Faster software discovery needs a faster transaction path

AI can accelerate software discovery without accelerating the purchase

AI is shortening the path to a software shortlist. More than 80% of buyers used AI chatbots for software recommendations over the past two years, according to G2’s “2026 Buyer Behavior Report.” Nearly half said AI had its greatest influence during shortlisting and evaluation. Faster discovery puts more pressure on the evaluation and transaction steps that follow.

G2 reports a difference between buyers who did and did not use AI for recommendations. Among those who did, 80% bought from their initial shortlist in at least three of their last five purchases. The figure was 65% among buyers who did not use AI for recommendations. Evaluation, internal approval, security review, contracting, billing, and payment still stand between a shortlist and completed revenue.

For executives, these findings separate two kinds of speed: how quickly buyers form and evaluate a set of choices, and how quickly an approved choice becomes a completed transaction. Improving discovery still leaves the transaction path to address.

The bottleneck is moving downstream

Evaluation already consumes more time than any other stage for many buyers. G2 reports that 40% now identify evaluation as the longest part of the software buying journey, up from 36% the previous year. The current breakdown shows where buyers spend the most time.

The scrutiny continues as organizations move toward a purchase. G2 identifies security review as a source of delay for 39% of buyers, budget approval for 32%, and implementation planning for 25%. These delays sit inside the buying organization and can persist even when vendor discovery is fast.

Finance is also taking a larger role in software decisions. Its participation in buying committees rose from 31% to 46% in a year, according to G2. G2 also found that 49% of buyers said their CFO had reversed a purchase the buying team had already approved, while three-quarters now expect positive ROI within six months of signing a contract. A product champion’s approval can therefore be one step in a broader process.

G2 found companies moving pricing logic, ROI calculators, case studies, benefits, and trials earlier in the process. Buyers facing financial scrutiny can use this material in internal discussions, while security and implementation stakeholders need evidence relevant to their decisions. Earlier access puts that information in stakeholders’ hands before the final transaction stage.

Downstream friction has different owners. Company policy, finance, procurement, security, and implementation planning can create delays inside the buyer’s organization. Sellers can add waiting when pricing requires a request, a quote needs manual preparation, or routine terms require negotiation. Leaders need to separate these causes before redesigning the process.

Routine transactions expose the largest avoidable mismatch

Cleverbridge’s “The Cost of Selling Software” reports a large gap between stated interest in self-service and adoption by sellers. Cleverbridge sells software-commerce infrastructure, so it has a commercial interest in wider use of self-service software purchasing. Its findings should therefore be read as vendor-sponsored evidence about buyer and seller behavior.

Cleverbridge also reports that waiting for a quote or pricing delays 36% of buyers, while 30% cite back-and-forth with sales over terms. These steps occur after discovery and can persist even when buyers already understand the product they want. They are a separate operational issue from finding and shortlisting software.

Cleverbridge reports that nearly three-quarters of buyers say purchases are significantly delayed or abandoned because of internal approvals or back-and-forth with vendors. Only 45% complete a routine software purchase within three business days, and a quarter wait at least a week. Internal approvals are included in these findings, so they do not isolate seller-created delay.

The seller still controls specific parts of the path. A buyer may have authority, accept standard terms, know the required configuration, and be ready to pay while still waiting for requested pricing or a manually prepared quote. Billing, payments, and coordination can add more processing. These are concrete targets for transaction redesign.

Self-service changes those mechanics for eligible routine purchases. A buyer can select a standard product or change, see the applicable commercial terms, complete checkout, and proceed without waiting for a seller to move the transaction between internal systems. Human assistance can remain available when needed. Company policy still determines whether the buyer can use that path.

Cleverbridge’s findings show that interest in independent purchasing extends beyond initial acquisition. Two-thirds of buyers would use self-service for renewals and upgrades or plan changes, 54% would use it for expansions, and 44% would buy new software independently. A transaction model centered on assisted new-logo sales therefore needs to account for other common purchase types.

Seller processing also consumes labor and money. More than half of sellers say they spend at least $1,000 internally processing a typical routine transaction, according to Cleverbridge, while more than a third spend six hours or more. Quoting, approvals, billing, payments, support, and coordination are among the activities involved. Leaders should ask which of those activities actually require human judgment for each eligible transaction.

Self-service is about transaction complexity

Price alone gives an incomplete picture of transaction complexity. More than half of buyers in the Cleverbridge research said they would be comfortable completing a routine self-service purchase worth at least $25,000. Another 17% would be comfortable doing so at $100,000 or more. Some buyers therefore accept self-service for substantial routine purchases.

A useful boundary is the amount of judgment and exception handling involved. A standard renewal with known terms, an additional seat, an upgrade, or another established product change can follow predefined rules. The customer may already understand the product, have an approved supplier relationship, and know the required change. Leaders can use those conditions to decide which transactions qualify for an automated path.

High-value purchases can also bring unusual legal terms, integration work, procurement requirements, or deployment risk. Those conditions create reasons for human review regardless of invoice amount. Price can remain one policy input, while routing reflects the exceptions and judgments people must resolve.

Keep people where the purchase genuinely needs them

Buyers still want expertise when transactions become difficult. Cleverbridge says more than half want human assistance available for large multi-year agreements, security or compliance reviews, procurement exceptions or special terms, and complex configurations. Nearly half want help with custom pricing and discounts. These cases can involve negotiation, interpretation, risk assessment, or solution design.

That supports a hybrid transaction model. Routine purchases can proceed autonomously when the customer’s policy permits it, while specialists handle exceptions and complex decisions. Security or compliance reviews can reach the appropriate expert, unusual procurement terms can go to people authorized to negotiate them, and complex configurations can receive sales or solution support.

The operating model must support the customer-facing checkout. Rules should determine which purchases qualify as routine and which conditions trigger review, while pricing, contracts, billing, and payments follow the same routing logic. A purchase button alone does not let an eligible buyer complete a transaction independently. Leaders should measure completion of the eligible transaction from selection through payment.

Main highlights

• Match transaction speed to discovery speed: AI is helping buyers form software shortlists faster, while evaluation, approvals, security reviews, contracting, and payments still slow completion. Commercial teams can examine these downstream stages for avoidable delays.
• Separate buyer delays from seller delays: Finance, security, procurement, and implementation requirements create internal friction, while manual quotes, pricing requests, and contract negotiations can add seller-side waiting. Revenue operations teams can identify the owner of each delay before redesigning the process.
• Automate eligible routine transactions: Buyers report strong interest in self-service, yet relatively few sellers offer it. Commerce teams can route standard renewals, upgrades, expansions, and other predefined purchases through self-service while retaining assisted paths for exceptions.
• Define self-service by transaction complexity: Purchase value alone does not determine whether a transaction needs human involvement. Sales and finance teams can set eligibility rules around standard terms, known configurations, established supplier relationships, and the amount of exception handling required.
• Put expertise where judgment matters: Large multi-year agreements, security reviews, procurement exceptions, complex configurations, and custom pricing still benefit from specialist involvement. Sales organizations can build hybrid models that automate routine processing and route complex cases to the right experts.