Editor’s note — 16 July 2026: This article was originally published on 26 June 2026 and has been substantially expanded with additional industry commentary from SecurityScorecard, CBTS, Arcova, Cofense, Reolink, Jabra and HCLTech.
As businesses race to integrate artificial intelligence into everyday operations, industry leaders are warning that the real challenge is no longer simply choosing the right AI model. It is ensuring the technology is supported by accurate data, effective governance, strong security and clearly defined outcomes.
The message comes as organisations worldwide mark AI Appreciation Day, with commentary from across the technology industry reflecting a conversation that has become noticeably more grounded. Digital Reviews Network has been following this evolution closely, from enterprise AI deployments through to AI-powered productivity tools entering everyday workflows.
Although the contributors operate across fields including enterprise AI, cybersecurity, home security and workplace collaboration, a consistent theme emerges: the next phase of AI will be determined less by increasingly capable models and more by whether organisations can deploy them responsibly, securely and usefully.
Contents
Better data remains the foundation of enterprise AI
AI data specialist iMerit says successful enterprise AI relies on far more than model performance alone.
“Many enterprises have moved beyond experimentation,” said Sudeep George, Chief Technology Officer at iMerit. “The focus is now on ensuring AI systems remain trustworthy through high-quality data, rigorous evaluation and ongoing human oversight.”
George, who has more than two decades of experience in computer vision, AI systems engineering and multimodal AI, has led the development of iMerit’s AI workflow automation platform, Ango. He is also recognised for his work in human-in-the-loop AI, reinforcement learning and foundation model evaluation.
Enterprise AI success depends on robust data pipelines, continuous model validation and governance rather than simply deploying the latest large language model. Human expertise also remains critical for reinforcement learning, red teaming and identifying edge cases that automated systems may miss. These same principles increasingly extend beyond enterprise infrastructure into consumer products, where AI features are becoming everyday differentiators.
The comments came one day after a landmark announcement for iMerit.
Global data and AI company EXL announced its intention to acquire iMerit in a deal valued at up to US$310 million, subject to customary closing conditions and expected to be completed during the third quarter of 2026.
The acquisition is intended to strengthen EXL’s enterprise AI capabilities by adding iMerit’s expertise in AI model training, evaluation and reinforcement learning.
EXL said the transaction would enhance its ability to deliver production-ready AI systems by combining its industry expertise with iMerit’s specialised AI data operations, foundation-model partnerships and Ango platform.
EXL Chairman and CEO Rohit Kapoor said organisations increasingly require industry-specific data and rigorous evaluation to deploy reliable AI in business-critical workflows. The acquisition is intended to help customers move from AI experimentation into enterprise-scale production.
For iMerit founder and CEO Radha Ramaswami Basu, the acquisition reflects a broader shift occurring across the AI industry.
“We see EXL as an ideal leader in this defining moment for AI,” Basu said. “Both companies share a belief that specialised high-quality data is the foundation of AI success. We are excited to multiply our impact through EXL’s industry expertise, complementary technology and trusted enterprise relationships.”
The combined business is expected to expand EXL’s capabilities across healthcare, financial services, insurance and other regulated industries where dependable AI deployment relies heavily on domain expertise, quality data and continuous evaluation.
From experimentation to AI maturity
While quality data provides the foundation, other technology leaders say organisations also need to understand precisely where AI is being used and what business problem it is intended to solve.
John Bruggeman, vCISO at CBTS, said AI Appreciation Day provides an opportunity to recognise the progress organisations are making across areas including network analysis, security triage, threat detection and customer experiences.
However, he said the next challenge is making that value dependable. “AI assistance is most effective when it is supported by clear governance, valid and reliable data, strong identity controls, and visibility into how tools are being used across the business,” Bruggeman said.
For many organisations, AI adoption is advancing more quickly than the operating model surrounding it. Employees are experimenting with tools and building more efficient workflows in real time, but without approved and practical pathways, that activity can move into the shadows and expose the organisation to additional risk.
Bruggeman said the number of AI pilots underway should not be treated as a meaningful measure of maturity.
“The measure of AI maturity should not be how many pilots are underway. It should be whether the organisation understands where AI is interacting with corporate data, business and customer workflows, user and administrative identities, and real-time decision making.”
Joseph Perry, Cybersecurity Researcher and Advanced Services Lead at Arcova, similarly argued that deploying AI does not itself prove that meaningful transformation has occurred.
“The real test of AI maturity is not how many tools a company has deployed or how often employees use them,” Perry said. “It is whether leaders can clearly explain what problem the technology solves, what it costs at scale, what data it can access and where human judgment is still required.”
Perry said implementation should not begin with a general mandate to use more AI. Instead, organisations should identify a specific process that is slow, repetitive or difficult to scale and assess whether AI can meaningfully improve the outcome.
They must then determine how the technology will fit into existing workflows, who remains accountable for the result and what will happen when the system gets something wrong.
“Transformation is the result of technological potential applied to the solution of human problems,” Perry said.
Organisations also need to look beyond the tools they formally deploy. Employees may already be using public models, vendors may quietly add AI capabilities to existing products, and threat actors are adopting the same technology to increase the speed and volume of their activity.
An organisation’s AI posture is therefore shaped not only by its own technology decisions, but by the way AI enters the business through employees, partners and adversaries.
Sonia Eland, Executive Vice President and Country Manager, Australia and New Zealand at HCLTech, said Australian organisations are now confronting a different challenge.
Rather than deciding whether AI has a place in the business, organisations are increasingly focused on deploying it effectively and responsibly while demonstrating measurable returns.
“The reality is that Australia’s biggest AI challenge is no longer access to technology; it’s execution,” Eland said.
She noted that while enthusiasm for AI remains high, many organisations are still navigating practical issues around governance, workforce capability, data quality and translating AI investments into meaningful business outcomes.
According to Eland, the organisations achieving the strongest results are moving beyond proof-of-concept projects and concentrating on practical applications that improve productivity, enhance customer experiences and allow employees to focus on higher-value work.
“Australia has a significant opportunity ahead, but success will ultimately depend on turning AI ambition into trusted, measurable results at scale.”
MJ Robotham, Director, APAC at NinjaOne, said the discussion around AI has shifted decisively from adoption to operational value.
Rather than debating whether AI belongs in the workplace, many IT teams are now using it to reduce the manual workload associated with managing increasingly complex technology environments.
“Most IT teams in the country are already realising the value of AI by reducing the manual work that slows them down,” Robotham said.
With organisations responsible for growing numbers of endpoints, applications and security risks, often without corresponding increases in resources, AI is increasingly being applied to automate repetitive administrative tasks, improve visibility across IT environments and allow technology teams to focus on more strategic initiatives.
However, Robotham cautioned that successful AI deployment depends on applying the technology deliberately rather than indiscriminately.
“The organisations getting the most from AI aren’t necessarily using the greatest number of AI tools. Rather, they are using AI deliberately to simplify operations, strengthen resilience and enable their IT teams to focus on higher-value work.”
Carla Ramchand, Managing Director for ANZ at Cognizant, said organisations should avoid treating AI as simply another software deployment.
While AI can automate individual tasks, she argues the greatest business benefits come from redesigning entire workflows rather than adding AI to existing processes.
“AI agents are now a part of our teams. They can manage work, make recommendations, and hand decisions back to people when human judgement is needed,” Ramchand said.
However, she noted that AI differs fundamentally from traditional software and cannot simply be installed and expected to understand how an organisation operates.
According to Ramchand, many organisations are widening the gap between AI investment and measurable business outcomes because they are failing to redesign business processes around the technology.
Instead, AI should be built around an organisation’s own processes, data, industry expertise and workforce knowledge, while maintaining clear governance and human accountability.
“The real opportunity with AI lies in looking at the process from the start to finish and rethinking how work should flow,” she said.
Together, the perspectives from CBTS, Arcova, HCLTech, NinjaOne and Cognizant suggest enterprise AI is entering a far more mature phase. Rather than measuring success by the number of AI pilots or tools deployed, organisations are increasingly focusing on redesigning business processes, improving operational efficiency and delivering measurable outcomes through responsible implementation.
Agentic AI expands the digital supply chain
Michael Centrella, Head of Public Policy at SecurityScorecard, said AI is changing not only how cybersecurity teams work, but how organisations think about risk across their broader digital ecosystems.
“Agentic AI is quickly becoming an extension of the vendor landscape,” Centrella said. “That creates new questions about what these tools can access, what actions they can take, and how organisations govern non-deterministic systems.”
AI agents can assist security teams by taking on repetitive and time-consuming work, closing knowledge gaps and allowing people to concentrate on higher-skill and more complex problems.
However, greater autonomy also introduces additional governance challenges. Organisations must understand how AI agents interact with vendors, software providers, partners and digital supply chains, along with the consequences when an agent takes an unexpected action.
Centrella said AI is also compressing both the time required to discover vulnerabilities and the time required to compromise affected systems.
Defenders will increasingly need to use AI to connect signals across suppliers and partners before relatively small exposures develop into larger incidents.
“Responsible AI is not only about innovation,” Centrella said. “It is about using powerful technology to build a more secure, transparent, and resilient digital ecosystem.”
AI is Reshaping Cybersecurity
Chance Caldwell, Senior Director of PDC Threat Services at Cofense, said AI Appreciation Day should also prompt cybersecurity leaders to consider how they will defend against a technology accelerating both innovation and risk. In email security, AI has become a force multiplier for attackers and defenders alike.
Security teams can use it to analyse patterns, automate workflows and scale detection. At the same time, threat actors can use AI to produce more polished, personalised and convincing phishing campaigns at far greater speed.
“The old tells, such as awkward grammar, generic messaging, or obvious errors, are becoming less reliable as AI helps attackers make malicious emails look routine, relevant, and business-like,” Caldwell said.
The growing sophistication of AI-generated content allows malicious campaigns to resemble ordinary workplace communication more closely. Attacks involving brand impersonation, credential phishing, QR codes and trusted-service abuse can also be designed to evade traditional email gateways and reach users directly.
Caldwell said AI-powered security tools remain valuable, but attackers continually develop techniques that may not match established patterns.
The organisations best positioned to respond will combine AI-driven defence with human intelligence.
“Employees who are trained to recognise and report suspicious activity are not the weak link; they are a critical signal source,” Caldwell said.
“The future of cyber defence is not AI replacing people, but AI strengthened by people.”
Aaron Bugal, Field CISO for APJ at Sophos, said AI Appreciation Day should also serve as a reminder that organisations have a responsibility to deploy AI securely.
While AI continues helping defenders analyse threats more quickly, it is also enabling attackers to execute familiar techniques at greater speed and scale.
“AI is changing cybersecurity. It helps defenders analyse threats faster and respond more efficiently, but also gives cybercriminals new ways to scale familiar attacks,” Bugal said.
According to Sophos’ latest ransomware research, identity-based techniques including phishing, malicious email and compromised credentials remain responsible for the vast majority of attacks across the Asia-Pacific region, despite rapid advances in AI.
Rather than replacing traditional attack methods, AI is making them more convincing, more scalable and more difficult to detect.
Bugal said organisations should avoid treating governance, visibility and security as afterthoughts when deploying AI.
Instead, they should be embedded into AI initiatives from the outset.
“The organisations that will benefit most are those using AI to strengthen human expertise, accelerate response and build resilience, rather than assuming technology alone will solve their cybersecurity challenges.”
Privacy and On-Device AI Become Competitive Advantages
While much of the AI conversation focuses on enterprise technology, artificial intelligence is also becoming increasingly important in consumer security products.
Nick Nigro, Vice President Sales Australasia at Reolink, said AI Appreciation Day 2026 arrives at an interesting point for the home security industry.
A year ago, the conversation largely centred on what AI could do, including object detection, false-alarm reduction and intelligent alerts. In 2026, the more important question is whether those capabilities are continuing to work for consumers.
After years of subscription fatigue and cloud-service outages, users increasingly expect security systems to understand context, surface genuinely important activity and protect their data.
“The industry’s broader shift toward on-device processing reflects where those expectations are heading,” Nigro said.
Processing footage locally can enable faster responses while limiting the privacy and security risks associated with sending recordings to cloud infrastructure.
Reolink’s Local AI Video Search follows that approach by running on the device rather than sending footage to external servers for analysis.
Nigro said keeping these capabilities subscription-free is also deliberate, as consumers increasingly expect useful AI features without an ongoing payment plan.
“The category is moving toward something people can genuinely rely on: capable, private by design, and accessible regardless of the hardware they already own,” he said.
Voice could become AI’s next major interface
AI has already begun transforming how people work, but Jabra believes the next major shift will involve how people interact with it.
Ling Lu, Head of New Product Development at Jabra, said speaking with AI at work could become as natural as typing into it is today.
Jabra research suggests voice interaction with AI will reach mainstream adoption by 2028.
Whether summarising meetings, reviewing documents, requesting feedback or brainstorming ideas, speaking with AI is expected to become a more common part of the working day. We’re already seeing this transition begin through tools that integrate directly with generative AI platforms to analyse meetings and conversations.
As AI develops from a productivity tool into a more active collaborator, keyboards are unlikely to disappear, but conversation may become a more intuitive method of interaction, particularly as AI takes on increasingly complex workplace tasks.
However, a voice-first future introduces a challenge already familiar to hybrid workers: poor audio.
“AI is only as effective as what it hears,” Lu said.
If the captured speech is unclear, distorted or overwhelmed by background noise, AI-generated transcriptions, summaries and responses may also be unreliable.
Lu said the next wave of AI innovation will therefore not be defined by smarter models alone.
Technologies that isolate voices, reduce environmental noise and provide consistently clear voice capture will become foundational to the way people work with AI.
“The businesses that recognise this early won’t just improve communication between people,” Lu said. “They’ll unlock more value from AI by ensuring every conversation starts with clear, accurate audio.”
Trust will define AI’s next chapter
Taken together, the commentary provided for AI Appreciation Day shows how the industry’s focus is expanding.
Better data remains essential, but it is only one component of dependable AI. Across the commentary, organisations consistently identified business transformation, operational efficiency, governance, cybersecurity, privacy and measurable business outcomes as the factors that will determine whether AI delivers lasting value.
Organisations also need governance that keeps pace with employee adoption, visibility over where AI interacts with corporate information, strong identity controls, effective cybersecurity and clear accountability when automated systems make mistakes.
In consumer environments, privacy, local processing and freedom from compulsory subscriptions are becoming important measures of value. In the workplace, the effectiveness of voice-driven AI may depend just as heavily on audio capture as the sophistication of the underlying model.
Across the contributions, few industry leaders presented AI as a simple replacement for human expertise.
Instead, they described technology that must be trained, governed, evaluated and supported by people.
AI Appreciation Day may celebrate the progress artificial intelligence has already enabled, but the message for 2026 is increasingly clear: the next phase of AI will not be defined by bigger models alone.
It will be defined by better data, stronger governance, human oversight, effective security and the trust of the people expected to use it.
