Be a part of our every day and weekly newsletters for the most recent updates and unique content material on industry-leading AI protection. Study Extra
Consulting corporations’ accelerating adoption of generative AI to automate information work is sending shockwaves throughout the {industry}, triggering workforce shakeups and layoffs.
This month, PwC lower roughly 2% of its U.S. workers, roughly 1,500 jobs, in audit and tax traces. EY eradicated 150 roles even because it introduced a $1.4 billion funding to construct an enterprise AI platform. Accenture slashed 19,000 positions (2.5% of its workforce) amid slowing progress and rising tech prices in 2023.
McKinsey & Firm is reportedly paying senior workers as much as 9 months’ wage to stop voluntarily, a technique {industry} observers hyperlink to a downturn in consulting spending accelerated by AI-driven change.
KPMG is realigning expertise as AI platforms change components of the audit course of and different routine duties. In November of final yr, 333 jobs had been lower, or roughly 4% of U.S. audit staff.
This escalating wave of AI-driven layoffs is mirrored in IBM CEO Arvind Krishna’s current acknowledgment that IBM has changed a number of hundred routine human sources roles with AI brokers. Krishna’s feedback underscore the unsettling actuality confronting high-performing staff: roles centered round “rote course of work” are quickly turning into out of date. Though IBM has reallocated some sources towards roles in software program growth and gross sales, the underlying message is evident. Workers more and more understand AI brokers as existential threats, fueling nervousness and driving many to construct shadow AI apps to protect their relevance defensively.
The underside line is that gen AI is redefining all types of information work a lot quicker than anybody, together with the {industry}’s elite consultants, leaders and companions, anticipated.
AI layoffs are sparking a survival mindset
Fearing they might be caught up in sweeping layoffs pushed by AI and automation, most of the {industry}’s elite consultants and excessive performers are reinventing themselves rapidly earlier than their roles vanish.
Groups within the hardest hit areas usually have dozens of shadow AI apps designed to enhance effectivity and workforce productiveness. Proposal and pitch automation, operations and workflow automation, monetary modeling, state of affairs evaluation, and consumer relationship managers being changed by firm-specific copilots are the place shadow AI prospers. Many are counting on Python-based shadow AI to construct customized automation instruments, bypass inner IT bottlenecks, and ship quicker, differentiated insights that shield their roles in an {industry} beneath stress from gen AI.
Python is turning into the language of reinvention
VentureBeat has realized that most of the top-tier strategists, entrepreneurs, practice leaders and their groups have gotten proficient in creating Python-based apps that may take evaluation and insights past the present genAI instruments supplied by IT. Groups creating these apps are proficient with Open AI, Google programmable search engines like google and yahoo, Google Gemini 2.5 Professional, Perplexity and different AI platforms’ API keys and calls. Platforms of selection for fine-tuning shadow AI apps embody Google Colab and Google AI Studio. Many are utilizing Replit to create standalone apps.
Constructing Shadow AI apps with enterprise-grade attain
By combining a sequence of APIs and search engine IDs from Anthropic, Perplexity, Open AI and Google, the pace, accuracy and acuity of insights, associates’ shadow AI apps ship attain past the present scope of legit, IT-approved copilots and chatbots. One SME chief confided to VentureBeat that the mixture of APIs and Python fine-tuning makes it potential to create apps so hyper-customized to a consumer’s targets that it’s saving him days of guide work aggregating and analyzing information.
Associates at high corporations globally have created dozens, and in some instances a whole lot, of distinctive Google Search Engine APIs and IDs to energy their Python apps. These APIs present exact, real-time integration of exterior information instantly into their shadow AI instruments, additional boosting their analytical edge.
Shadow AI is rapidly rising as the brand new consulting stack
An evaluation by Cyberhaven of AI utilization throughout three million staff discovered that 73.8% of office ChatGPT accounts had been private moderately than company, indicating that almost all consultants flip to those instruments independently. Cameron Coles, VP of Advertising and marketing at Cyberhaven’s weblog put up final month, AI Utilization at Work Is Exploding — However 71% of Instruments Put Your Information at Danger, gives insights into what sort of information is most frequently shared throughout shadow AI apps and the speedy progress of the class.
Coles writes, “AI utilization at work continues its exceptional progress trajectory. Prior to now 12 months alone, utilization has elevated 4.6x, and over the previous 24 months, AI utilization has grown an astounding 61x. This represents one of many quickest adoption charges for any office know-how, considerably outpacing even SaaS adoption, which took years to attain related penetration ranges.”
Inside high consulting corporations, the proliferation of self-built, unauthorized apps continues to be explosive. Itamar Golan, CEO of Immediate Safety, notes, “We see 50 new AI apps a day, and we’ve already cataloged over 12,000,” highlighting how quickly these shadow instruments are rising. He just lately instructed VentureBeat throughout an interview that “many default to indiscriminately coaching on proprietary information inputs,” exposing corporations to substantial threat. Inside information confirms that 70–75% of consultants now usually depend on generative AI apps, instantly attributing productiveness positive aspects to shadow AI apps. It’s develop into the weapon of selection for consulting’s high expertise, enabling them to provide distinctive work in a fraction of the time.
VentureBeat interviewed Golan, Vineet Arora, CTO of WinWire, and senior leaders at fourteen main international consultancies to know the breadth of shadow AI adoption.:
Estimating the true scale of shadow AI in consulting
Whereas most enterprise instruments nonetheless fail to detect the dimensions of shadow AI use, discipline interviews and telemetry from Immediate Safety, WinWire and interviews with 14 top-tier consulting corporations make one factor clear: shadow AI is now not a fringe phenomenon. It’s rising as a parallel tech stack constructed from the bottom up by consultants themselves.
VentureBeat’s estimate incorporates:
- Immediate Safety telemetry, which detects ~50 new shadow AI apps per day and has already cataloged 12,000+ instruments throughout consulting corporations globally.
- WinWire enterprise AI information, masking Gemini, GPT-4, Claude 3 and Colab-based deployments.
- Cyberhaven utilization information, which reveals that 73.8% of ChatGPT office accounts are unauthorized, and enterprise AI utilization has grown 61x in 24 months.
- 14 govt interviews with CTOs, CISOs, AI leads and companions throughout Tier 1 corporations.
Solely actively deployed, production-grade instruments are counted, not one-off prompts, short-term Google Colab notebooks, or ChatGPT browser classes. These numbers replicate a validated decrease certain, probably far in need of the true whole.
Shadow AI app panorama in consulting, 2025 (Verified Estimate)
| Use Case Class | Estimated Shadow AI Apps (Q2 2025) | Major Instruments Used |
| Pitch & Proposal Automation | 12,000 | GPT-4, Gemini, Replit, Colab |
| Market Segmentation & Concentrating on | 9,000 | Perplexity, Gemini APIs, RAG apps |
| Analysis Assistants & Data Bots | 15,000 | Claude 3, Gemini Professional, Google Search APIs |
| Consumer-Dealing with Chatbots & Brokers | 7,500 | OpenAI Assistants, LangChain, customized LLMs |
| Workflow & Productiveness Automation | 13,000 | Python automations, Sheets, Zapier |
| Monetary Evaluation & Situation Fashions | 18,000 | Monte Carlo fashions, Gemini + Python |
| Whole (Validated Estimate) | 74,500+ |
Sources: Immediate Safety, WinWire, Cyberhaven, VentureBeat interviews with 14 international corporations
Shadow AI progress trajectory: What comes subsequent
Shadow AI is scaling quicker than any sanctioned inner platform, and most corporations haven’t any actual strategy to sluggish it down. Based mostly on a conservative 5% month-over-month progress charge, the variety of actively used shadow apps may greater than double by mid-2026.
What began as remoted productiveness scripts has developed into one thing extra sturdy. Shadow AI is now not a fringe toolset. It’s now a parallel supply stack. It operates exterior IT, with out formal governance, but it powers most of the high-value outputs corporations ship to purchasers each day.
Projected shadow AI app progress in consulting
| Quarter | Projected App Rely | Drivers of Progress |
|---|---|---|
| Q2 2025 | 74,500+ | Verified energetic apps from Immediate Safety, WinWire, and interviews |
| Q3 2025 | 90,000 to 95,000 | Progress in Gemini and Claude apps, partner-led growth |
| This autumn 2025 | 110,000 to 115,000 | Shadow instruments develop into embedded in consumer supply workflows |
| Q1 2026 | 130,000 to 140,000 | Emergence of gray-market copilots and self-maintained apps |
| Q2 2026 | 150,000 to 160,000+ | Shadow AI evolves into an ungoverned parallel supply stack |
These projections exclude one-off use of ChatGPT or Gemini in browser classes. They replicate persistent apps and workflows constructed utilizing APIs, scripting, or automated brokers developed inside consulting groups.
How one can strategically handle shadow AI dangers
Shadow AI is prospering as a result of conventional IT and cybersecurity frameworks aren’t designed to trace its use. IT and safety groups in practically each enterprise VentureBeat spoke with have three to 5 occasions the variety of initiatives they will full this yr. Whereas getting a brand new copilot out is a excessive precedence, it may well face useful resource and approval hurdles in consequence.
Arora underscores that “most conventional administration instruments lack complete visibility into AI apps,” enabling unauthorized AI to quietly embed itself inside enterprise workflows.” Arora’s insights reveal an underlying fact: Workers aren’t appearing maliciously; they’re petrified of being let go whereas concurrently overwhelmed with work, leveraging AI to deal with escalating workloads, shrinking deadlines, and relentless efficiency expectations.
Quite than stifling AI adoption, Arora advocates proactive empowerment via strategic, centralized governance. By institutionalizing clear oversight, organizations can harness AI securely, reworking shadow AI from an unseen risk right into a managed asset.
A blueprint for governance
Consultancies’ senior administration groups want a transparent, sensible roadmap to get in entrance of shadow AI dangers and harness its strategic potential. Arora outlined an in depth governance framework throughout a current interview with VentureBeat, explicitly designed for enterprises navigating the complexities of shadow AI:
- Shadow AI audits are desk stakes:
Repeatedly take stock of all unauthorized AI exercise via strong community monitoring and detailed software program asset administration.
- Create an Workplace of Accountable AI:
Centralize AI governance capabilities spanning coverage creation, vendor assessments and threat evaluation, and preserve an accepted AI instruments catalog accessible to all groups.
- Get AI-aware safety controls in place instantly:
Deploy specialised Information Loss Prevention (DLP) instruments and real-time inference monitoring able to detecting delicate information leaks particular to AI functions in real-time.
- Go all in on making use of zero belief to AI architectures:
Undertake strict output validation protocols, anonymize or tokenize delicate inputs, and rigorously handle information flows to attenuate publicity and stop unauthorized information coaching.
- Discover out the place the roadblocks are to getting extra gen AI instruments out now:
Each group can enhance on the pace at which it deploys new applied sciences. Discover out the place the gaps and roadblocks are holding the consultancy again from delivering more proficient copilots and chatbots. It’s important to get a roadmap outlined for IT and DevOps to work on for internally steered Python apps, fine-tuned to consumer wants.
- GRC integration and steady coaching:
Combine AI governance inside current governance, threat, and compliance (GRC) frameworks, and constantly seek the advice of on safe, compliant AI practices.
- Keep away from blanket bans, it’s gas for much more shadow AI app growth:
Acknowledge that outright AI bans inevitably backfire, rising shadow AI proliferation. As an alternative, quickly deploy safe, sanctioned alternate options that allow compliant, productive innovation.
Initially an underground productiveness hack, shadow AI has emerged as a decisive consider how top-tier consultants ship differentiated consumer worth. Pushed by a stark survival crucial amid widespread AI-triggered layoffs, elite expertise now depends on Python-driven, generative AI-powered options, enabling uniquely tailor-made consumer insights and speedy responses to their purchasers.
Consulting corporations which are sluggish to adapt or hesitant to strategically harness these improvements strategically threat forfeiting their future aggressive edge. The trail ahead calls for not prohibition however considerate, safe integration of shadow AI and the transformation of potential dangers into decisive strategic benefits.
