We build and operate intelligent systems where humans and AI work together.
Intelligence². Smart Humans × Smart AI.
Your business runs on data, CRM and increasingly on AI. The strategy lives in one place and the execution in another. We help your teams explore the new ground, then connect the two.
An AI and data company that proved itself in production first.
We design AI and data architectures, build the systems that run them, and operate those systems alongside our clients’ teams. We work across consumer brands, publishers, ecommerce operators, SMEs and global enterprises. Some of our work shows up as products you can buy. Some of it shows up as a working plan, an installed system, and a team that knows how to keep it running. The principle is the same in every engagement: humans and AI keep each other sharp, both stacks are vital, and the system has to earn its place on a P&L. We explore before we operate, and we operate what we have learned to trust.
The infrastructure layer. Sendability, our agentic email and CRM platform. BrandExpand, our human and AI content engine. Custom AI agents and data architectures we ship for clients who need a system that does not exist on the shelf. Internal tools that graduate into products when they earn it. Built and run by a small team that uses what it ships.
→ See what we buildThe production layer. Right now an agent inside our infrastructure is deciding whether to throttle a campaign. A human operator is reviewing it. The tension between them is the work. We operate deliverability infrastructure for enterprise senders, agentic content systems for publishers and creators, data pipelines for CRM teams, and our own Amazon FBA P&L (Throwback Traits), the lab where we test agentic operations on inventory we own before we recommend the same patterns to clients.
→ How we run productionThe thinking layer. We work as an extension of your AI, Data, CRM and Dev team in three movements. We help you explore (where AI fits in your work, what is genuinely new ground for your team, what to try this quarter without disturbing production). We help you compose (which workflows your humans should keep, which ones agents should hold, where the two should meet). We help you operate at confidence (a coherent system, a measured cadence, an architecture your team understands and can run). The work follows the conversation, not a deck template.
→ How we adviseThe public layer. The IESE Fellowship on agentic accountability, AI ethics and business sustainability (2026 to 2029). The Intelligence² Insights newsletter. Seth & Noēsis, the book on Human and AI co-evolution forthcoming May 2026. Speaking, writing, frameworks for the wider conversation. Advised by Dela Quist.
→ Read the researchWe work as an extension of your team.
We collaborate closely as an extension of your AI, Data, CRM and Dev team. Our approach is responsible, participatory and diligent. We align with your goals while protecting what matters to your customers: data, relevance and autonomy. We bridge AI partners with human operators because both stacks, biological and artificial, are vital to enhancing intelligence in production.
We deliver from inside your team’s process. Your tools, your meetings, your decision flow. Not a report handed over from outside.
We try things well before we put them into production.
We ship and we run, then we advise. Most relationships are measured in years, not engagements.
Explore well, then operate.
The hard part of bringing AI into a team is not the technology, it is the rhythm. Organisational research has known this since James G. March’s 1991 paper on Exploration and Exploitation: companies that only exploit what they know stagnate, and companies that only explore never compound results. Alex Osterwalder and the Strategyzer team have spent the last decade turning that idea into a portfolio practice that runs both engines in parallel. We bring the same logic to AI adoption. Most teams arriving at AI in 2026 are pushed straight from headlines into deployment, skipping the part where humans and agents learn each other’s edges. That is the part we like best. Small experiments before large rollouts. Measured trials before committed architectures. A team that has explored together is a team that operates with confidence later.
Production-grade choices, model-agnostic by design.
Three rules guide every infrastructure decision: it has to work in production, it has to respect the geography of the data, and it has to keep the client in control of their own intelligence over time.
Our architecture principle. Each client decides where their data lives and under which law. EU clients run EU infrastructure. Canadian clients run Canadian servers. Sovereignty is structural, not a checkbox on the contract.
Who hires us.
We work with consumer brands, publishers, affiliates, SMEs setting up their first CRM, ecommerce operators, and global enterprise brands including names like Nike. The settings change. The principle of building intelligent systems where humans and AI keep each other sharp does not.
“Data Innovation is the unique type of partner that thinks at the level of strategy and ships at the level of operations. We brought them in to become more efficient on how we run customer communications across multiple channels with millions of consumers that each engage in different proportions with our many brands. They worked side by side with our team and our Media Agency and we got a clearer model, more transparency and in-depth solutions to many aspects of our CRM. And we keep evolving our frameworks.”
Advised by Dela Quist on email and CRM strategy. “Florin’s ability to navigate complex data & systems and deliver actionable strategies is unmatched.”
A note from Florin.
I started Data Innovation in 2019 because I noticed there were many sub-optimal solutions in CRM. Customer acquisition was costly and getting harder. Activation, monetization and loyalty were where the unrecognised value sat. Six years later the question is the same and the conditions have changed. The systems are smarter. The decisions are faster. The temptation to remove the human from the loop is louder than it has ever been. Same as the “desire” to automate everything.
I do not believe that is the right tactic. I also do not believe gut feeling and instinct alone are a sufficient answer in 2026. I believe data can lie and humans too can “hallucinate”. Yet my strongest belief is that humans and AI are two different forms of intelligence that can co-exist, collaborate and co-create, and the companies that win the next decade will design their systems and their teams around that fact. There is a way to interact with AI so that we co-evolve, so that AI improves at speed and humans become deeper, sharper and better at the work that is ours to do.
Co-evolution is a working practice, not a slogan. It means humans keep the abilities that make us who we are: judgement, taste, narrative sense, the willingness to disagree with a chart. It means AIs keep the abilities that make them useful: pattern at scale, tireless recall, calm under load. The teams I see doing this well are the ones that explore together first. They try a small thing, they look at what happened, they notice what each side learned, and only then do they decide what to put into production. They do not rush. They are not behind. They are usually ahead, because their systems compound and their people know what they are running.
We run with a small, deliberate team. Iván on data engineering. Ziyad on platform, infrastructure and Mautic. Joel on integrations, agents and AI. Me on strategy and the deliverability craft. We collaborate with veteran experts when an engagement calls for them, including Dela Quist, who has shaped this industry for longer than most of us have been in it. We work with the AI partners we trust most, Anthropic and Google as our first options, while staying fluent in open-source models and the platforms our clients already run.
The thinking on this site comes from somewhere. Personally, that includes AI strategy training at MIT Sloan, both Artificial Intelligence: Implications for Business Strategy and Implementing Agentic AI: Building Your Organizational Playbook, programs at Singularity University, the IESE Research Fellowship I currently hold (2026 to 2029, Observatory for AI and Ethics in Organizations), and the daily practice of running production systems for clients who do not tolerate vague fearful answers. The book that comes out in Q2-Q3 is the long version. The newsletter is the short version. If you are working on a system where the human and the machine have to keep getting smarter together, we are probably trying to solve some of the same problems. Talk to us.
Florin Armasu · Data Innovation Founder & CEO · IESE Research Fellow
Talk to us about Sendability.
Our agentic email and CRM platform. Where Profitable Deliverability lives in practice. EU-hosted. Modular by design.
Go to Sendability →Talk to us about your AI strategy.
We work with founders, CTOs, heads of strategy and operations teams. The conversation starts with where you are. The work follows.
Book a strategy callOr read first: Insights · subscribe to Intelligence² Insights · pre-order Seth & Noēsis.
Thinking on AI, data and intelligent systems.
The Tableau For Email Marketing Analytics Playbook That Actually Works
Third Party Cookie Death CRM Impact: The Board-Level Conversation You Need to Have
The GEO Generative Engine Optimization Framework Senior Leaders Are Using Now
Email Deliverability Best Practices 2026: What Changed and What Still Works
Questions we hear often.
Quick answers to questions you may have. If you do not find what you are looking for, get in touch.
Contact →What does it mean to work with you as an extension of our team?
You mention AI and data, but where does email fit?
What is Geo-Patriation?
Do you work with companies that already have a data team?
Our practice draws on Ethan Mollick on the jagged frontier of AI capability, Jensen Huang on accelerated computing and intelligence factories, Andrej Karpathy on the rise of agentic software, Demis Hassabis on AI as scientific method and solving the world’s biggest problems together, Dela Quist on the email economy and the inefficiencies of the ecosystem that runs it, James G. March and Alex Osterwalder on exploration and exploitation, and Charles Sanders Peirce on creative sympathy as the primary force of evolution, rather than chance or mechanical necessity.