Engageably

05 — Journal

← Back to Journal

Listen before the tool decision, every time

May 17, 2026 · 8 min read · Method · Benjamin Rodriguez

Executive TL;DR

  • Picking the tool before you study the work does not reduce risk. It compounds it.

  • A portfolio of pilots is not a strategy. The diagnostic gap is why most stall before they reach the P&L.

  • Governance bolted on after the fact will not fit the real process. The tool shapes the framework, not the work.

  • Listening sessions held before any vendor enters the room produce a different roadmap. Finance can fund it in tranches. Operations can absorb it. The initiatives that would have died quietly get killed early, on purpose, before the money is spent.

The Macro Picture

The sequence has been hijacked. Vendors demo before anyone defines the problem. A champion sees a slick agent, forwards a deck, and the initiative starts from the tool instead of the work.

The numbers show what that costs. 88% of organizations report regular AI use in at least one business function. But nearly two-thirds have not begun scaling AI across the enterprise. Only 39% report EBIT impact at the enterprise level (McKinsey & Company, The State of AI in 2025, 2025). MIT put a sharper point on it. 95% of organizations deploying generative AI saw zero measurable P&L return. The cause is a workflow learning gap, not model quality (MIT Media Lab Project NANDA, via Fortune, 2025). The failure is the sequence.

Vendors are selling executives certainty about solutions before anyone has sat with the people doing the work long enough to know what is broken.

My Thesis

Structured listening, done before any tool is evaluated, is the single most valuable act available to an operating executive starting an AI initiative. Not a workshop. Not a requirements call run by a vendor. A real diagnostic phase, owned inside the company, that produces a written record of how the work actually happens before anyone signs a license.

The tool is not the strategy. The diagnosis is.

For the CEO, this is the only way to know whether the initiative coheres or whether it is six demos in a trench coat. For the Chief AI Officer, it is the only way to know whether deployment will land in governed conditions.

Default Behavior Versus Effective Practice

The default starts with a demo. Someone inside the company sees one, gets excited, circulates the deck, and the initiative begins from that moment. Listening either never happens, or it shrinks into a requirements call the vendor runs. Which means the vendor is shaping the diagnosis of the problem their product is going to solve. The conflict of interest is not subtle.

Effective practice inverts it. Internal leadership owns the listening sessions. The work is to watch the job, map friction, and triangulate between the people who do it, the people who manage it, and the people who audit it. No vendor in the room. No tool on the table.

When the sequence is right, the downstream conversations sharpen. Procurement scopes tighter. Integrations get narrower. Governance fits the process that exists, not the one in the pitch deck.

P&L Impact in the Next 12 to 24 Months

The listening deficit shows up as financial exposure, not just operational friction. 42% of companies abandoned most of their AI initiatives in 2025, up from 17% in 2024. The average sunk cost per abandoned initiative reached $7.2 million (S&P Global Market Intelligence, 2025). Gartner projected that at least 30% of generative AI projects would be abandoned after proof of concept by the end of 2025, with deployment costs of $5 million to $20 million (Gartner, 2024). The exposure has four parts.

1. Sunk cost on misaligned tools. Licenses, integration work, change management, and training spent on a tool whose fit was never tested against the real workflow. When the company pauses or reverses the deployment, finance recognizes that spend as impairment, not deferred value.

2. Velocity tax. The delay cost when the team has to redesign a deployment mid-cycle because the original brief was wrong. Every week the rework runs is a week the competing initiative pulls ahead. The cost is real even when it never hits a write-down line.

3. Lost leverage. Companies that understand their own workflows before vendor conversations keep pricing and scope leverage. Companies that do not surrender it. Once a tool is embedded in a half-mapped process, the renewal conversation is no longer a negotiation.

4. Governance rework. The cost of retrofitting controls onto a system the company deployed before mapping its process boundaries. This is the most expensive category, because it usually arrives during an audit or an incident, when the response window is short and the options are narrow.

Structural Risks and Governance Gaps

The audit trail starts at discovery, not deployment.

Regulators and internal risk functions want to see that the company understood the problem before it chose the solution. Organizations that cannot produce a pre-tool diagnostic record are exposed when a regulator or auditor later examines the deployment. Not because the tool is wrong. Because the decision process is invisible. The artifact that protects the company is the one most teams skip.

Listening surfaces what a vendor matrix cannot.

Data access boundaries. Exception handling. Human override conditions. Escalation paths when the model is uncertain. These are not features on a comparison sheet. They are constraints you have to know before the system is designed, and they only emerge through structured conversation with the people closest to the work. Deloitte’s most recent survey shows only one in five companies has a mature governance model for autonomous AI agents, even as agentic usage climbs fast (Deloitte, State of AI in the Enterprise 2026, 2026). That is not a technology gap. It is a listening gap.

Skipping discovery compresses the control window.

When the tool decision comes before the workflow understanding, the team sets deployment timelines against a scope no one ever validated. Controls designed for the imagined process do not fit the real one. By the time the gap shows up in live use, pulling it back is expensive and the operational risk is already concentrated. Gartner expects over 40% of agentic AI projects to be canceled by the end of 2027, citing escalating costs, unclear value, and inadequate risk controls. It also warns that vendor agent washing is clouding tool decisions further (Gartner, 2025). The compression is structural.

Operating Model and Talent Shifts

Listening-first requires someone inside the company to own the diagnostic phase as a real capability, not a pre-sales courtesy. That is a role definition and a budget question.

The skill set is not the same as AI implementation. Workflow observation, friction synthesis, and stakeholder triangulation are closer to investigative work than to engineering. Companies that conflate the two will staff discovery with people whose incentive is to reach a tool decision, because that is what their performance review measures. The discovery then collapses back into procurement, and the sequence problem repeats. McKinsey is direct on the point. AI high performers are 2.8 times more likely to report fundamental workflow redesign, 55% versus 20%. Workflow redesign is the single biggest driver of EBIT impact from generative AI (McKinsey & Company, via CX Today, 2025). Redesign is what listening produces.

A listening-first company produces fewer pilots and more funded programs. The scoping work has already killed the initiatives that would have stalled.

The Playbook

Four moves. Each one shifts the sequence so the diagnosis precedes the tool, and the funding precedes the demo only once the company understands the work.

1. Run a pre-tool diagnostic before any vendor enters the room. Owner: CEO or CAIO Horizon: Immediate KPI: Workflow friction points documented before first vendor demo.

2. Separate listening from procurement. Owner: COO or CAIO Horizon: 30 to 60 days KPI: Discovery sessions run internally, not by a vendor.

3. Map human override and exception conditions before scoping automation. Owner: CAIO KPI: Exception conditions documented per process in scope.

4. Tie funding tranches to diagnostic completion, not demo milestones. Owner: CFO Horizon: Next planning cycle KPI: No tranche released without a written diagnostic artifact on file.

Executive Next Step

Before the next AI funding decision, ask one question. Where is the written diagnostic that says we understand the work this tool will touch. If no one can produce it, the initiative is not ready to be funded. It is ready to be listened to.

Sources

← Back to Journal