Markets converge before company labels agree. The strongest early signal is not shared vocabulary; it is multiple products moving toward the same buyer job, workflow control point, commercial unit, and proof requirement. Track those dimensions over time and treat category names as evidence, not as the boundary of the market.

What this guide covers

  • why vendor categories lag product reality;
  • four dimensions that reveal convergence;
  • a method for separating adjacency from direct substitution;
  • a convergence map and decision rules.

Category labels are deliberately strategic

A company chooses language that helps it enter a comparison, escape one, or enlarge the budget it can address. Two products may use different labels while competing for the same workflow. Two others may use the same fashionable label while solving different jobs.

The official offers make the boundary problem visible. GitHub Copilot plans describe capabilities for software work, Replit pricing combines building and deployment economics, Gumloop pricing exposes automation-oriented units, and Intercom's pricing FAQ describes customer-service economics. These pages establish current offers, not a universal taxonomy. Whether the products converge depends on the decision being made.

Track four convergence dimensions

1. Buyer job

What progress is the buyer trying to make? “Use AI” is not a job. “Resolve support demand without proportionally adding agents” and “ship an internal workflow without waiting for engineering capacity” are closer.

2. Control point

Where does the product sit in the workflow? A tool that drafts an answer, one that executes a task, and one that governs the system may touch the same process but own different control points.

3. Commercial unit

What does the vendor meter: seat, task, execution, token, minute, resolution, environment, or outcome? Commercial units reveal which value and cost boundary the vendor is attempting to own.

4. Proof requirement

What must the buyer believe? Individual productivity may need a convincing trial. An enterprise system may need controls, auditability, reliability, deployment choices, and procurement evidence. Convergence becomes more consequential when proof standards also align.

Build a convergence map

Do not put companies in mutually exclusive boxes. Score their current movement against a named job.

CompanyBuyer job overlapControl-point overlapUnit overlapProof overlapDirectionEvidence date
Ahighmediumlowmediummoving closerYYYY-MM-DD
BmediumhighhighlowstableYYYY-MM-DD
ClowlowmediumhighadjacentYYYY-MM-DD

Use qualitative scores with written definitions. False numerical precision is not analysis. The value comes from retaining the observations that caused a score to move.

Distinguish three market relationships

  • Surface adjacency: products appear in similar content or use the same label but do not displace each other in a real decision.
  • Workflow adjacency: products own neighboring steps and may partner, bundle, or expand into one another.
  • Decision convergence: the same buyer can fund one instead of the other for a defined job.

Only the third is direct competition for that decision. Workflow adjacency still matters because the owner of the strongest control point may expand, change the bundle, or turn a partner into a feature.

Look for coordinated movement, not isolated launches

Confidence rises when several dimensions move in the same direction:

  • a product launch expands control of the workflow;
  • pricing adopts the same value unit as an adjacent category;
  • integrations become native orchestration rather than handoff;
  • enterprise controls appear around the new job;
  • positioning addresses the adjacent buyer;
  • customer proof demonstrates substitution or consolidation.

A single integration or landing page should not redraw the market. Persistent, corroborated movement should.

Translate convergence into choices

Use convergence analysis to choose a review, not to create a dramatic market map.

PatternUseful strategic question
same job, different control pointwhich position captures durable workflow leverage?
same control point, different unitwhich commercial model better matches value and cost?
same unit, different proofare we serving a different risk tolerance or segment?
all four moving togetheris a new direct peer set forming?
partner moving toward our jobshould the relationship, boundary, or roadmap change?

The EXVIV market map supplies a starting taxonomy, while the company directory and Evidence Compare help inspect the participants and movements. Use the monitoring guide to watch only relevant control points and the living competitor matrix to keep categories revisable.

Keep the category falsifiable

Write inclusion and exclusion rules. Name the job, buyer, control point, and substitutes. Add a company only with evidence that it participates in that decision. Remove it when the product or buyer boundary changes. Record disputed cases instead of forcing agreement.

This makes a market definition useful even before consensus language emerges. The map can evolve without pretending that yesterday's category name was an objective fact.

Sources and further reading

Method note

This is a decision-based market definition, not a claim that the cited companies all compete in every use case. Offer terms and product boundaries change; verify them at the analysis date and preserve alternative classifications.