundaries in mind. This creates real exposure for enterprises that adopt a slick-looking solution without checking if it can actually operate within EU or UK rules. Jain sees these regulatory changes as a positive step, particularly as AI becomes more embedded in customer interactions. "I actually think it's very important that you do have the right regulation, and there's a very good reason why these things are there."

This point extends into data sovereignty too. With geopolitical tension pushing some organisations to reconsider public cloud dependence, Jain suggests flexibility- the ability to run on-prem, in a private cloud, or across different regions depending on the customer; this is becoming one of the more undervalued priorities on a CIO's list.

Connected Customer Experience With AI

The conversation closes on something every customer feels but rarely thinks about, which is the disjointedness of switching between voice, chat, email, and messaging with the same company. Jain's push is for a single AI agent that carries context across every channel and every language. He uses his own background in Switzerland, a country with four national languages, to illustrate the scale of the problem. Without a unified system, a company could end up managing a dozen separate AI agents just to cover language and channel combinations.

His final advice focuses on what enterprise leaders can put into practice now. Start by actually listening to your customer data to figure out where automation adds value and where human time matters most. Then be honest about whether building AI in-house is really core to your business, or if it's a distraction from what you're actually good at. As he notes, insurance companies aren't in the business of building voice agents and trying to keep pace with shifting regulation on your own is an expensive, ongoing commitment few teams are set up for.

Jain’s perspective is his focus on making AI useful rather than simply more autonomous. His approach puts the customer experience first, while recognising that trust, human judgment and the right guardrails still matter. For businesses considering what comes next, that balance may be the difference between adopting AI for the sake of it and using it to create a better customer experience. If you would like to learn more, visit enterprisebot.ai or connect with Pranay Jain on LinkedIn.

Takeaways

• The shift from cost centre to strategic asset in customer support.
• How AI can augment human agents and improve emotional connection.
• Regulatory challenges and compliance in AI deployment.
• Data sovereignty and AI sovereignty considerations.
• Building connected, multi-channel customer experiences.
• The importance of governance and safeguards in AI systems.

Chapters

00:00 Introduction to autonomous customer experience and guest Pranay Jain

01:55 Transforming customer support from a cost centre to a strategic asset

02:54 Using AI to handle mundane tasks and enhance human agent capabilities

04:53 Balancing automation with human connection in customer journeys

05:58 Analysing customer journeys to identify emotional impact areas

06:52 Automating verification and wrap-up to improve efficiency

07:55 Regulatory landscape: EU AI Act, UK FCA, and Spain's requirements

08:54 Compliance challenges and designing regulation-friendly AI solutions

10:53 Data sovereignty, data security, and flexible deployment models

13:09 Creating a seamless, multi-channel customer experience

14:09 Ensuring connected AI interactions across voice, chat, and email

15:10 The importance of multilingual AI agents for global support

23:09 Final advice for enterprise leaders: customer-centric, compliant, and scalable AI strategies

For years, voice has been the channel most enterprises have quietly dreaded. The reason stems from it being harder to automate than chat. It's also harder to scale across markets, and far less forgiving if the experience is clumsy. Yet according to Pranay Jain, CEO and Co-founder of Enterprise Bot, that reluctance is exactly why so many businesses are still losing customers to frustrating IVR menus while their competitors move ahead with something smarter.

In a recent episode of Tech Transformed, Jain sat down with Trisha Pillay to explore the changing role of voice AI in customer service. Together they unpacked the technological advances making it possible, right to the practical strategies businesses need to use it effectively.

Jain, who built Enterprise Bot alongside his wife and co-founder after a detour through real estate, explained where agentic voice AI is heading and why treating it as a bolt-on to existing digital channels is a mistake. He pointed out that voice was never really solved; businesses simply learned to live with its limitations. As conversational AI becomes better at understanding intent, maintaining context and taking action rather than reciting scripted responses, the calculus for enterprise AI deployment is now changing.

Why Voice Still Carries the Business

The figures Jain cites indicate the shift is already underway. In much of Europe, somewhere between 60 and 80 per cent of customer requests still arrive by phone, while chat accounts for a sliver of total contact centre volume in markets like Switzerland. This may surprise people who think digital channels have already taken over. They haven't, especially when customers have complex or urgent problems to be attended to.

Jain explains that customers choose their channel based on how important or difficult their issue is. A quick balance check or a password reset happens over chat or an app. But when something feels uncertain, urgent, or emotionally loaded, people still reach for the phone. They want to be heard, not routed through a flowchart. That's why a robotic voice bot can be more damaging than a mediocre chatbot. Customers often turn to voice when they want a more personal interaction, so a cold or scripted response can make them feel ignored and not valued rather than helped.

This is also where the case for treating voice, chat, and email as one connected system rather than separate builds becomes obvious. Jain described meeting a client running a twenty-person team just to maintain their AI stack while only ten people handled actual customer service. This was a sign that channel-by-channel and language-by-language builds spiral out of control fast. A single underlying AI that can operate consistently across languages and touchpoints isn't a nice-to-have, in his view; it's the only version of enterprise AI that scales without collapsing under its own maintenance burden.

Voice AI’s Real Value

Jain says that the interesting thing about voice AI is not just that it sounds natural but what it can actually do. Most companies just use voice AI to answer questions. Jain says that about 80 per cent of customer calls are about transactions, not just information. Customers don't call just to ask questions; they call to get something done, like a refund or a change to their account.

Enterprise Bots work with Generali Switzerland to show the difference between answering questions and actually resolving issues. When the AI could complete transactions on its own, customer satisfaction went up. When the AI couldn't complete transactions, it was still helpful by directing customers to the right person, which saved time and improved accuracy

Scaling AI Starts With Governance

Jain warns that none of this works without taking compliance and data privacy seriously. Companies in services have to follow rules about where they can use generative AI. The EU has rules about what data can be processed and how, and some things, like sentiment tracking, are allowed in some places but not others. Adding in industry rules, the technical build is secondary to getting the governance right.

Language is another challenge. Building a system for the US market might mean supporting English and Spanish. Building for Switzerland means supporting four national languages, plus dialects and regional variants. Jain says that many companies fail because they can build a demo that works in one language but can't make it work in real-world conditions.

His advice for companies considering AI is to start with a pilot, involve IT and security from the beginning and make sure the technology is solving a real business problem. Jain says that successful AI adoption is not about using the advanced technology but about showing that it can deliver real value before scaling it up. If you would like to find out more about this, visit enterprisebot.ai or follow Pranay Jain on LinkedIn.

Takeaways

• The importance of voice AI in customer interactions.
• Multilingual and multi-channel AI support.
• Regulatory and data privacy challenges.
• Transactional vs informational AI use cases.
• Real-world deployment success stories.
• Technical considerations for scalable AI.
• Customer experience and emotional engagement.
• Future trends in voice AI technology.

Chapters

00:00 Introduction to Voice AI and Its Growing Importance

02:57 Why Voice Remains a Key Customer Channel

04:46 The Evolution of Voice AI and Emotional Understanding

07:20 The Role of Voice in Multilingual Europe

09:22 From Answering Questions to Taking Action

11:01 ROI and Business Impact of Voice AI

12:14 Key Factors in Choosing AI Solutions

14:28 Real-World Customer Success with Generali Switzerland

18:04 Handling Multilingual and Dialectal Variations

21:00 Regulatory and Security Considerations

23:42 Common Mistakes and Best Practices in AI Governance

The fact that OpenAI has quickly adopted Temporal, a rapidly expanding AI ecosystem, and has even entered into a partnership with Crystal Palace FC shows that its business strategy is to pursue community-first innovation.

For years, tech enterprises competed against each other on the grounds of innovative features. However, the competition has changed recently. Now the winners are those who build communities.

In the recent episode of the Tech Transformed podcast, host Christina Stathopoulos, Founder of Dare to Data, is joined by Melissa Herrera, Senior Developer Advocate at Temporal, and Les Jackson, Staff Developer Advocate, to discuss the pivotal role of the community in technology.

They further explore how Temporal's open-source philosophy fosters developer engagement, the impact of community feedback on product development, and the significance of partnerships, such as with OpenAI and Crystal Palace. The conversation emphasises the importance of authentic community relationships and the future direction of Temporal, highlighting the need for continuous integration and collaboration with developers.

Takeaways

• Community is a central part of Temporal's growth.
• Temporal's philosophy is rooted in open-source software.
• In-person community interactions are invaluable for developers.
• Feedback from the community directly shapes product direction.
• OpenAI's adoption of Temporal led to significant scaling.
• Partnerships should focus on community engagement, not just transactions.
• Temporal's collaboration with Crystal Palace merges tech and sports communities.
• Investing in community fosters trust and collaboration.
• Continuous integration with developer tools is essential for success.
• Authentic community relationships drive technology innovation.

Chapters

00:00 Introduction to Tech Transformed Podcast

02:45 The Importance of Community in Technology

05:13 Feedback from Developers: Shaping Product Direction

07:51 OpenAI's Adoption of Temporal: A Case Study

10:02 Unique Partnerships: Temporal and Crystal Palace

15:12 Looking Ahead: Future of Temporal and Community Engagement

19:38 Final Thoughts: Investing in Community

Temporal, Enterprise AI, Developer Communities, Developer Advocacy, Open Source, OpenAI, AI Agents, AI Infrastructure, Durable Execution, AI Orchestration, Enterprise Software, Developer Experience, AI Workflows, AI Adoption, Community Led Growth, Developer Led Growth, Temporal SDK, Software Engineering, AI Engineering, Enterprise Technology, Crystal Palace, Tech Transformed

In this day and age, there is probably no meeting or conversation in the tech world that doesn't concentrate on how artificial intelligence (AI) can improve the work streams of organisations worldwide. There will definitely be a moment in almost every AI strategy meeting where someone says, "We need agents."

In this episode of Tech Transformed, Sara Maldon unpacks with host Christina Stathopoulos how she has heard this sentence countless times. As the head of AI and automation at Make, she's learned to treat that sentence as a starting point because most of the time, the business problem sitting underneath it doesn't actually need an agent at all.

This conversation is less about hype and more about when a workflow needs autonomy, and when a simple if-this-then-that rule is doing the job just fine. Maldon's answer comes from two and a half years of running AI transformation inside a company that was automating things long before "agentic" became a buzzword.

Where Agentic AI Fits on the Automation Spectrum

On the other end of the spectrum is agentic AI. Rather than following a fixed sequence of instructions, it gives an LLM a goal, the right tools, and enough context to decide for itself what needs to happen next. Maldon points to a simple example. A car-leasing company needed product images pulled automatically from manufacturers' websites. A traditional scraper did the job well enough until BMW redesigned its website. Overnight, the automation failed because the image had moved. "An agent doesn't care where the picture is on the page," she explains. "It just knows it needs to get the BMW photo."

This indicates that Agentic AI isn't valuable because it's more "intelligent"; it's beneficial because it can adapt when the environment changes instead of breaking down the moment something shifts. Even Make's own AI sales agent isn't fully autonomous. It's deployed across 60 sales representatives, but, as Maldon points out, it's "95 per cent deterministic." Most of the workflow still follows predefined rules. The agent steps in only when a judgment call is needed, deciding a conversation should be escalated, followed up by email, or moved into a Slack channel. The plumbing remains deterministic because the agent simply adds decision-making where it creates the most value.

Building With AI

Maldon believes the people best equipped to build workflows are those who understand the business firsthand. Operations teams, customer success, marketing, and other business users see where processes slow down because they work with them every day.

She knows this because she has been through the same journey herself. Before leading AI and automation at Make, she trained as a lawyer, not a developer. Today, she describes herself as one of the company's most active builders.

For Maldon, technical expertise isn't the defining quality. Domain knowledge comes first, followed by an instinct for improving processes, a desire to solve problems for other people, and enough curiosity and persistence to keep experimenting until something works.

This philosophy shapes how Make develops automation internally. Employees rarely begin inside the automation platform itself. Instead, they sketch ideas in Claude, using it to think through the workflow and refine the logic. If an automation proves useful beyond one or two people, the AI team then helps turn that prototype into something production-ready using AI co-worker, Maia by Make. Starting from scratch is becoming the exception rather than the rule. The first draft already exists; the job is to refine, test, and scale it.

Scaling Automation Across the Business

Here's the part leaders don't want to hear: the hard part isn't the AI. It's everything around it. Maldon points to Make's sales agent again, the one handling escalations, follow-ups, and CRM updates after every call. Building the agent itself took two days. Planning and mapping the deterministic pipeline around it took four weeks. Rolling it out to 60 people, with training, feedback loops, and adoption? Four months.

Patience, she believes, is the least-discussed skill in AI transformation and maybe the most necessary. Not because the technology is slow, but because people aren't robots; they need time to trust a new process before they'll actually use it.

The HR team's onboarding redesign makes the point without needing sales numbers to prove it (though the numbers help - a 10 per cent lift in AI adoption from something as small as a personalised pre-start video). Ninety-six per cent of Make's employees now run an AI agent they built themselves. Not because leadership mandated it, but the tools got low enough friction, and the culture rewarded people for trying. Maldon's advice to leaders stalling on where to start is to stop assessing tools and pick one. Run a hackathon or set aside an afternoon to experiment for a small section of your enterprise. The market moves faster than any evaluation process can keep up with anyway, so get a foot in the door, then figure out the rest as you go. If you would like to find out more, visit make.com or follow Sara Maldon on LinkedIn.

Takeaways

• Organisations should start small and iterate quickly with AI projects.
• Building a culture of builders and curiosity accelerates AI adoption.
• Patience and soft skills are crucial for scaling automation.
• Agentic AI allows for more flexible and resilient workflows.
• Empowering non-technical teams to build with AI democratises innovation.

Chapters

00:00 Introduction to AI transformation and automation at Make

00:57 Meet Sara Maldon and her role at Make

01:59 Understanding the spectrum: deterministic vs agentic AI

02:52 The agentic spectrum and real-world examples

07:04 Building automation in the age of AI

09:54 The rise of builders: no-code and operational roles

12:02 Scaling automation across organisations

16:13 Customer success story

18:58 How Make uses AI internally to innovate

22:01 Advice for leaders starting with AI and automation

24:04 Conclusion and key takeaways from the episode

Every enterprise leader has heard the numbers by now as billions poured into licences, seats and tokens, with productivity gains still trailing well behind the promise. On a recent episode of Tech Transformed, host Christina Stathopoulos sat down with Jay Richman, Chief Product and Technology Officer at Multiverse, to unpack why so many organisations remain stuck in what he calls "tinker mode" and what it actually takes to move past it.

Richman's view is shaped by a career spent at the sharp end of three separate technology shifts. A decade at Spotify saw him build out the company's advertising business and early subscription platform from scratch. Roughly four years at Amazon followed, working on agentic AI systems within the advertising division, effectively rebuilding the old-fashioned creative agency model using specialised generative media tools. Four months ago, he relocated from New York to London to join Multiverse, drawn by a mission he sees as almost the reverse of his previous work. Rather than using people to make machines smarter, the task now is training machines to make people smarter. That framing sits at the centre of this conversation.

The AI Adoption Gap

Richman suggests that the widely reported gap between AI spend and AI-driven output isn't really a technology failure; it's a human one. Tools sit unused, licences go unopened, and organisations struggle to move employees from curious onlookers to confident, habitual users. Multiverse's answer is what it calls the "adoption layer", which is a diagnostic approach that pinpoints where skill gaps actually sit across a workforce, builds tailored learning paths around them, then embeds coaching directly inside the tools people already use, whether that's Claude, Gemini or ChatGPT.

Crucially, Richman distinguishes between what he wryly terms "token maxing" and value creation. Blunt incentives, leaderboards, usage dashboards and mandates to switch from manual coding to AI-assisted development. All this can help people over that first hurdle, and he admits to having used some of these tactics himself. But the real measure of success has to be defined customer by customer, tied to a specific business outcome, not simply how many prompts someone fired off in a week. As he puts it, the value of any given AI initiative can only be judged against what that particular organisation set out to solve in the first place.

Human Skills AI Can't Replace

One of the most interesting parts of the discussion explores how generative and agentic AI are reshaping team structures. Richman describes a flattening of traditional boundaries: product managers writing code, designers contributing to technical documents, and engineers drafting strategy papers. Functions that were once tightly specialised, such as a dedicated user-research team and a standalone copywriting role, are increasingly being absorbed as skills within more generalist, multi-hatted employees. At the same time, he notes, the opposite is happening at the model layer, with AI agents becoming ever more specialised. The result, in his telling, is smaller teams working with far greater autonomy and considerably less coordination overhead than in years past.

Asked what he prioritises when hiring now, Richman is candid that mastery of a single discipline matters to him less than certain behavioural traits: curiosity, high agency, sound judgment. He shares an interview technique he relies on, asking candidates how they spent their time during the pandemic lockdowns, as a rough proxy for whether someone tends to seize initiative or simply wait for circumstances to dictate their next move.

Moving Fast Without Cutting Corners

On responsible deployment, Richman insists culture has to start at the top. Leaders, he argues, need to be visibly hands-on and honest about their own gaps in knowledge rather than issuing mandates from a distance. At Multiverse, that meant lifting caps on AI coding tools, resetting the expectation that new work should be AI-generated by default, and investing heavily in quality assurance and evaluation frameworks so that human reviewers still hold a meaningful checkpoint before anything reaches customers. He contrasts this with the slower, more cautious approach he sees at many other organisations — one he believes is not only less efficient but also less likely to stick once the initial push fades.

Measuring What Actually Matters

Tying this back to the earlier point about value over volume, Richman suggests the real test of an AI initiative is how closely it maps to a business's own stated objective, not adoption figures in isolation. That means starting with what a customer or team is actually trying to achieve, building learning around the specific gaps standing in the way, and only then judging success by whether the resulting behaviour change shows up in the metrics that organisation already cares about.

The conversation closes on a personal note, with Richman sharing the advice he'd give his own children as they weigh careers in a fast-shifting landscape: run towards disruption rather than away from it. Having entered the workforce at the tail end of the dot-com boom, he sees clear parallels with the present moment, a period, in his words, where nobody can credibly claim expertise, which makes it as good a time as any to jump in. For organisations still puzzling over why their AI investment hasn't translated into measurable returns, Richman's message is simple: the technology was never really the difficult part. If you would like to find out more, please visit multiverse.io or connect with Richman on LinkedIn.

Takeaways

• Successful AI adoption hinges on people, not just technology.
• The AI adoption layer helps bridge the gap between investment and productivity.
• High agency and curiosity are key traits of successful AI practitioners.
• Leadership must lead by example and foster a culture of experimentation.
• Rapid technological change requires teams to be flexible and multi-skilled.

Chapters

00:00 Introduction to AI transformation and people-centred approach

00:54 Jay Richman's background at Spotify, Amazon, and Multiverse

03:47 The challenge of AI adoption and pilot mode

07:04 The concept of the AI adoption layer and its purpose

12:54 Distinguishing token maxing from value maxing

16:03 Impact of generative AI on product and engineering teams

19:09 Building teams in the AI era: blurring roles and skills

21:49 Qualities and skills for future AI teams

25:49 Leading responsibly with AI and fostering a culture of experimentation