Matthew Talbot

@talbot_mcj

Co-CEO, Complexio Ltd

Monaco

Joined February 2012

603 Following

677 Followers

2.8K Posts

Matthew Talbot

@talbot_mcj

4 months ago

"Where are we on this?" The most-asked question inside any enterprise. Slack, email, standups, board prep. Every time, someone stops what they're doing to pull context from five places and assemble an answer. Finding out what's happening takes longer than deciding what to do about it. Stevie answers it directly. One question, one answer, grounded in the operation as it actually stands. Whether the context lives in a database, a document, an email thread, or a midnight message from a port agent. Every answer shows its sources. That's the condition that makes executives willing to act on it. #EnterpriseAI #Stevie

talbot_mcj's tweet photo. "Where are we on this?"

The most-asked question inside any enterprise. Slack, email, standups, board prep.

Every time, someone stops what they're doing to pull context from five places and assemble an answer.

Finding out what's happening takes longer than deciding what to do about it.

Stevie answers it directly. One question, one answer, grounded in the operation as it actually stands. Whether the context lives in a database, a document, an email thread, or a midnight message from a port agent.

Every answer shows its sources. That's the condition that makes executives willing to act on it.

#EnterpriseAI #Stevie

Matthew Talbot

@talbot_mcj

6 months ago

Every operation generates a continuous record of itself - equipment behaviour, process timing, exception patterns, the rhythm of how work actually moves through a business day, and the majority of organisations are capturing it as measurement rather than reading it as memory. Monitoring tells you when something crosses a threshold. Intelligence tells you what that crossing means in the specific context of your operation - what preceded it, what followed last time, and whether the pattern is accelerating or resolving. Those are different classes of information, and they lead to fundamentally different decisions. The industrial AI platforms converging on this are treating time series data as event streams rather than data points. The automated discovery of what constitutes a meaningful event, without having to define it in advance. An alert you configured tells you what you already expected to look for. A pattern your system discovered tells you something your operation knows that you didn't. That distinction is what Stevie is built on. Rather than surfacing alerts, it reads the operational record of a business - drawn from its own data, grounded in how that organisation actually runs - and surfaces what deserves leadership attention each day. Leaders can query it, ask follow-up questions, and go as deep as the situation requires. Not a dashboard. A conversation with your operation's memory. That self-knowledge compounds the longer it runs, it's an advantage that widens quietly. #EnterpriseAI #OperationalIntelligence #Complexio

talbot_mcj's tweet photo. Every operation generates a continuous record of itself - equipment behaviour, process timing, exception patterns, the rhythm of how work actually moves through a business day, and the majority of organisations are capturing it as measurement rather than reading it as memory.

Monitoring tells you when something crosses a threshold. Intelligence tells you what that crossing means in the specific context of your operation - what preceded it, what followed last time, and whether the pattern is accelerating or resolving. Those are different classes of information, and they lead to fundamentally different decisions.

The industrial AI platforms converging on this are treating time series data as event streams rather than data points. The automated discovery of what constitutes a meaningful event, without having to define it in advance. An alert you configured tells you what you already expected to look for. A pattern your system discovered tells you something your operation knows that you didn't.

That distinction is what Stevie is built on. Rather than surfacing alerts, it reads the operational record of a business - drawn from its own data, grounded in how that organisation actually runs - and surfaces what deserves leadership attention each day. Leaders can query it, ask follow-up questions, and go as deep as the situation requires. Not a dashboard. A conversation with your operation's memory.

That self-knowledge compounds the longer it runs, it's an advantage that widens quietly.

#EnterpriseAI #OperationalIntelligence #Complexio

Matthew Talbot

@talbot_mcj

6 months ago

The major platforms are accelerating on agentic AI. The direction is right, AI that acts and coordinates without waiting to be prompted is where operational value lives. What's being underweighted is access. Every agent operates on permissions. What it can read, what it can trigger, what decisions it can make without human confirmation. Those boundaries are set at deployment - often quickly, often broadly - because the goal is to demonstrate capability. The organisations getting durable value are treating access architecture as a first-order problem. Defining what an agent should know and where human judgment stays in the loop before anything goes live. The capability conversation has accelerated. The governance conversation determines whether that capability compounds or creates complexity. #AgenticAI #Governance #Complexio

Matthew Talbot

@talbot_mcj

6 months ago

We can see from the market that the enterprise AI conversation has matured faster than most expected. A year ago, the priority for leadership teams was demonstrating an AI position - pilots, partnerships, innovation labs. The audience was the board. Today, the questions coming from those same organisations sound completely different. They're asking about auditability. About how AI handles exceptions. About whether the system can explain why it made a specific recommendation, and what evidence supports it. The operational questions signal something important: enterprises are done treating AI as a strategic narrative. They want it embedded in workflows where decisions carry real weight. This operational reality is what @ComplexioAI was rooted in. Our focus from the beginning has been on building AI that comprehends context deeply enough to be trusted with consequential work - tracking dependencies, surfacing reasoning, attaching evidence to every output. Stevie is live across multiple companies now, and what we hear consistently is that the value shows up in the conversations people have with it. They ask questions about their business and get answers grounded in their own data, with the reasoning visible. That transparency changes adoption speed entirely. Quiet depth compounds. We're seeing that play out every day. #EnterpriseAI #Operations #Complexio

talbot_mcj's tweet photo. We can see from the market that the enterprise AI conversation has matured faster than most expected.

A year ago, the priority for leadership teams was demonstrating an AI position - pilots, partnerships, innovation labs. The audience was the board. Today, the questions coming from those same organisations sound completely different. They're asking about auditability. About how AI handles exceptions. About whether the system can explain why it made a specific recommendation, and what evidence supports it.

The operational questions signal something important: enterprises are done treating AI as a strategic narrative. They want it embedded in workflows where decisions carry real weight.

This operational reality is what @ComplexioAI was rooted in. Our focus from the beginning has been on building AI that comprehends context deeply enough to be trusted with consequential work - tracking dependencies, surfacing reasoning, attaching evidence to every output.

Stevie is live across multiple companies now, and what we hear consistently is that the value shows up in the conversations people have with it. They ask questions about their business and get answers grounded in their own data, with the reasoning visible. That transparency changes adoption speed entirely.
Quiet depth compounds. We're seeing that play out every day.

#EnterpriseAI #Operations #Complexio

Who to follow

JPFinnell

@JPFinnell

Bringing out the best in business harnessing global strategic partners @SAP | Building community and developing early talent @GrowSAP @JA_USA @NFTE @Wharton

Marty Beard

@martyjbeard

tech is cool. #philosophy is cooler.

Lori Breitzke

@loribreitzke

The Payments Expert- Keeping you ahead.

Matthew Talbot

@talbot_mcj

7 months ago

One month into 2026, and we’re building what makes AI usable in the real world: clarity people can trust. We built Stevie last year after seeing the same pattern across enterprises: execution is only part of the equation. What people need is AI that communicates clearly, attaches evidence, and explains its reasoning. That’s what builds trust. That’s what this work demands. The landscape shifts, client needs evolve, and what worked six months ago becomes the baseline. If you’re not learning faster than the market moves, you’re already behind. 2025 reinforced that. Every client engagement taught us something new, and every challenge we solved revealed the next opportunity to improve. That’s how meaningful products get built: stay close to the work, listen hard, and evolve fast. Stevie is a conversational layer for enterprise intelligence. Ask it about your business, and it responds from your data with evidence you can stand behind. That’s the mindset at Complexio: always improving, always listening, and ready for whatever comes next. Proud of what the team delivered. Excited for what we’ll create next. #EnterpriseAI #AgenticAI #Automation #Complexio

talbot_mcj's tweet photo. One month into 2026, and we’re building what makes AI usable in the real world: clarity people can trust.

We built Stevie last year after seeing the same pattern across enterprises: execution is only part of the equation. What people need is AI that communicates clearly, attaches evidence, and explains its reasoning. That’s what builds trust.

That’s what this work demands. The landscape shifts, client needs evolve, and what worked six months ago becomes the baseline. If you’re not learning faster than the market moves, you’re already behind.

2025 reinforced that. Every client engagement taught us something new, and every challenge we solved revealed the next opportunity to improve. That’s how meaningful products get built: stay close to the work, listen hard, and evolve fast.

Stevie is a conversational layer for enterprise intelligence. Ask it about your business, and it responds from your data with evidence you can stand behind.

That’s the mindset at Complexio: always improving, always listening, and ready for whatever comes next.

Proud of what the team delivered. Excited for what we’ll create next.

#EnterpriseAI #AgenticAI #Automation #Complexio

Matthew Talbot

@talbot_mcj

7 months ago

This distinction between analysis and visibility is critical. Most organisations don’t actually struggle with whether AI can be useful, they struggle with whether it can be trusted. And trust breaks the moment “understanding the context” is treated as the same thing as “exposing the context.” As Danny highlights in the episode, AI has to understand the full operational reality to be genuinely effective: the handoffs, constraints, exceptions, and the sensitive inputs that shape decisions. But that doesn’t mean every user (or every workflow) should inherit broad visibility just because the system needs depth. The scalable approach is simple, let the system understand broadly, but reveal narrowly. What people see, and what the system can do, should stay governed by role, entitlements, and audit trails. #EnterpriseAI #Automation #Governance #Understanding #DigitalTransformation

Complexio

@ComplexioAI

7 months ago

Deep context does not have to mean open access. To drive genuine efficiency, an intelligent system must be able to analyse the full scope of a workflow, including the sensitive data that shapes decision-making. However, the ability to analyse must be distinct from the permission to view. Complexio is engineered to bridge this gap. It allows the system to leverage the organisation’s full operational reality to build and execute workflows, while strictly governing the output. The architecture ensures that while the system understands the necessary context, the employee only interacts with the data that their specific product entitlements and role permit them to see. This is governance built for the reality of modern enterprise operations: absolute precision in who sees what, without losing the value of the whole. Listen to the Complexio Podcast. Spotify: https://t.co/Ai56TuuCh7 Apple Podcasts: https://t.co/52SgQYTFIT #EnterpriseAI #Podcast #Automation #IntelligentAutomation #DataGovernance

205

Matthew Talbot

@talbot_mcj

8 months ago

@ComplexioAI We often mistake systems of record for reality. A database sees the transaction but misses the urgency and intent buried in unstructured communication. We capture this invisible layer, ensuring AI understands the business context before it acts. #EnterpriseAI #Context #Complexio

Matthew Talbot

@talbot_mcj

9 months ago

This is the gap we are focused on. An intelligence layer that learns how your business really operates from live data. With that shared context, automation and orchestration stop being brittle one-offs and become governed, testable and safe to scale. #EnterpriseAI #Complexio

Complexio

@ComplexioAI

9 months ago

Most enterprises struggle to see a return on AI. The issue is not a lack of tools, but a lack of business context. Current AI adds friction through prompting and siloed systems. Complexio removes it by understanding how and why your business operates. #Complexio #EnterpriseAI