Agentic AI: Adobe’s Finance Chief Leads the Charge in Automating the Back Office
In a significant shift reshaping corporate leadership, Adobe’s finance chief, Dan Durn, is spearheading an ambitious initiative to transform the company’s finance organisation into a cutting-edge proving ground for agentic artificial intelligence. This bold strategy involves deploying autonomous software agents to handle a diverse range of critical tasks, from forecasting financial results and meticulously scanning contracts to managing an overwhelming volume of emails.
This internal push mirrors Adobe’s broader commitment to agentic AI, a technology that empowers users to select and customise AI models, integrate them with proprietary and company data, and direct these agents towards specific business objectives. For customers, this means a more personalised and powerful AI experience. Internally, Durn, who also oversees technology, security, and operations, has adopted a parallel approach. By integrating AI into the inherently data-intensive and rules-based finance function, and by consolidating finance, IT, and security under a single leadership umbrella, Adobe is creating an environment where pilot projects can swiftly transition to full production.
“Accuracy is non-negotiable,” Durn emphasises. This unwavering focus on precision underpins Adobe’s substantial investment in structured data and robust governance frameworks, ensuring that the company can innovate at speed without compromising accuracy.
The rapid ascent of AI is undeniably altering the landscape of corporate leadership. It is accelerating executive turnover and placing a premium on leaders who can demonstrate swift, tangible results. Even seasoned executives are feeling the heat from investors, who are increasingly demanding aggressive adoption of AI technologies. Recent high-profile leadership changes within the industry underscore the market’s dwindling patience for perceived hesitation. Adobe itself reported a more than threefold year-over-year increase in annualised revenue from its AI-first products in the first quarter of fiscal 2026, a testament to the growing demand for AI-driven solutions. Across the corporate world, this dynamic is fostering a new internal proving ground where executives are increasingly evaluated on their effectiveness and speed in deploying AI to fuel growth, enhance efficiency, and drive innovation.
AI in Action: Transforming Finance Operations
Within the finance department, Durn has categorised AI applications into three key areas: forecasting, anomaly detection, and general productivity enhancements.
- Forecasting: AI agents excel at identifying intricate patterns and subtle signals within vast datasets that would be exceptionally challenging for human analysts to detect quickly. This allows for more proactive and informed financial planning.
- Anomaly Detection: These intelligent agents are designed to flag any performance metrics that deviate unexpectedly from the norm, whether positively or negatively. By identifying these anomalies early, the finance team can intervene promptly, preventing potential issues from escalating or capitalising on unforeseen opportunities that might otherwise be lost in the sheer volume of data.
Productivity Gains: Durn highlights this category as currently yielding the most impactful results, citing three prominent use cases:
Extracting Information from PDFs:
This is one of the most mature applications of agentic AI at Adobe. The finance team leverages Adobe’s PDF Spaces, a collaborative digital workspace, to consolidate collections of documents such as investor transcripts, quarterly reports, and analyst research. An agentic AI assistant then swiftly processes these documents, surfacing key themes, insights, and messaging cues in a matter of minutes, a process that previously took hours. A recent Forrester TEI study underscored the efficiency gains, finding that Acrobat’s agentic AI Assistant boosts document summarisation and analysis efficiencies by an impressive 45%. Durn stresses the significance of this capability, noting that “the world’s information lives in PDF,” and AI’s ability to transform this static content into actionable insights is invaluable.Halving Contract Review Time:
Adobe is revolutionising its contract review processes across finance and procurement functions, including revenue assurance, contract operations, product fulfillment, and vendor management, through the implementation of agentic AI. Instead of finance professionals painstakingly scrutinising every clause, an AI assistant now scans thousands of contracts. It intelligently highlights provisions pertinent to each specific function and flags any non-standard terms. This system has dramatically reduced review times by approximately 50%. Beyond speeding up individual reviews, it enables teams to query the entire contract repository with unprecedented ease. For instance, they can quickly identify all contracts featuring auto-cancellation clauses or specific foreign-exchange adjustment windows. Adobe developed its initial prototype by April 2024 and began onboarding teams in January 2025.Automating “Common” Inboxes:
A third significant area of AI implementation involves the automation of “common inboxes” – shared email addresses that handle high volumes of internal and external communications for departments like sales, treasury, finance, and supplier inquiries. Adobe has deployed an agentic AI assistant capable of automatically tagging, prioritising, routing, and, under specific predefined criteria, auto-responding to emails. Typical queries managed by this system include supplier billing discrepancies or standard credit-quality inquiries directed to the treasury department from platforms like Salesforce.
In 2025 alone, this system successfully auto-responded to approximately 300,000 emails across 19 different inboxes, translating to over 5,000 hours of saved manual labour. This automation frees up valuable human resources to concentrate on more complex and strategic issues. The development of this tool took roughly six months, with beta testing commencing around August 2024 and a full rollout in January 2025. Durn is quick to point out that the primary benefit is not headcount reduction, but rather the enhanced ability to scale operations more efficiently as Adobe continues to grow.
A Decade in the Making: From Grassroots Ideas to AI Adoption
Durn attributes these successful finance use cases to Adobe’s long-standing commitment to AI and a culture that fosters innovation from the ground up. The company has been investing in machine learning and AI for over a decade, initially focusing on understanding customer behaviour and integrating intelligence into its products. This foundational work has been instrumental in paving the way for the current advancements in generative and agentic AI.
He highlights that many of the most effective AI applications emerge from actively seeking input from employees across the organisation, asking them where AI could alleviate pain points or improve their daily work. Currently, the demand for AI solutions outstrips the available capacity, necessitating a rigorous prioritisation process focused on applications with the greatest potential impact.
When evaluating AI investments, Durn’s primary consideration is “organisational velocity” – the capacity of back-office functions to keep pace with the rapid innovation occurring in product development. He argues that if finance fails to embrace AI, it risks becoming a significant bottleneck, impeding the company’s overall growth trajectory. The actual financial outlay for these initiatives is relatively modest, he notes, with a substantial portion of the effort dedicated to change management and process redesign, built upon Adobe’s existing technological infrastructure.
Durn’s perspective on change management aligns with recent research from McKinsey. The consultancy’s findings suggest that to fully realise the benefits of AI, organisations must move beyond fragmented approaches and pursue a “double transformation” – encompassing both technological advancements and organisational restructuring. This includes a fundamental reimagining of how work is executed across various functions and workflows. While a significant 88% of surveyed organisations are actively experimenting with AI, fewer than 20% report achieving tangible, bottom-line results, according to the research.
AI’s Impact on Executive Workflow
For his own daily operations, Durn primarily relies on AI for insight generation. In the lead-up to earnings announcements, his team utilises an AI-powered workspace that ingests pre-earnings research reports, Adobe’s filings, and transcripts from peer companies. This enables the surfacing of key themes and the anticipation of likely investor questions. Subsequently, scripts and question-and-answer preparations are run through AI models with built-in guardrails. This process rigorously tests whether the messaging effectively addresses the identified themes and prompts critical self-reflection with questions like, “If I were an investor, what are my key takeaways?” Durn views this as a valuable mechanism for validating instincts, ensuring clarity, and sharpening Adobe’s communication strategy with the market.
