The New Playbook for Enterprise AI Contracts

This article is sponsored by UpperEdge and was written, edited, and published in alignment with our Emerj sponsored content guidelines. Learn more about our thought leadership and content creation services on our Emerj Media Services page.

Enterprise AI spend and outcomes are diverging, and available data quantify the gap.

The difficulty for CIOs is whether they can price, negotiate, and budget for AI before the ground shifts again.

The U.S. Government Accountability Office reported in April 2026 that federal agencies more than doubled their use of AI between 2023 and 2024, while officials across the agencies GAO reviewed cited “the difficulty of determining pricing and overall cost of AI adoption” as a distinct, named acquisition challenge — separate from the technology’s performance. The report notes that agencies increasingly buy AI “as a service,” in which a vendor supplies capability on an ongoing basis rather than as a fixed-price product, replacing a budget line with an open-ended commitment.

A separate GAO review documented that agencies were effectively locked into vendors — not because of any restrictive clause, but because the cost of re-architecting around a competitor had become prohibitive once systems were built around one provider’s tools. A related audit found that 10 vendors account for roughly 73% of the most widely used federal software licenses, with one vendor, Microsoft, representing more than 31% of total spend — a concentration that leaves buyers negotiating from a structurally weaker position.

Brookings’ tracking of federal AI contracts shows potential award value in one major contract category growing from $311 million to $1.9 billion, and in another from $5 million to $2.2 billion, within roughly two years — growth rates that outpace the multi-year terms most agreements are still written around.

These findings describe a market where the bill, the leverage, and the contract term are all moving at different speeds — and rarely in the buyer’s favor.

Emerj’s AI in Business podcast recently featured John Belden, Chief of Strategy and Research at UpperEdge; Adam Mansfield, Practice Leader at UpperEdge; and David Cost, Chief Digital Officer at Rainbow Apparel, in a three‑episode series examining how AI is reshaping enterprise technology.

This article examines four insights executives can use to protect cost, flexibility, and leverage as AI reshapes enterprise technology.

Reversible transformation decisions: Structure AI and ERP commitments so they can be undone or redirected, enabling leaders to adjust as technology, vendor roadmaps, and regulatory conditions shift faster than any traditional transformation model can keep pace with.

Evidence‑based negotiation leverage: Ground every AI commercial discussion in hard usage and value data, replacing vendor projections with enterprise evidence so pricing, consumption tiers, and risk allocation are set by facts rather than vendor assumptions.

Short‑cycle commercial commitments: Replace multi‑year agreements with short terms, explicit exit clauses, and enterprise‑controlled proof‑of‑value cycles, ensuring vendors must continuously earn renewals as AI reshapes the economics of build‑vs‑buy.

Independent accountability for SI and vendor productivity: Require system integrators and software vendors to disclose their AI roadmaps and submit to periodic capability and productivity audits, aligning compensation with measurable improvements instead of static deliverables written for a pre‑AI environment.

Listen to the full episodes below:

Episode 1: The Hidden Risk in Every Enterprise AI Vendor Contract – with John Belden of UpperEdge

​Guest: John Belden, Chief of Strategy and Research at UpperEdge

Expertise: IT Strategy, Digital Transformation, IT Governance & Risk, Enterprise Technology

Brief Recognition: John Belden is a technology and business transformation executive with more than 25 years of experience leading enterprise IT strategy, governance, and large-scale transformation programs. He currently serves as Chief of Strategy and Research at UpperEdge, where he leads research initiatives and advises organizations on IT-enabled transformation, governance, risk management, and technology strategy. Prior to UpperEdge, John held multiple executive leadership roles at The Timken Company, including Vice President of Project ONE and Vice President of Information Technology. He also co-hosts the Insights for IT Negotiations podcast, covering enterprise technology, AI, and IT sourcing strategies. John holds a Master’s degree in Computer Science from Kent State University.​

Episode 2: The Pricing Shift Reshaping Enterprise AI Spend – with Adam Mansfield of UpperEdge

​Guest: Adam Mansfield, Practice Leader at UpperEdge

Expertise: IT Contract Negotiation, SaaS & Cloud Strategy, Vendor Management, Enterprise Software Procurement

Brief Recognition: Adam Mansfield is an enterprise technology advisor with more than 15 years of experience helping organizations negotiate complex software, cloud, and IT services agreements. He is a Leadership Team Member and Practice Lead at UpperEdge, where he advises enterprise executives on negotiations involving major technology providers including Microsoft, Salesforce, ServiceNow, and leading AI vendors. Prior to UpperEdge, Adam led contract negotiation and benchmarking engagements at AMR Research. Earlier in his career, he negotiated software and consulting agreements at Skillsoft. Adam holds both an MBA from the Suffolk University Sawyer Business School and a JD from Suffolk University Law School.

Episode 3: How Commerce Leaders Avoid Renewal Traps and Vendor Drag – with David Cost of Rainbow Apparel

Guest: David Cost, Chief Digital Officer at Rainbow Apparel​

Expertise: AI Strategy, Digital Transformation, E-commerce Technology, Enterprise Architecture

Brief Recognition: David Cost is a digital transformation and technology executive with experience spanning AI, e-commerce, enterprise architecture, and digital strategy. He currently serves as Chief Digital Officer at Rainbow Apparel, where he leads the company’s digital platform, AI initiatives, and technology strategy across its omnichannel retail business. During his tenure, he led the organization’s migration to Shopify, expanded its e-commerce capabilities, and introduced AI-enabled workflows across marketing, personalization, and digital operations. Earlier in his career, David co-founded PriceSCAN, an early comparison-shopping platform, and began in management consulting, focusing on decision support systems and information processing.

Reversible transformation decisions

AI has made long‑range transformation planning inherently unstable, and the series opens with John Belden, arguing that modern programs must be engineered so major decisions can be unwound.

Belden frames transformation as a sequence of commitments made under shifting conditions — platform maturity, regulatory pressure, SI delivery models, pricing volatility — and insists that leaders must define how each commitment can be reversed before making it.

His core message: reversibility is not a mindset; it’s a structural design choice.

“Most transformations fail not because the initial decision was wrong, but because the organization had no way to change direction once reality shifted. A reversible decision has a trigger, a pivot path, and a cost profile you understand before you execute it. If you can’t articulate those elements, you’re not managing uncertainty — you’re surrendering to it.”

—John Belden, Chief of Strategy and Research at UpperEdge

Adam Mansfield extends Belden’s point into commercial structure. He argues that reversibility only exists if contracts allow it — meaning short terms, renegotiation triggers, consumption protections, and pricing tied to observable outcomes.

Mansfield’s view is that governance and deal architecture must be designed together, or strategic pivots become legally impossible even when they’re operationally necessary.

In retail, David Cost states that he has seen AI initiatives succeed when they’re treated as controlled experiments with defined exit ramps rather than multi‑year commitments. He describes building transformation roadmaps with kill switches, alternative vendor paths, and rapid evaluation cycles so teams can move fast without locking the business into a single AI approach or commercial structure.

Belden’s practical framework for reversible decisions:

Define the irreversible commitments — identify the few decisions that cannot be unwound and minimize them.

Map the pivot paths — document how each major decision could be redirected if conditions change.

Set explicit trigger signals — determine what evidence would justify a pivot (pricing shifts, roadmap slippage, regulatory changes).

Assign decision ownership — clarify who makes the call and how quickly the organization must respond.

Pre‑calculate the cost of reversal — understand the operational and financial impact before committing.

Evidence‑based negotiation leverage

AI has pushed enterprise pricing into a volatile, consumption‑driven model, and Adam Mansfield argues that the only way for leaders to negotiate from strength is to anchor every commercial discussion in hard usage and value data. He stresses that vendors themselves cannot accurately model future AI consumption, which means the enterprise must walk into negotiations with its own evidence — not vendor projections — to prevent unpredictable spend and misaligned commitments.

Mansfield frames evidence as the foundation of leverage: leaders must know exactly where current spend is under‑ or over‑utilized, which capabilities drive measurable value, and where consumption patterns contradict vendor assumptions. Without that baseline, AI pricing becomes guesswork:

“AI pricing is built on uncertainty, and vendors will always try to make their assumptions your reality. The only way to negotiate from strength is to bring evidence they cannot dispute — usage patterns, value delivered, and the true criticality of each capability. When the enterprise owns the facts, the vendor loses the ability to dictate the future.”

—Adam Mansfield, Practice Leader at UpperEdge

Mansfield’s guidance becomes most practical when he outlines how leaders should prepare before entering any AI‑related commercial discussion. His approach is a negotiation sequence rather than a static checklist — a way to convert evidence into leverage:

Audit current usage to expose under‑ and over‑leveraged spend.

Quantify business value so pricing reflects actual impact, not vendor narratives.

Challenge vendor assumptions by comparing their projections to enterprise evidence.

Model realistic consumption scenarios based on observed patterns, not theoretical ones.

Anchor negotiations in facts so pricing tiers, risk allocation, and commitments reflect reality.

John Belden reinforces Mansfield’s point by directly linking evidence to transformation governance. For Belden, usage and value data are not just negotiation tools — they are the backbone of decision‑making. He argues that steering AI programs without recurring evidence reviews is indistinguishable from steering by intuition, and that governance forums should be built around measurable consumption, roadmap delivery, and risk signals.

David Cost describes how he has seen in negotiations that shifts dramatically when teams arrive with a clear picture of which capabilities actually drive revenue, margin, or cycle‑time reduction. Cost emphasizes that evidence allows leaders to strip out non‑essential features, challenge bundled AI upsells, and insist on pricing that reflects real business impact rather than theoretical vendor value.

Short‑cycle commercial commitments

AI is collapsing the value horizon of enterprise software, and David Cost argues that long, multi‑year contracts no longer make sense for anything outside fo

[truncated for AI cost control]