maddaisy – cultivating innovation

Tag: consulting

  • Open Protocols Promise AI Flexibility, But Platform Partnerships Define Delivery Advantage

    Enterprise AI orchestration faces a fundamental tension between technological flexibility and delivery complexity. While EY’s Canvas platform now processes 1.4 trillion lines of audit data annually and Model Context Protocol deployment spans 10,000+ enterprise servers, the promise of vendor flexibility through open standards is colliding with a harder reality: successful AI delivery depends on platform partnerships, not protocol freedom.

    As we reported in June, consulting firms that embed in outcome-validated platforms capture the most strategic margin. The April 2026 data now reveals why: Adobe’s formal partnerships with nine system integrators and seven AI model providers signal that vendors view professional services firms as mandatory delivery channels, not optional implementers.

    From Tool Selection to Strategic Partnership

    The shift is structural, not incremental. Where enterprises once procured AI tools through IT departments, they now select strategic partners whose vendor relationships define delivery capability. Choosing an agentic AI vendor in 2026 is fundamentally different, according to AI practitioner Kai Waehner. “Unlike a CRM or an ERP, an AI vendor is not just a tool you deploy. It is a strategic partner whose safety culture, governance model, and long-term ambitions will directly influence the reliability and trustworthiness of your most critical business processes.”

    Adobe’s April announcement exemplifies this shift. The company has established formal partnerships with Accenture, Capgemini, Cognizant, Deloitte Digital, EY, IBM, Infosys, PwC, and TCS – creating an ecosystem where consulting firms become the primary go-to-market channel for enterprise AI adoption. The partnership structure positions consulting firms as the implementation layer that handles change management, governance architecture, and cross-functional coordination required for enterprise deployment.

    The Open Standards Paradox

    Model Context Protocol (MCP) and Agent-to-Agent (A2A) standards are being deployed across thousands of enterprise environments, promising interoperability between model providers like OpenAI, Anthropic, and Google. The technical flexibility is real – enterprises can now orchestrate multiple AI models without rebuilding integration layers for each vendor switch.

    But this flexibility introduces governance complexity that few organisations are prepared to manage. HCLTech’s survey of 467 senior executives reveals that 43% of major AI initiatives are expected to fail, driven not by model quality or tool access, but by “governance gaps, technical debt, integration pitfalls, or vendor lock-in” management failures.

    Open protocols solve the vendor selection problem but create an architecture management problem. Enterprises now need specialists who can design multi-vendor governance frameworks, manage the integration dependencies, and navigate the trust implications of distributing critical workflows across multiple AI providers. This requirement favours consulting firms with proven platform delivery experience over those offering point-solution deployment.

    Platform-Centric Delivery Wins

    The evidence from production deployments confirms platform-centric delivery as the competitive advantage. Salesforce Agentforce deployments achieve 84% case resolution improvements in production environments, while EY’s Canvas platform supports 130,000 professionals across 160,000 global engagements. These outcomes reflect integrated platform approaches that bundle orchestration, governance, and change management into a single delivery framework.

    The time-to-value pressure is intensifying this platform preference. HCLTech’s research shows that nearly 50% of enterprise leaders expect measurable value from AI investments within 18 months. Custom integrations of multiple point tools rarely meet this timeline, while platform-based approaches with proven governance models can deliver results within enterprise expectations.

    For professional services firms, this creates a clear strategic choice: build delivery practices around proven platforms with established vendor partnerships, or compete on custom integration capability in an increasingly commoditised market.

    Governance as Competitive Differentiator

    As we noted in May, regulatory compliance is forcing governance onto the critical path of AI delivery. The April 2026 production data shows this creating a deeper structural shift: governance capability is becoming the primary differentiator between consulting firms that capture strategic margin and those competing on implementation hours.

    Waehner’s framework positions vendor selection around two dimensions: trust in the vendor’s AI capabilities and tolerance for vendor lock-in. But this analysis misses the third dimension that consulting firms must navigate: governance architecture across multi-vendor environments. As Capgemini’s Mark Roberts notes, 2026 represents a shift “from innovation theatre to a more mature focus on real, practical deployment” where “integration rather than invention” defines success.

    Professional services firms that position themselves as governance and integration specialists rather than tool deployers capture the highest-value engagements. They architect for compliance, design for multi-vendor orchestration, and manage the organisational change required to sustain agentic AI at scale.

    What This Means for Professional Services

    The implications centre on three shifts evident in the production data and vendor partnership announcements. First, vendor partnerships have become part of go-to-market value propositions. The formal ecosystem partnerships announced by Adobe, Microsoft, Google, and AWS create delivery channels that differentiate consulting firms’ ability to scale AI implementations.

    Second, as Waehner’s vendor selection framework demonstrates, technical interoperability through open protocols does not reduce complexity – it redistributes it. MCP and A2A standards enable flexibility if governance and integration architecture can support it, but the 43% failure rate for major AI initiatives suggests most enterprises cannot yet manage multi-vendor orchestration effectively.

    Third, the data suggests that execution gaps rather than tool limitations drive the high failure rates for AI initiatives. Professional services firms that focus on change management, stakeholder alignment, and governance maturity rather than tool sophistication are better positioned to deliver within enterprise timelines and capture strategic consulting margin.

    The promise of open standards eliminating vendor lock-in reflects a structural trade-off rather than a pure benefit. This points to a shift from vendor dependency to architectural complexity and integration risk. But for consulting firms with the governance and platform delivery capability to manage that complexity, this shift creates sustainable competitive differentiation in an increasingly crowded market.

  • Enterprise AI Adoption Is Outcome-Driven, Not Architecture-Driven

    Enterprise AI adoption in 2026 is being driven by measurable business outcomes, not by technical architecture debates or vendor flexibility. The evidence is clear: platforms delivering concrete value capture market share, and consulting firms embed where results are proven. 

    The numbers tell the story. According to a fifthrow.com analysis, EY’s Canvas platform now processes 1.4 trillion lines of audit data annually across 160,000 global engagements. Salesforce’s Agentforce deployments are achieving 84% improvements in case resolution. Adobe’s customer experience orchestration platform has mobilised 13 consulting partners including Accenture, Capgemini, Deloitte Digital, EY, IBM, Infosys, PwC, and TCS to scale deployment. These aren’t technology choices — they’re business results that enterprises are willing to invest in.

    Enterprise procurement of AI platforms follows the same logic as ERP or CRM decisions: which vendor delivers measurable impact? As industry analyst Kai Waehner observes, “Unlike a CRM or an ERP, an AI vendor is not just a tool you deploy. It is a strategic partner whose safety culture, governance model, and long-term ambitions will directly influence the reliability and trustworthiness of your most critical business processes.” The difference is that AI outcomes depend more heavily on integration depth and governance alignment than traditional software did. Enterprises that want EY’s audit processing capabilities or Salesforce’s case resolution improvements cannot achieve them through loose integrations or multi-vendor orchestration — they need platform embeddedness.

    Consulting firms have followed this logic. Adobe’s April announcement formalising partnerships with major consultancies isn’t a commercial lock-in scheme — it’s a recognition that the platform’s value multiplier lives in implementation expertise, not in the software itself. These firms embed because that’s where their clients are getting results.

    The Real Bottleneck: Implementation, Not Architecture

    Open standards like Model Context Protocol (MCP) and Agent-to-Agent (A2A) have been deployed across 10,000-plus enterprise servers. They serve the same infrastructure role that TCP/IP played in networking or ODBC in databases — necessary technical foundations that enable interoperability. Their existence doesn’t change the fact that enterprises achieve competitive advantage through proprietary business logic, not through architectural neutrality.

    The actual constraint on enterprise AI success isn’t technology selection. According to HCLTech’s survey of 467 senior executives at billion-plus-dollar enterprises, 43% of major AI initiatives are expected to fail. This failure rate has nothing to do with whether platforms use open standards or proprietary integrations. It reflects organisational readiness: data quality, change management maturity, governance frameworks, and cross-functional alignment.

    This is where consulting value lives. Technology availability has far outpaced organisational capability to deploy it. As Mark Roberts, Head of AI Future Labs at Capgemini, noted in January, “2026 is a moment of truth for AI. After years of headlines, investment and experimentation, the mood is shifting: innovation theatre is giving way to a more mature focus on real, practical deployment.” The gap between EY’s 1.4 trillion data points processed and the industry’s 43% failure rate isn’t a technology problem — it’s a change management problem. Consulting firms that embed with winning platforms don’t do so to enforce lock-in; they do so because that’s where they can help clients bridge the implementation gap.

    Why This Matters

    The shift from vendor-neutral advisory to platform-embedded partnership reflects market maturation, not advisory capture. Enterprises prioritising measurable outcomes over architectural purity are making rational decisions. Consulting firms following those outcomes are allocating their expertise sensibly.

    The remaining question isn’t whether consulting firms should partner with platforms — the evidence suggests they should, where partnerships deliver client value. The question is whether they can transparently manage those relationships whilst maintaining credibility around trade-offs. That requires clarity about commercial relationships and disciplined case management, but it doesn’t require artificial neutrality across platforms that deliver different outcomes.

    Enterprise AI deployment in 2026 is driven by results, not by technology debates. Consulting firms are embedded where the results are. That’s not a problem to solve — it’s the market working as it should.

  • AI Safety Report 2026: Computing Power Crosses Critical Threshold as Enterprise Deployment Gaps Emerge

    Computing power for AI training runs has likely surpassed 10^26 FLOP in 2025, marking the first measurable milestone in a new era of AI capability assessment. This finding emerges from the International AI Safety Report 2026, the first comprehensive international assessment of AI capabilities versus safeguards, produced by over 100 AI experts from 30+ countries.

    For enterprise leaders considering AI agent deployment, the report provides something the industry has lacked: an evidence-based framework for evaluating actual capabilities rather than theoretical projections. The timing matters because boutique consultancies and tech firms are increasingly being asked to advise on AI implementation without clear benchmarks for readiness assessment.

    From Speculation to Measurement

    The 10^26 FLOP threshold represents more than a technical milestone – it provides the first concrete capability marker that enterprises can use to gauge where AI development stands relative to their implementation timelines. As Inside Privacy notes, “The Report does not make specific policy recommendations; instead, it synthesizes scientific evidence to provide an evidence base for decision-makers.”

    This evidence-based approach addresses a critical gap in enterprise AI planning. Where previous assessments relied on vendor projections or theoretical models, the international panel has produced measurable benchmarks that consulting firms can use to evaluate client readiness for AI integration.

    The report’s synthesis methodology – drawing insights from experts across 30+ countries and international organisations – suggests that future AI safety assessments will focus on observable capabilities rather than hypothetical scenarios. This shift gives consultancies clearer implementation guidance when advising enterprise clients on AI adoption strategies.

    Enterprise Implementation Reality Check

    While the computing power milestone captures headlines, the report’s capability gap analysis provides more immediate practical value for business leaders. The assessment reveals specific areas where current AI systems fall short of enterprise deployment requirements, particularly around autonomous agent reliability and contextual decision-making.

    These findings matter because enterprise software buyers often face vendor claims about AI agent capabilities without independent benchmarks for evaluation. The international panel’s evidence-based framework provides consultancies with tools to assess whether specific AI implementations match client operational requirements.

    The regulatory landscape adds urgency to this capability assessment. As Nature reports, at least 30 AI-related laws were passed globally in 2023, with 40 more in 2024, while US states passed 82 AI-related bills in 2024. Enterprise leaders need concrete capability frameworks to navigate this evolving compliance environment.

    What the Framework Actually Measures

    The report moves beyond broad AI safety discussions to examine specific capability areas that directly impact enterprise deployment decisions. Rather than theoretical risk scenarios, the assessment focuses on measurable performance gaps in areas like task persistence, error recovery, and contextual reasoning – factors that determine whether AI agents can reliably handle business-critical processes.

    For consultancies working with enterprise clients, this granular capability assessment provides a foundation for realistic implementation planning. Instead of advising clients based on vendor demonstrations or pilot project results, firms can reference internationally validated benchmarks to identify specific capability gaps that need addressing before full-scale deployment.

    The evidence-based approach also helps separate genuine capability advances from marketing positioning. When vendors claim breakthrough performance, consultancies can reference the report’s framework to evaluate whether claimed improvements address real operational requirements or represent incremental refinements.

    Forward-Looking Implementation Strategy

    The report’s impact extends beyond immediate capability assessment to longer-term strategic planning. By establishing evidence-based evaluation criteria, the international panel has created a framework that will likely influence how regulatory bodies assess AI deployment risks and how enterprises structure their adoption timelines.

    This suggests that successful AI implementation strategies will increasingly depend on measured capability assessment rather than theoretical potential. Consultancies that master the report’s evidence-based evaluation approach will be better positioned to guide clients through the complex decisions around AI agent deployment timing and risk management.

    The 10^26 FLOP threshold provides a concrete reference point for tracking future capability developments, but the report’s lasting value lies in its methodology for connecting measurable AI capabilities to real-world deployment requirements. As enterprise AI adoption accelerates, this evidence-based framework offers consultancies and their clients a more reliable foundation for implementation planning than speculative projections or vendor promises.

  • Enterprise AI Finds Its Middle Ground: How Pega Blueprint Solves the Governance Gap

    Enterprise software development has a governance problem. Teams want the speed of AI-powered “vibes coding” – where natural language commands generate working applications – but enterprise requirements demand the control and predictability that casual AI tools cannot provide. Pegasystems’ latest update to its Pega Blueprint platform offers a glimpse of how this tension might resolve.

    The March 2026 Blueprint update transforms what was originally a linear design-to-handoff tool into a continuous conversational interface. Users can now modify enterprise applications through natural language – via text or speech – while maintaining the governance standards that large organisations require. It is a practical solution to what consultancies increasingly encounter: clients who want AI development speed but cannot sacrifice compliance and control.

    From Static Design to Living Conversation

    When Pega Blueprint launched in February 2024, it addressed a specific pain point: the slow, expensive upfront process of designing enterprise applications. The original version followed a conventional pattern – users described their application idea, Blueprint’s AI generated a structured starting point, then handed it to developers.

    The conversational update represents a different approach entirely. Rather than a one-time design exercise, Blueprint becomes what the company calls a “continuous copilot” – an interface that allows ongoing modification and refinement through natural language while preserving enterprise-grade security and governance requirements.

    “Organizations can create workflows more quickly, improve data and logic, and preserve control and predictability across mission-critical applications,” according to the company’s announcement. This combination – speed with governance – addresses what has become a fundamental challenge for enterprise AI adoption.

    The Consulting Opportunity  

    For consultancies, this development signals both opportunity and competitive pressure. The broader market context supports this view: recent Capgemini research shows that 85% of corporate clients plan to engage with non-bank providers within the next year, while only 23% believe traditional banks meet current expectations.

    The data reveals a broader pattern: organisations are seeking more agile, technology-forward partners. Traditional providers struggle – 82% of banking executives report no revenue gains from new products, and 51% see no expected cost reductions from innovation initiatives. Only 29% of IT budgets are directed toward transformative technologies.

    This creates space for consultancies that can effectively bridge the gap between AI capabilities and enterprise requirements. The challenge is not just technical – it is organisational. As Fortune reported, AI companies have discovered they need consultants to help sell their AI agents, as effective AI implementation requires significant organisational transformation: cleaning up data, redesigning workflows, and strategic thinking about competitive advantage.

    The Implementation Reality

    The consulting industry’s relationship with AI has evolved in an unexpected direction. Rather than eliminating consulting roles, AI complexity has created new demand for implementation services. OpenAI employs approximately 70 “forward deployed engineers” for customer implementation, and Anthropic maintains a similar number of implementation specialists.

    “AI still suffers from a trust deficit – most boards would still rather put their faith in advice from McKinsey or BCG than ChatGPT,” the Fortune analysis noted. This trust gap creates opportunities for consultancies that can position themselves as essential partners in AI implementation.

    The Pega Blueprint evolution illustrates this dynamic clearly. The platform promises that completed blueprints can be deployed as working workflows “in minutes,” but the enterprise governance layer – security, compliance, audit trails, role-based permissions – requires careful implementation and ongoing management.

    Market Signals

    Financial markets are taking a measured view of these developments. Citigroup raised its price target for PEGA stock to £75 from £73, citing “stable Q4 software results” and the company’s position in “defensive end markets.” The modest adjustment and “stable” language suggest measured progress rather than breakthrough momentum.

    This reflects a broader pattern in enterprise AI adoption: incremental evolution rather than revolutionary transformation. Conversational interfaces for development are not new, but solving the enterprise governance problem while maintaining development speed represents meaningful progress.

    What This Means for Consultancies

    The Blueprint update signals several trends worth monitoring. First, expect similar conversational interfaces to appear across enterprise platforms. The pattern – natural language interaction with robust governance – addresses a real market need.

    Second, the consulting market appears to be splitting between firms that can deliver AI-assisted solutions with enterprise governance and those constrained by traditional delivery models. The 85% of corporate clients planning to engage non-traditional providers suggests demand for more agile implementation partners.

    Third, the governance gap creates a consulting opportunity. Organisations need partners who understand both AI capabilities and enterprise requirements – not just one or the other. This requires a different skill set from traditional systems integration: understanding AI model behaviour, data governance for machine learning, and the organisational change required for AI-assisted workflows.

    The Pega Blueprint evolution represents enterprise AI finding its practical middle ground. For consultancies, the question is whether they can navigate this balance effectively – delivering AI-powered innovation without sacrificing the control and predictability that enterprise clients require. The market opportunity appears significant, but it demands a more sophisticated approach than either pure AI enthusiasm or traditional enterprise delivery.