By Krupesh Bhat, Founder and CEO, Melento (formerly SignDesk)
Every CEO knows how to measure revenue. Every CFO knows how to measure cost. Every COO has a familiar set of productivity metrics. Yet very few organisations can tell you how much time their business loses while work is simply waiting. That is the Waiting Tax.
It sits between functions, systems and decisions: an approval sitting in an inbox, a contract waiting for legal review, a supplier waiting to be onboarded, a loan application waiting for compliance, or an employee waiting for information before the next step can begin. Individually, these delays appear insignificant. Across thousands of transactions, they become a material drag on growth.
For decades, organisations have treated the resulting waiting as the unavoidable cost of scale. The result is a paradox: an organisation can make every department more productive while making the business as a whole slower.
I believe AI is challenging that model. Its most consequential contribution may not be helping individual employees complete tasks faster but allowing work to move continuously across functions. That is the beginning of what I call the throughput economy.
Over the past two decades, enterprises have invested heavily in ERP, CRM, workflow and collaboration platforms. These technologies digitised work, but much of the underlying operating model remained unchanged.
Consider a customer onboarding journey. Sales may complete its work efficiently, but the customer still waits for legal, compliance, finance and operations to complete theirs. Similarly, a supplier may have submitted every required document, yet remain inactive while procurement, finance, legal and quality conduct separate reviews.
Each function may be operating within its SLA. The customer or supplier experiences one organisation, not four departments. This is enterprise latency: the time work spends waiting between productive actions.
As organisations become larger and more regulated, latency compounds. We have become adept at managing these delays through escalation emails, SLA dashboards and follow-ups.
The case for throughput has become stronger because three forces are converging: complexity, regulation and customer expectations.
Businesses today operate across more systems, markets, partners and regulatory frameworks than ever before. At the same time, regulators expect stronger controls and better auditability. Customers, meanwhile, increasingly expect the speed and simplicity of the best digital experience they encounter anywhere.
BFSI illustrates the tension particularly well. A lender needs to collect documents, verify them, perform KYC and risk checks, satisfy compliance requirements and obtain approval. But if each activity happens sequentially, the customer can spend days waiting for a decision that may require only minutes of actual human judgement.
AI-led orchestration can change that equation by connecting document verification, compliance checks and approval workflows, allowing activities that can happen concurrently to do so, while routing exceptions to the right human decision-maker.
Manufacturing presents a similar challenge. A delay in onboarding one supplier can have consequences far beyond procurement, affecting inventory, production planning and customer commitments.
In both cases, the competitive advantage comes from compressing the time between intent and outcome, without compromising the controls that make the process trustworthy.
Automation improved this model by removing repetitive manual tasks. AI introduces something more significant: the possibility of coordinating work across the enterprise through digital labour.
Documents can be interpreted and verified. Business rules can be applied. Approvals can be initiated. Exceptions can be detected. The next action can be triggered without someone having to manually monitor the previous one.
The important change, therefore, is not simply that a task takes ten minutes instead of twenty. It is that the ten-hour gap between two tasks can potentially disappear. That is the difference between productivity and throughput.
Productivity asks how efficiently a resource performs an activity. Throughput asks how efficiently the organisation converts work into outcomes.
For years, the management question has been, ‘How do we improve productivity?’ AI requires leaders to add another question: Where does work wait?
Leadership teams adopting AI-led workflows should begin by mapping the points where value repeatedly gets trapped: approvals, handoffs, document dependencies, exception handling and decisions that unnecessarily happen in sequence. They should then ask four practical questions.
Where is the largest operational latency?
Which delays genuinely protect the business, and which exist because processes were designed for an earlier operating environment?
Where can AI coordinate work without removing necessary human judgement?
And, most importantly, can we measure the business outcome?
This requires a shift in how AI investments are evaluated. Leaders should increasingly examine cycle time, first-time-right rates, exception resolution, decision velocity and the time between customer intent and business outcome.
The emerging AI economy may reward flow. The winners will be companies that understand where work gets stuck and redesign those points of friction.
The competitive question will increasingly move from “How much work can our organisation handle?” to “How quickly can our organisation turn demand into outcomes?” That is the essence of the throughput economy.
And from what I have seen building AI-led enterprise workflows, the most valuable AI transformation may ultimately be less about making individual people faster and more about making the organisation itself move faster.
The real competitive advantage will belong to those who learn to treat waiting not as an unavoidable cost of doing business, but as an operational problem that can finally be engineered away.