The new hot job in Israeli high-tech: AI managers receiving up to 90,000 shekels per month
High-tech companies are appointing dedicated AI managers to lead cross-company change: from development to operations and recruitment. The market is competing for managers who connect strategy to business value with salaries reaching up to 90,000 shekels per month, even though practical experience with autonomous agents is still rare. Senior AI managers and recruitment experts explain what the role looks like on a day-to-day basis and how to get into it.

The waves of layoffs in high-tech continue, and companies, as happened with Pantera this week, admit that artificial intelligence and the changing nature of work play a significant role in this. If you have looked at job postings in high-tech recently, you have surely noticed that 'AI' is attached to almost every role. But what is the most senior role in the field? Meet the VP AI, Vice President of Artificial Intelligence. High-tech companies in Israel are appointing dedicated AI managers to lead cross-company change: from product and development to customer service, operations, recruitment, and payroll.
Ziv Peled, Chief AI and Customer Officer at AppsFlyer, took on the CAIO role in the last year after more than 12 years at the company. He joined AppsFlyer in October 2013 as the seventh employee and built the Customer Success organization from two employees to a global organization of about 180 people in more than 20 offices worldwide. Along the way, he also served as VP Product for two years, until the end of 2019. Today, he holds two roles simultaneously: leading the AI transformation in the company and responsibility for the Customer Success organization.
'The mandate I received from the CEO, Oren Kaniel, is to lead the AI transformation throughout the company, with the goal of becoming an AI Native company,' Peled tells mako.
As part of his role, Peled centralizes all of the company's operational and business activity under him — areas of responsibility previously managed by the Chief Operating Officer (COO). This includes managing the end-to-end activity chain, from the stage of establishing contact with the customer to the actual receipt of payment, alongside the marketing and sales, finance, and computing infrastructure optimization departments using artificial intelligence (AIOps). His day-to-day work focuses on developing autonomous AI agents and automation capabilities adapted for employees, establishing internal platforms, and training 'AI Ambassadors' who will lead the change in every department. Simultaneously, he is responsible for evaluating suppliers and new tools in the market, measuring the degree of use and productivity of the tools in the company, and redesigning management workflows.
Peled says that AppsFlyer implemented the enterprise version of Claude (Claude Enterprise) as the company's main work tool, a move that started with about 600 marketing and sales personnel and subsequently expanded to all employees. The practical use of the tool is reflected in an internal customer recognition system that aggregates information from 16 data sources and provides sales personnel with a focused business snapshot before meetings, alongside daily information summaries that are automatically generated for managers. In addition, every employee in the organization can independently run automated processes and pre-defined capabilities — moves that, according to the company, significantly shorten preparation time for meetings, improve the quality of decision-making, and accelerate response times to customers.
To keep up with new tools, Peled describes a combination of field research, supplier evaluations, internal benchmarks, and an AI Champions community that brings knowledge from the field. According to him, he recently returned from a trip to the USA that included about 30 structured meetings with AI Native companies and strategic customers. 'Are we keeping up? It's a race, but our approach is not to chase every tool but to build an organizational capability to learn quickly,' he says.
To the question of how AI affects recruitment, Peled says that AppsFlyer is still hiring and has not reduced staff this year. What has changed, according to him, is the thinking process before opening a position: can AI answer the need, and what human profile is really required for the role. 'Our goal is not to replace people with AI but to grow smarter, so that every employee does more, with more impact,' he says.
Dr. Naomi Onkolos-Spiegel, from the Software Engineering Department at the Braude Academic College of Engineering in Karmiel, explains that the very appearance of these roles is an acknowledgment that implementing AI in an organization is not a one-time technological project. 'Today, AI is no longer just a technical auxiliary tool; it is becoming a companion, a partner that accompanies the employee both in the personal space, in decision-making and thinking, and in the team space, in collaboration and joint work,' she says.
According to her, the main challenge in implementing AI is not choosing the model but adapting the uses for different people in the organization. A senior manager, a developer, and a customer service representative need different levels of access, adapted prompts, and different use scenarios. 'Without this differentiation, any implementation encounters resistance, overt or quiet, or superficial use,' says Onkolos-Spiegel. A successful VP AI is one who knows how to map organizational requirements, identify how AI as a personal companion integrates with team functioning, and build adapted implementation paths instead of a uniform approach that fits no one.
Salaries are also in line. Nitzan Ron, CEO of MostWanted, which specializes in locating and recruiting senior executives, says that the VP AI role marks the new stage of the artificial intelligence revolution. According to him, if two or three years ago companies mainly looked for AI experts and innovation managers, today more organizations are looking for managers who know how to connect strategy, product, data, development, and management, and turn AI into a business advantage.
Salaries are also in line. According to Ron, in Enterprise companies, compensation packages for VP AI roles usually range from 75,000 to 90,000 shekels gross per month, in Scale-up companies from 58,000 to 72,000 shekels, and in startups from 46,000 to 65,000 shekels, alongside options, shares, and bonuses. However, according to him, in the world of senior executives, salary tables are mainly background noise: the final package is determined according to the candidate's salary and compensation structure in their last role, and according to the strategic value they are expected to bring to the organization.
'The most sought-after candidates are not necessarily AI researchers or Machine Learning personnel,' says Ron. In most cases, these are senior managers who grew up from the worlds of development, product, data, or technology, and in recent years have led significant implementations of artificial intelligence in organizations. Ultimately, companies are looking less for a technology expert and more for a leader who knows how to harness people, build a strategy, and turn AI capabilities into measurable business value.
Lital Yaron, CEO of iLeadx, says that in recent months there has been high demand for roles such as AI product development managers, VPs of Product, and VP AI, alongside an increase in salary ranges. According to her, in Israel, these roles currently range around 60,000-65,000 shekels per month, and in companies that also operate in the international arena, offers are approaching 75,000-85,000 shekels.
Yaron says that in almost all of these jobs, there is currently a requirement for experience in Agentic AI, meaning autonomous AI agents. Companies are looking for product managers who know how to connect APIs, lead initial pilots, manage risks, and understand system architecture. However, according to her, only about 15% of companies have so far reached full implementation of agentic AI in production, and most of them are still in pilot stages. Those who are already seeing a real return on investment see it mainly in streamlining internal processes and cost savings, while the commercialization of the technology to end customers is still in its infancy.
'Companies that are looking for five years of concrete experience are turning to a market where it does not yet exist,' says Yaron. According to her, smart companies look at product managers who come from the worlds of Internal Tooling, automation, and DevTools, and evaluate candidates also by potential and direction of development, not just by exact fit for every item in the job description.
Shlomit Lavin, VP AI at Navina, came to the role from the worlds of research, with a background that combines behavioral sciences and computer sciences, and subsequently Data Science and AI product development. According to her, for the last 15 years, she has been in management and executive roles in the fields of research, Data Science, and AI. She joined Navina in November 2025 as VP AI, after years of managing Research and Data Science organizations and developing AI products in technology companies.
Lavin says that the role at Navina was built from the start with two hats: leading innovation and AI in the product, alongside leading the change in how the company itself uses AI. According to her, this connection was born from the understanding of Shi, the company's CTO and founder, that in order to develop good AI products, AI researchers and engineers must also be sophisticated users of the technology and understand how it changes work processes, decision-making, and product definition.
'The simple way to describe the role is that it connects two axes: AI in the product and AI within the company,' says Lavin. On the product axis, Navina uses AI to process large amounts of scattered medical information and turn it into an accessible, relevant, and understandable clinical picture for the doctor. Within the company, according to her, AI is used in software development, research, data analysis, meeting summaries, document preparation, customer service, and internal work processes.
Lavin emphasizes that in medicine, the goal is not for a model to make a medical decision instead of the doctor, but to provide them with relevant information, highlight important connections, and allow them to dedicate more time to the judgment itself. Therefore, a large part of her role is dedicated to measurement and supervision mechanisms: how to evaluate quality, what requires human approval, where more autonomy can be allowed, and how to connect the use of AI to the company's business goals.
At Navina, they also established an AI Champions group with representatives from different departments, which tests new tools, shares knowledge, identifies relevant uses, and helps employees learn and achieve business goals through AI. In addition, the company is developing roles and processes around the transformation, including AI Transformation Lead, intended to help teams actually change the way they work.
One of the directions Navina is currently investing in is building infrastructure for agents and automated processes, alongside human supervision at critical junctions. Lavin describes, for example, a future scenario where a malfunction coming from a customer is automatically translated into a development task, undergoes investigation, receives a code fix proposal, and is tested before a new version is released. According to her, the company does not yet allow such an end-to-end process without control, but has already streamlined parts of it, and the goal is to gradually connect the stages while maintaining human approval in the right places.
'The most significant improvement is the ability to get faster from a description of a need to an initial product that can be tested, improved, and implemented,' says Lavin. Instead of starting every task from a blank page, employees can use AI to generate a draft, analyze data, suggest solutions, write initial code, or summarize complex material. According to her, human performance does not disappear, but its center of gravity changes: as AI takes more of the technical execution, employees are required to be better at defining the problem, understanding the context, communication, and criticism of the result.
At the cyber company Cycode, which recently appointed a VP AI & ML, they also see the role as part of the change in the management echelon. Mor Rosenstein, the company's HR manager, says that the revolution does not rely only on advanced algorithms.
'Placing professional and technological leaders in the AI field at the senior management level is no longer a privilege, but a strategic necessity that reflects a commitment to building an independent and advanced research arm at the highest global level,' she says. According to her, the role requires a combination of research and technological depth with leadership and a culture of innovation, and affects the product, the team, and the entire company.
Roni Gurevich, VP AI&ML at Cycode, adds that his role is to lead the company's transition from an AppSec platform that provides tools for security personnel to a platform where AI agents perform an increasing part of the work for them. According to him, security teams still define the policy and boundaries, but part of the process is already moving to automation.
'In practice, security personnel are forced to go over thousands of findings, understand what is really important, and chase development teams to ensure they are addressed,' says Gurevich. According to him, AI agents can check if a vulnerability is really exploitable, prioritize it accordingly, and in appropriate cases also generate a fix and open a Pull Request.
Gurevich says that the AI team at Cycode built an infrastructure that allows product teams to develop agents, give them tools to perform tasks independently, and build feedback mechanisms that improve their performance over time. Alongside this, according to him, safety mechanisms were developed intended to ensure that the agents' answers are based, verifiable, and rely on the correct information.
Gurevich comes from a background of Machine Learning and Data. According to him, for more than a decade he has been building systems that take millions of images and texts and extract from them structured information that can be used for decision-making. Among other things, he dealt with understanding products from the images and text that describe them, first at Donde, which was acquired by Shopify, and subsequently at Shopify itself.
'The problem I am working on has not changed much over the years: to turn unstructured information into understanding and decisions,' says Gurevich. The tools, on the other hand, have changed dramatically, and in recent years the pace and scope of the change are something we have not seen before.
Gurevich warns that an unorganized integration of AI into products might create three main problems: sensitive information that flows to places that should not receive it, unfounded answers that are presented to the customer as if they were a fact, and AI costs that are not managed efficiently and might get out of control. On the other hand, according to him, when the company defines a common language and provides teams with infrastructure for correct building, every team can focus on the value it wants to give the customer and on the correct use of AI to achieve it.





