Integreon Logo
Contact Us

Category:

Shaping a Lasting AI Strategy in a Fast-Changing World

Hero image: A woman in a blazer types on a laptop at a desk in a modern office with colleagues in the background.

In an article by CIO.com, Integreon’s CTO John Wei makes a simple but important point. Nearly every company now has access to the same AI tools, so the tools themselves are no longer what sets a business apart. The advantage comes from how you decide to use them. John compares it to the airline industry. Most airlines fly the same kinds of planes under the same rules, yet some run circles around others on price, service, and reliability. The plane is not the difference. The way the business is built around it is. AI, he argues, is the plane. Your strategy is the route. He explains how the technology has finally settled down enough to plan around. AI can now take on much bigger jobs in one go, like reviewing an entire contract or a full set of code at once. The cost of running it has stopped swinging wildly, so companies can budget for it like any other expense. And the market is narrowing to a handful of serious providers, which makes choosing tools less of a gamble. One point stands out. Despite all the talk about AI replacing people, John points to evidence that it is mostly being used to support human work, not take it over. The hard parts of a job, like judgment, catching mistakes, and knowing when something looks wrong, still need a person. Most companies are using AI to grow and try new things rather than to cut staff. From there, John turns to the questions leaders actually need to answer. Not “which tool should we buy,” but “where is the business going, and how does AI help us get there?” Who do we want to become? What are we choosing not to do? Which parts of our business might AI shake up, and are we honest enough to admit it? He closes with a warning about two traps: refusing to change out of old habits, and chasing every new tool just because it is new. The companies that pull ahead will be the ones that pick a direction and stay with it. John lays out the full argument, along with the complete list of questions every leader should be asking, in his article for CIO.com: Shaping a lasting AI strategy in a fast-changing world

Everyone Claims, “AI Document Review.” No One Means the Same Thing in eDiscovery

Hero Image: Woman at a desk using a computer with a document-review interface

“AI enabled document review” has fast become the most overused and least defined phrase in the industry. With promises of faster timelines, substantial cost savings, and maintained accuracy, it’s a hot topic for a reason. Alternative Legal Service Providers (ALSPs), eDiscovery companies, and many law firms claim to offer “AI Review.” On the surface it sounds like a shared capability. The reality is that the phrase has increasingly become a catch-all marketing label used to signal innovation, attract attention, and capture market share, despite often describing fundamentally different workflows, technologies, and levels of human involvement. Legal teams should absolutely be embracing AI. As data volumes continue to increase alongside mounting pressure on legal budgets, AI plays an essential role in modern document review and eDiscovery workflows. However, when the same phrase is used to describe everything from light-touch document prioritization to deeply integrated AI led review models, meaningful comparisons between providers become increasingly difficult. When every proposal relies on the same language and buzzwords, the operational nuances that actually determine cost, risk, and defensibility are often lost. The result is a growing gap between what legal teams believe they are purchasing and how document review is ultimately being delivered. AI is Only as Effective as the Human Judgement Behind It One of the most significant differences between these models is the checks and balances integrated into the workflow alongside the quality and role of the “human-in-the-loop.” AI must be trained, its decisions must be calibrated, and its output must be validated if the final work product is to be accurate and defensible. At one end of the spectrum, AI is used to support large scale human review workflows by prioritizing documents, surfacing potentially relevant material, or accelerating reviewer decisions. At the other end, some providers are moving toward highly autonomous workflows with minimal human intervention beyond exception handling or final quality control. Neither approach is wrong, both have value depending on the circumstances of the matter in hand. The issue is they carry very different implications for cost, speed, defensibility and risk. The appropriate model depends heavily on the nature of the matter and the risk appetite of the legal team. A lower risk internal investigation may be well suited to a highly automated workflow designed for cost efficiency and speed. A high-stakes regulatory investigation or litigation may require significantly greater human effort and judgement. This is where the quality and placement of the “human-in-the-loop” becomes critical. The key distinction is how human judgment is incorporated into the workflow, who is making those decisions, and how the outputs are validated. The sophistication of the technology matters but understanding how the workflow is designed around risk and defensibility matters more. Technology is only ever as good as the process, workflow, and human judgement behind it. The Workflow Matters More Than the Label There is no single way AI is integrated into document review workflows. Providers may use similar language while deploying fundamentally different workflows behind the scenes. These differences extend beyond the technology itself. Operational differences change how work gets done, how human judgment is applied and how results are defended. In some workflows, AI is used to accelerate early stage review, helping teams identify relevant documents or prioritize what should be reviewed first or reducing the volume requiring human assessment. In others, it’s embedded more deeply into the process, influencing substantive review decisions, escalation paths, validation protocols and how outputs are ultimately validated. As a result, two providers describing their offering as “AI Review” may in practice be delivering vastly different levels of oversight, accountability and auditability. This may in turn impact accuracy, defensibility and how legal teams evaluate the work product. That distinction becomes particularly important when facing a challenge from opposing counsel, regulators or the court. A provider relying on automation without human oversight may struggle to explain the logic behind a production. In contrast, models that integrate human judgment at key intervals provide a much stronger narrative for the process, turning a technical output into a defensible legal position. The question is no longer simply whether AI was used. It is how the workflow was designed, executed, governed and documented. The Risk of Assuming Equivalence Highly automated review models may appear more cost effective on paper because they reduce the number of human review hours involved in the process. In the right circumstances, that efficiency can deliver significant value. However, lower upfront review costs do not automatically translate into lower overall risk or lower total cost. Treating different models as equivalent creates significant risk. Two providers may both use the label “AI enabled review” while offering completely different levels of automation, validation and documentation. Teams often fall back on surface level comparisons or guess what the service actually does, leading to “expectation gaps” where a legal team assumes a level of human oversight that isn’t actually part of the provider’s workflow. If an overly aggressive workflow results in over production, inconsistent privilege determinations, missed context or the need for substantial downstream re-review, the initial efficiency gains can quickly erode. In some matters, the operational cost of correcting errors or defending review decisions may outweigh the savings generated by increased automation. This is why legal teams can no longer evaluate document review models based solely on speed, reviewer counts or headline cost reductions. Selecting the appropriate workflow requires a clear understanding of how AI is being applied and how human judgement is being integrated: How is AI used at each stage of the review? Where does human oversight sit within the process? How are decisions validated and documented? What audit mechanisms are in place? How does the process stand up under scrutiny? These are not new considerations in document review but they are more important now that AI is at the centre of the work. Transparency, governance and defensibility are more critical than ever. Beyond the Labels Legal teams need to drill down into what “AI Review” actually means by asking more detailed questions about how AI

Legal AI Adoption Tips And Takeaways From Dot-Com Bubble

Hero image: A woman in a blazer typing on a laptop at a desk with holographic charts.

In the late 1990s and early 2000s, the dot-com bubble was in full swing. The internet exploded, the stock market boomed, and then came the market correction and the fall of many internet-focused companies. At the time, I was serving in a senior legal role for Hewlett-Packard, which gave me a unique vantage point from the legal strategy and business perspectives. Today, those of us who worked in legal and technology during that time are experiencing an undeniable sense of déjà vu. The current investment frenzy surrounding legal artificial intelligence looks remarkably similar to the dot-com bubble. So, what happens when capital allocation overpivots during the emergence of transformative technologies? Every massive technology cycle, just like this one in legal tech, creates legitimate long-term winners, irrational capital allocation and secondary distortions across the broader ecosystem. By examining past market behaviors, corporate legal functions and law firms can navigate the current environment without falling victim to market turbulence. Modern Indicators A mature market typically follows a steady, calculated progression of growth. In the current legal AI market, however, we’re seeing clear characteristics of speculative bubble behavior, as valuations are far outweighing traditional revenue metrics. For instance, Harvey AI reached a reported $3 billion valuation in 2025, with subsequent funding indications suggesting a valuation in excess of $11 billion, all on fairly modest revenue.[1] In March, Legora announced that it raised capital at a valuation north of $5 billion.[2] Meanwhile, generalist AI developers are rapidly entering legal, as highlighted by Anthropic rolling out its Claude for Legal solution[3] alongside a corporate valuation of $965 billion.[4] Beyond the billions in valuations, marketing strategies reflect classic bubble behavior. Legal AI vendors are producing Hollywood-caliber marketing campaigns — featuring A-list celebrities such as Jude Law[5] — to try to court law firm partners and chief legal officers. This mirrors the time when enterprise software-as-a-service companies wooed chief information officers with stadium concerts and open bars. When marketing budgets and valuations outpace foundational enterprise integration, market correction usually follows. Historical Echoes The fact that the dot-com bubble burst did not invalidate the transformative effect of the internet. The internet fundamentally changed how global commerce is transacted and how information is consumed. The bubble was a product of timing, inflated valuations and unrealistic adoption timelines. Capital simply flooded the market ahead of sustainable economics. When the Nasdaq peaked in March 2000 and subsequently lost roughly 78% of its value by 2002, the financial carnage was severe.[6] Yet, companies like Amazon.com Inc., Google LLC and Salesforce.com Inc. grew to be industry titans. On the other hand, companies like Pets.com and Webvan experienced rapid collapse. Pets.com became the ultimate symbol of speculative excess by spending heavily on highprofile branding before reaching maturity and sustainability.[7] Webvan collapsed because infrastructure costs and premature adoption assumptions outpaced market maturity, even though its core thesis of online grocery delivery became mainstream 20 years later.[8] A similar dynamic is playing out today. AI is genuinely transforming how content is created, managed and monetized across the global legal landscape. It’s automating manual processes that historically required human intervention due to poor workflows, infrastructure silos or bad data quality. The massive capital entering legal AI reflects the expectation that technology will reshape the delivery models of a $1 trillion industry.[9] However, just as in the early stages of the internet boom, the market will inevitably yield many casualties and only a few scaled winners. Likely Winners and Losers The Platform Contenders A fierce competition is underway to become the industry’s foundational legal AI platform. Legal-focused platforms like Harvey AI, Legora, CoCounsel and Litera are fighting for market dominance, and they’re facing competition from established enterprise software giants like Microsoft Corp., ServiceNow Inc. and Workday Inc., which aim to be the central hub connecting and managing automated AI tasks across the whole company. Concurrently, generalist large language model providers are entering industry verticals to avoid commoditization, as seen with Anthropic and OpenAI. There will be a few scaled winners of the platform race that can consolidate market share. This consolidation can also accelerate from the merging or acquisition of competitors. Point Solutions Point solution vendors are focusing on narrow use cases like contract creation, redlining and document review, or deep industry verticals like clinical trial agreements. To be successful, they must be exceptionally easy to use, execute targeted tasks well and avoid being absorbed by larger platforms. Many point solutions will struggle to scale quickly enough to sustain operations in an ecosystem changing at an exponential rate, and will likely experience the greatest initial fallout. Infrastructure Providers During the dot-com era, infrastructure providers like Cisco Systems Inc., Oracle Corp. and Akamai Technologies Inc. experienced massive growth. Even when individual internet startups failed, the underlying demand for networking, databases and hosting exploded. The massive overinvestment in telecom infrastructure and data centers was eventually absorbed by the market. Today, hardware and infrastructure providers like Nvidia Corp., Microsoft and Amazon Web Services are reaping the rewards of the AI boom; they remain positioned for growth regardless of which individual legal AI applications survive. The Advisory Ecosystem Every technology bubble creates a secondary economy of consultants, integrators, trainers and transformation advisers competing to monetize the transition. In the late 1990s, traditional consulting firms like Accenture PLC, KPMG Consulting and Razorfish rapidly repositioned themselves as internet transformation advisers. Today, major service providers, global consultancies and alternative legal service providers are similarly pivoting to lead AI enablement efforts. The Real Gaps: Why Workflow Lags Behind Hype A major challenge that comes with any transformative technology is the gap between marketing hype and the realities of deployment. In the late 1990s, many people believed that brick-and-mortar stores would close, newspapers would cease to be published and businesses would digitize instantly. Sure, the technology worked, but organizational change took years to fully take effect. Within the legal sector, several systemic frictions are slowing widespread adoption: Economic models: Legal AI pressures traditional law firm billable hour structures, requiring a nuanced evolution of firm-client economic relationships.

How to Escape Pilot Purgatory and More Efficiently Roll Out AI

Hero Image: Integreon branded editorial content.

The legal industry is currently witnessing a massive divergence between the speed of technological evolution and the reality of departmental adoption. Technology is simply advancing faster than legal departments can effectively implement it. The next wave – agentic AI – promises to automate and execute multi-step workflows and change the nature of in-house legal work end-to-end. The problem is that 52% of legal departments are still in pilots with generative AI, with fewer departments (23%) saying they are fully deployed and operational, according to the latest Blickstein Group Law Department Operations (LDO) survey. Meanwhile, 20% of legal departments are actively deploying AI agents for contract negotiation, with another 40% exploring the technology. Given the rapid pace of technology change, legal departments need a strategy for assessing, piloting and rolling out – or moving on – quickly and effectively. The Middle-Mile Problem Why do pilots stall? The root cause is often the “middle mile” problem. Organizations frequently treat AI as a standard software purchase or a tool to be installed rather than a fundamental workforce transformation. This traditional approach is failing for three reasons: Fragmentation: An overwhelming number of providers has created a tech stack that cannot communicate with itself. Safety Brakes: Security and accuracy concerns remain a significant hurdle for 44% of legal professionals, acting as a permanent brake on deployment. The Shiny Object Trap: The rush to explore agentic models without a foundational data strategy leads to pilots that look good in a vacuum but fail in production. To bridge the gap between piloting and deployment, legal leaders need a new roadmap. A Roadmap to Escape Pilot Purgatory 1.- Establish a Rigorous Assessment Framework With a flood of new technologies hitting the market, it is easy to get lost chasing the latest flashy feature. Successful departments apply the discipline of an objective framework: survey the market, prioritize high-value use cases, and apply consistent criteria to every tool. If a tool doesn’t meet the threshold, move on quickly. 2.- Accept a Continuing Work-in-Progress The legal tech stack will never be complete. Many teams fall into paralysis by analysis, waiting for the perfect, all-in-one solution. In reality, decisions must be made to solve today’s challenges, with the understanding that these tools may be re-evaluated or replaced in two to three years as the landscape shifts. 3.- Leverage Trusted Partners to Shortlist You don’t have to vet every tool yourself. Law firms and Legal Service Providers (LSPs) are exposed to a wide array of technologies across different clients. Leverage their experience to shortlist vendors. Often, a provider can run a pilot within your environment or conduct a joint proof-of-concept, saving your internal team hundreds of hours in the vetting phase. 4.- Define Success with Go/No-Go Timelines Every pilot must have a shelf life. Before a single login is created, establish what success looks like using quantifiable metrics. Set a strict period for testing and reporting. At the end of that period, there are only two options: Go or no-go. 5.- Prioritize Quantifiable ROI Use Cases Focus on pilots that directly reduce outside counsel spend or internal hours on high-volume tasks. If you cannot establish a current baseline or demonstrate time savings against your current state, do not waste time piloting it. For example, rather than exploring AI, map out a specific process like M&A due diligence where agents handle the first 80% of the heavy lifting. 6.- Empower Change Champions Technology doesn’t fail; adoption does. Identify change champions within your team – those who are naturally tech-curious – to lead the pilot. Their enthusiasm generates the internal momentum necessary to push a tool past the pilot phase and into the daily workflow of the rest of the department. Agentic Adoption Realities The most common reason pilots fail is a misalignment of expectations. Many leaders expect 100% automation and are disappointed when a human still needs to intervene. Setting reasonable expectations for agentic adoption requires mapping out the process and understanding the priority areas where agents will offer the most lift and where humans will supervise most effectively. Success looks like agentic AI handling the first 80% of the work, leaving the final, high-value 20% for human experts. In addition, the coming wave of agentic AI requires legal departments to calibrate how they are thinking about data and workflows. Consider piloting a data cleanup agent. The most sophisticated LLM in the world is useless if your contracts are trapped in disorganized repositories or inconsistent formats. By focusing on the data layer first, you ensure your legal AI has a clean foundation to act upon. The 2026 Deadline Many legal departments are still in pilot mode, but we are rapidly approaching a tipping point. By the end of 2026, the gap between piloters and deployers will become a significant competitive disadvantage. Those who have built an AI-enabled workforce will operate at a speed and cost-basis that traditional departments simply cannot match. The goal for the modern GC isn’t to have the most expensive or best AI; it’s to have the best AI-enabled workforce. To get there, you have to stop piloting and start flying. Vice President, Enterprise Solutions lntegreon About the author Scott Bien, J.D., is Vice President, Enterprise Solutions, at Integreon. He earned his law degree from Northern Kentucky University, Salmon P. Chase College of Law.