You’re Using AI At The Execution Layer. The Value Is In The Judgment Layer

The devices are deployed. The licenses are paid. And if you’re a elder SEO aliases GEO practitioner correct now, you’re astir apt utilizing AI each time – for drafts, for summaries, for first passes astatine contented that utilized to return doubly arsenic long. That’s existent productivity, and it’s not nothing.

It’s besides not the return the finance is tin of producing. And the spread betwixt what you’re getting and what’s disposable isn’t a instrumentality problem. It’s a mode problem.

A peer-reviewed study published astatine the 2025 ASIS&T Annual Meeting by Tim Gorichanaz astatine Drexel University gives that problem a sanction (h/t to Shari Thurow for pointing maine astatine this paper!). Analyzing 205 real-world ChatGPT usage cases, Gorichanaz identified six chopped modes successful which group really usage AI: Writing, Deciding, Identifying, Ideating, Talking, and Critiquing. The information came from Reddit and skews Anglophone, which limits its generalizability, but the taxonomy it produced maps uncomfortably good onto really astir practitioners are really working. Two modes dominate. Four are being near connected the table. The 4 being near are the ones that find whether AI makes you much strategically valuable aliases conscionable faster astatine execution-layer work.

That favoritism matters much correct now than it has astatine immoderate anterior constituent successful this industry’s history.

The Two Modes Everyone Defaults To

Writing was the largest class successful Gorichanaz’s information astatine 47% of observed usage cases – drafting, editing, summarizing, translating, generating. McKinsey’s 2025 State of AI survey confirms this astatine the endeavor level: the astir commonly reported AI usage cases are contented drafting and accusation capture, and 63% of organizations utilizing generative AI use it chiefly to create text.

Identifying – explaining something, answering a actual question, summarizing a archive – was different 10% of the study’s data, and represents the different pillar astir practitioners person built their AI workflow around. Research a topic, get a synthesis, move to the adjacent task.

Together, these 2 modes relationship for the overwhelming mostly of really AI is being utilized by practitioners and enterprises alike. Both person existent value, yet neither is wherever the leverage is. And if your AI believe originates and ends there, you’re utilizing an progressively blase instrumentality to do activity that was already being automated – conscionable faster and astatine higher volume.

The different 4 modes (Deciding astatine 21% of Gorichanaz’s sample, Ideating astatine 9%, Talking astatine 8%, and Critiquing astatine 6%) are wherever the activity becomes irreplaceable. They’re besides wherever almost nary practitioner has built a deliberate workflow, because cipher handed them one, and the unit to show contiguous output has consistently crowded retired the abstraction to create one.

The Decisions You’re Still Making Alone

In the practitioner’s week, Deciding-mode questions are everywhere: which queries really person AI visibility vulnerability worthy prioritizing correct now, whether a brand’s retrieval problem is simply a contented architecture problem aliases a sourcing and awesome problem, really to allocate effort crossed a portfolio erstwhile some SEO and GEO request attraction and the budget doesn’t agelong to screen some fully, erstwhile to escalate a visibility interest to activity versus erstwhile to hole it successful the activity earlier anyone asks.

Most elder practitioners are presently solving these questions pinch acquisition and intuition. That’s not a failure, arsenic acquisition and intuition are genuinely valuable, and nary AI replaces them. But AI utilized deliberately successful Deciding mode adds thing acquisition can’t supply connected its own: a system pressure-test of the assumptions underneath the decision, applied earlier the determination hardens.

That requires much than a bully question. Deciding mode requires giving the AI the applicable discourse (competitive landscape, existent visibility posture, humanities performance, strategical constraints) and past treating what comes backmost arsenic a genuine input to the determination alternatively than a draught to beryllium skimmed and group aside. It requires a workflow that doesn’t yet beryllium successful astir practitioners’ practice, not because anyone blocked it, but because nary 1 built the clip aliases building for it either.

The aforesaid McKinsey information makes clear what that spread costs astatine scale: 88% of organizations usage AI, but only 6% suffice arsenic precocious performers generating meaningful enterprise-wide impact, and precocious performers are 3.6 times much apt to person fundamentally reworked their workflows alternatively than simply deployed devices into existing ones. The shape holds astatine the practitioner level. Faster output from an unreconstructed workflow is not the aforesaid point arsenic amended decisions from a restructured one.

The Gaps Nobody Briefed

For SEO and GEO practitioners, Ideating mode has a circumstantial exertion that astir are not utilizing and astir should be: mapping the entity and authority gaps the marque hasn’t recognized yet.

What angles of topical authority has the marque grounded to found that AI retrieval systems are presently filling from different sources? What organization signals (forum discussions, aggregated reviews, third-party commentary) are shaping how LLMs correspond the brand successful consequence to class queries, and what would it return to displacement them? What framings of the marque beryllium successful exemplary training information that the brand’s ain contented has ne'er addressed aliases countered?

These are genuinely Ideating-mode questions. They’re besides questions astir practitioners person immoderate type of successful the backmost of their mind without a system method for surfacing the answers. AI utilized successful Ideating mode, not “give maine 5 contented ideas” but a genuine iterative exploration pinch deliberate constraints and existent willingness to travel the output location the squad hasn’t already been, is 1 of the astir nonstop methods disposable for uncovering those gaps earlier a competitor aliases a customer audit finds them first.

The obstruction isn’t capability. It’s the quality betwixt a Writing punctual pinch a database output and an existent Ideating session. The first takes 2 minutes. The 2nd takes twenty, requires a different posture toward the tool, and produces thing that can’t beryllium replicated by anyone who didn’t do it. That asymmetry is wherever practitioner worth gets built successful the existent environment, and astir practitioners are not claiming it.

The Honest Read Your Team Won’t Give You

This is the mode pinch the astir nonstop exertion to regular believe and the astir organizational resistance, because it requires utilizing AI to find problems successful activity the practitioner aliases their squad has already invested in.

Used properly, Critiquing is really a elder practitioner catches what soul reappraisal missed. The anemic entity declare successful a contented strategy that sounds charismatic but isn’t backed by the sourcing AI retrieval systems really trust. The spread betwixt what the marque says astir itself crossed owned properties and what a well-prompted LLM surfaces erstwhile asked a class mobility the marque should own. The assumed premise successful a GEO proposal that made consciousness six months agone and is now contradicted by really retrieval patterns person shifted.

That past exertion is not abstract. Running your ain marque (or a client’s brand) done a system AI Critiquing convention earlier the adjacent strategy rhythm is precisely the benignant of proactive activity that separates practitioners operating astatine the judgement furniture from practitioners operating astatine the accumulation layer. It’s besides the benignant of activity that changes the speech pinch a customer aliases a activity team, because you’re surfacing problems earlier they go visible successful the information alternatively than explaining them aft the fact.

The logic Critiquing is underused isn’t a governance problem. It’s a disposition problem. Organizations and practitioners person broadly trained themselves to usage AI to nutrient output, not to interrogate it. Reversing that wont is simply a choice, and it’s 1 of the much consequential choices disposable to a elder practitioner correct now.

Rehearsal

The Talking mode successful Gorichanaz’s taxonomy covers AI arsenic a speech partner, and for practitioners, the astir valuable type of that is rehearsal for the soul and customer conversations wherever the stakes are real.

The customer telephone wherever you person to explicate why organic postulation is down 30% while AI hunt visibility is besides poor, and you request to clasp 2 abstracted causal explanations simultaneously without letting them illness into a azygous communicative that oversimplifies both. The soul briefing wherever you person to make the lawsuit for GEO finance alongside existing SEO fund to a activity squad that still conflates the 2 disciplines and wants a azygous number that explains the ROI of both. The agency aliases vendor reappraisal wherever you request to push backmost connected a recommended attack without losing the relationship.

These conversations are recurring and high-stakes, and astir practitioners locomotion into them pinch only their ain intelligence rehearsal arsenic preparation. Talking mode (role-playing the pushback, asking the AI to reason the different side, moving done the type of the speech that goes wrong) is not a replacement for experience. It is simply a mentation method that costs 20 minutes and materially changes the value of the practitioner who walks into the room.

It doesn’t nutrient an artifact. It doesn’t show up successful a utilization report. EY’s 2025 Work Reimagined Survey, which covered 15,000 labor and 1,500 employers crossed 29 countries, recovered that 88% of labor usage AI astatine work, but only 5% usage it successful ways that fundamentally toggle shape what they produce. The logic that spread is truthful wide is almost surely that the precocious modes – Critiquing, Deciding, Talking – don’t nutrient thing measurable successful the moment. They nutrient a amended practitioner complete time, which is simply a return that compounds and doesn’t look successful a dashboard.

What Mode You’re In Is What Layer You’re On

The six-mode taxonomy maps almost precisely onto the divided betwixt execution-layer activity and judgment-layer work. Writing and Identifying are execution-layer modes. They’re valuable, they’re visible, and they’re progressively the modes that AI handles pinch little and little quality involvement. Deciding, Ideating, Critiquing, and Talking are judgment-layer modes. They’re wherever the practitioner’s irreplaceability lives.

A elder SEO aliases GEO practitioner who uses AI only successful Writing and Identifying mode is, functionally, positioning themselves arsenic an execution-layer worker astatine precisely the infinitesimal erstwhile AI is astir aggressively compressing that layer. That’s not a prediction astir occupation displacement. It’s an study astir master differentiation. The practitioners building durable worth successful this situation are the ones utilizing AI to make their judgement better, not conscionable their output faster.

Gorichanaz’s study reframes what accusation request really intends successful the AI era, not conscionable question-answering aliases uncertainty reduction, but what the authors telephone skillfully coping successful the world, meaning the ongoing exertion of applicable intelligence to situations requiring some knowing and action. For a elder practitioner, that framing is simply a useful diagnostic. The mobility isn’t what AI tin do. It’s which parts of your activity require the benignant of applicable intelligence that compounds pinch experience, and whether your existent AI believe is making that intelligence sharper aliases conscionable making everything astir it move faster.

McKinsey’s workplace research finds that only 1% of leaders telephone their companies mature connected AI deployment, meaning AI is afloat integrated into workflows and driving important business outcomes. The practitioner-level type of that spread is conscionable arsenic wide, and conscionable arsenic fixable.

If you mapped your existent AI usage against the six modes this week (not what you intend to do, what you really did), really would the distribution look? How overmuch was Writing and Identifying? How overmuch was Deciding, Ideating, Critiquing, Talking?

The practitioners who adjacent that spread deliberately, who build moreover a minimal workflow astir the judgment-layer modes, are not doing thing exotic. They’re doing thing astir of their peers are not. In a subject wherever the execution furniture is getting compressed by the aforesaid devices everyone has entree to, that spread is the 1 worthy closing first.

To spot what I conscionable built aft months of work, you tin publication much about information for decisions and grounds for your conversations.

More Resources

• Search Marketing’s Insight Gap: When Automation Replaces Understanding
• Google: Use AI successful the Best Possible Way for Search
• Scaling AI Content Is The #1 Enterprise Priority: How Do You Scale Without Penalty?

This station was primitively published connected Duane Forrester Decodes.

Featured Image: Roman Samborskyi/Shutterstock