Operational intelligence for the work industrial companies cannot see.

The operating thesis

Plexo helps oil & gas service companies uncover where operational inefficiencies are hurting their margins, redesign those processes, and eventually execute eligible workflows with AI agents.

Category: AI-native operational transformation for mid-market industrial companies — creating demand by making continuous process discovery and redesign economically viable for companies that often do not do it systematically today.

The initial wedge has service-like characteristics, but the intended model resolves toward recurring software and agent execution rather than remaining consulting.

“McKinsey that interviews your entire company — then deploys agents to fix what it finds.” This is a metaphor, not a claim that Plexo currently provides McKinsey-equivalent outcomes.

Initial ICP

Oil & gas service providers with approximately 200–2,000 employees.

Economic buyer: CEO. Champions: COO; Innovation / Transformation leaders. High operating costs make operational efficiency important to EBITDA and, where capacity is constrained, potentially revenue capacity.

Leadership often knows which processes create friction but lacks granular visibility into where time is lost, why a workflow is inefficient, what it costs, which process should be fixed first, and what should be redesigned versus automated.

One concrete workflow surfaced is the manual work required to keep documentation in order and upload or update it across customers’ service/vendor portals.

The problem is not that these companies lack automation ideas. They lack an accurate picture of how work actually happens.

• Validation priority: quantify hours, people, frequency, cost of delays, and economic impact for the first target workflows.
• Warm access to ~15 ICP companies comes through former Zent customers, family relationships, and close personal networks in oil & gas.
• ~15 is warm access — not customers, pipeline, commitments, or revenue.

A discovery gap

Making organization-wide operational discovery cheap and scalable enough to become continuous rather than episodic.

Traditional process transformation can require consultants or internal teams to interview employees, conduct workshops, document workflows, and manually identify opportunities.

Process-mining software is strongest where work leaves clean system event logs, while industrial workflows also span people, conversations, spreadsheets, portals, field operations, and undocumented decisions.

Plexo’s thesis is that AI changes the economics of operational discovery: voice agents interview employees, models synthesize those conversations and supporting information, and Plexo reconstructs how work actually happens.

AI interviewing itself is not a durable moat. Plexo needs vertical operational data, historical process knowledge, measured improvement outcomes, and execution context.

• Alternatives: Horizon / Ontora / Foaster; Celonis / process mining; SAP Signavio / BPM; consultants; internal transformation teams; and doing nothing.
• Closest AI-native competitors validate the category. Plexo must differentiate through vertical depth, quantified economics, and ultimately execution.
• The belief that management commonly does nothing despite known friction is a validation priority for the initial discovery motion.

Sources: Horizon — AI-powered discovery platform · Ontora — employee interviews and synthesis · SAP Signavio — process transformation suite

Three capability curves

Voice agents can conduct structured conversations at low enough marginal cost to make interviewing an organization more scalable than consultant-led discovery.

Modern language models can synthesize unstructured conversations and documents into structured operational representations.

Agentic systems increasingly interact with the software companies already use, creating a path from understanding workflows to executing eligible parts of them.

The opportunity shifts from “Can AI execute a task?” toward “Does AI understand what work needs to be executed, including exceptions and context?”

• Planning assumption: $25K average initial discovery engagement. 40 engagements/year × $25K = $1M annual discovery revenue.
• Hypothetical: 50% expand into a recurring relationship at $100K ACV: 20 customers × $100K = $2M ARR.
• Validation priorities: initial ICP count by geography, pricing, attach/expansion behavior, and comparable economics.

Current product flow

Define an objective such as “Understand our customer-documentation process.” Plexo identifies or is given relevant employees, voice agents conduct interviews, and the system cross-references perspectives.

The output is a reconstructed current-state process, bottlenecks, handoffs, redundancies, friction, quick wins, automation opportunities, and a redesigned future-state process.

Today’s scope includes interviews, extraction, synthesis, reconstruction, comparison, initial bottleneck identification, and report generation. Human QA, process validation, and operational judgment remain part of delivery.

• Human QA, process validation, and operational judgment remain explicit delivery gates.
• The product flow describes application scope; it is not presented as customer traction or a measured delivery claim.

Aspirational planning targets, not forecasts

End of 2027 target: Plexo is the best operational-discovery and redesign product for oil & gas service companies. Planning targets: $4M revenue, ~$2M EBITDA, 50–75 customers, and <20 headcount.

End of 2031 target: Plexo has expanded into an operating layer for AI-native industrial companies. Planning targets: $25M+ revenue, $10M+ EBITDA, 250+ customers, and <75 headcount.

Customers enter through discovery and can expand into continuous optimization and execution. The targets are the operating ambition against which the model will be tested.

Pre-revenue validation plan

Status: pre-revenue.

A recurring problem hypothesis in oil & gas service operations is that leadership can see operational friction but may lack a systematic way to determine where time is being lost, what it costs, and what should be improved first.

Initial distribution is founder-led and warm-network driven. The immediate goal is to convert ~15 warm paths into discovery conversations and design partners; a repeatable long-term channel is not yet proven.

• Current starting point: ~15 warm ICP paths; paying customers: 0. Warm access is not pipeline, commitments, or revenue.
• The validation program focuses on ICP interviews, workflow economics, willingness to pay, paid design partners, and a repeatable founder-led conversion motion.
• Key question: will a CEO pay for operational discovery/redesign as a standalone wedge, or must Plexo sell a tightly quantified EBITDA/capacity outcome from day one?

Working economic model

Initial pricing hypothesis: a $25K fixed-fee Operational Efficiency Diagnostic, potentially covering interviews, current-state reconstruction, bottleneck analysis, quick wins, automation opportunities, future-state redesign, and eventually quantified impact.

Long-term recurring relationship hypothesis: $50K–$150K+ ACV depending on workflows Plexo continuously monitors, improves, or executes.

Variable costs include voice/inference, human delivery, and implementation. The economic thesis is that discovery should increasingly scale with voice/inference usage plus declining human QA.

• Planning target: 70%+ discovery gross margin as AI handles more interviews and analysis without proportional analyst headcount.
• Planning target: 80%+ recurring gross margin where workflows are predominantly automated.
• The operating model will be built from observed interview volume, voice and inference usage, QA, onboarding, implementation, ACV, conversion, expansion, churn, and hiring inputs.

Land and expand

The initial engagement creates operational context that should make subsequent work inside the same company faster and more valuable.

Expansion path: one process → multiple processes → department → cross-functional operations → continuous optimization → agent execution.

More workflows per customer and higher-value execution revenue are the two intended compounding mechanisms.

The more of the organization Plexo understands, the easier it should become to understand, redesign, and automate the rest.

Operational Efficiency Diagnostic

Plexo enters through a narrowly scoped operational-discovery engagement. Planning price: $25K. Target time-to-value: 2–4 weeks. Both are hypotheses.

Output: current-state map → bottlenecks/friction → quick wins → automation opportunities → redesigned future state → eventually quantified operational impact.

Resolution: paid diagnostic → implementation → recurring optimization / agent execution → more processes → broader operational layer.

• Closed doors: generic SMB process mapping; arbitrary AI automation agency work; becoming a diagramming/BPM product; horizontal enterprise GTM from day one.
• Intended stickiness comes from execution and accumulated operational context; this is a hypothesis, not demonstrated retention.

Founder–market fit

Plexo emerged from building Zent and deploying AI automation systems for companies. The recurring challenge was not only building automation, but understanding work that was undocumented, held in people’s heads, crossed tools, and contained hard-to-capture exceptions.

Before an AI agent can reliably operate a company workflow, it needs an accurate representation of how that work actually happens.

Team archetypes: CEO / sales creates demand; Bautista Silvano, technical / product, creates value; Francisco Sobral, product / operations, creates leverage.

The team explored due-diligence/certification-oriented workflows before moving toward the underlying process-discovery layer and narrowing GTM to oil & gas services. This is iteration evidence, not product-market fit.

Biggest risk: the team has identified a real operational problem before proving it is a standalone budget. Discovery must be tied quickly and credibly to measurable EBITDA or capacity improvement.

The ask

The central question is how to turn Plexo’s technical capability into the right economic machine: an initial discovery wedge that creates value and gross profit quickly while resolving into high-margin recurring revenue.

• Pricing: price around customer value and EBITDA/capacity impact, not seats, hours, or raw usage.
• Wedge design: pressure-test an engagement that is easy for a CEO to buy, valuable in weeks, gross-profitable, and evidence-generating for expansion.
• Revenue architecture: turn one-time discovery into recurring optimization and agent-execution revenue without becoming a consulting company.
• GTM and economics: turn warm access into a rigorous founder-led motion, then build a repeatable channel and capital-efficient model.

We know how to build AI systems, and we have warm access to the first customer set. We want Gateway X to help us build the economic machine around Plexo.