Zapote is your experienced AI advisory on the client side, with real technical depth. We define vision, priorities and architecture with you, prepare sound build-vs-buy decisions and support delivery: with your team, your vendors or ours.
Pick the situation that fits best. You will see where we typically start and what you have in hand after a few weeks.
Together with management and business units, we translate your corporate goals into an AI vision with clear priorities, rather than a list of tools.
We collect, structure and assess your ideas by value, feasibility, data availability and risk, and often find the opportunities nobody has written down yet.
Not every process needs an AI agent. We assess where classic automation is enough, where AI should assist people and where controlled autonomy creates real value.
We review your pilots for value, architecture, data quality, evaluation and operability, and tell you honestly which ones should scale and which should not.
We design what your product or process looks like with AI: from the perspective of users, data and architecture. The result is an MVP scope your team can build.
Agents that operate systems and prepare decisions need more than a good prompt. We clarify where agents make sense, how autonomously they may act and which controls stand in front of them.
We assess proposals, platforms and agencies with technical rigor. We receive no commissions or referral fees from vendors. We disclose our own delivery proposals.
External technical leadership of agreed AI initiatives: We coordinate technical workstreams with your teams and vendors and build up your internal capability.
Governance belongs in the architecture, not in a slide deck. We review how your AI systems handle identity, permissions, data flows, human oversight and traceability.
From the first vision to steering delivery, from one source. The same person who shapes the strategy with you also owns architecture, vendor selection and quality. There is no gap between slide and code.
We translate your corporate strategy into an AI vision that management, business units and IT stand behind together, with clear priorities and the courage to leave things out.
Structured discovery with your business units: we find the use cases with measurable value and assess them honestly for feasibility, data availability and risk.
What does your product or process look like when AI is used well? We design the target picture from the perspective of users, business and technology, and cut it into deliverable steps. This includes automating existing processes with AI.
We design an architecture that fits your IT landscape, can be operated securely and does not break with the first model change. We prototype ourselves where it speeds up decisions. This includes agentic AI with identity, policies and human oversight.
Build it yourself, buy standard software or deliver with a partner? We make the decision transparent: with a robust TCO model. We disclose vendor ties and our own commercial interests.
Can your team do it? Will the agency deliver what it promises? We assess capabilities, proposals and solutions with technical rigor, so you start with the right crew.
An independent outside view before a system goes into production or budget is scaled. We review code, architecture, evaluation, security and compliance readiness.
We sit on your side of the table. We coordinate agreed technical workstreams with your teams and vendors and make progress and risks visible.
External technical leadership of agreed AI initiatives, or a sparring partner for executive management and the board, within the agreed scope.
Strategy without delivery stays on the slide. That is why we see your AI initiatives through to production, in the model that fits your organization.
Ideal if you want to build AI as an in-house capability. The goal is that your team gradually no longer needs us.
Ideal when capacity is sourced externally. You get a technically versed advisor on your side of the table.
Ideal for fast results without an AI team of your own. Zapote delivery is one option. We compare it with alternatives and explain our recommendation.
No large project on speculation. You start with a clearly scoped entry and decide afterwards how to continue.
We understand your starting point, goals and constraints. You get an honest assessment, even if it is: “You don’t need us for this.”
Vision & Roadmap Sprint, Opportunity Discovery or Architecture & Vendor Review: clearly scoped, with a fixed outcome and low risk for you.
You get: a documented basis for decisions for management and relevant control functions.
As Fractional AI Lead, delivery coordination or Architecture & Quality Assurance, we take on technical leadership of delivery for agreed initiatives, with your team, your vendors or ours.
Shared goal: putting AI solutions into productive use – with agreed quality gates, governance and a documented handover.
We build internal capability, processes and roles until your team or your new Head of AI takes over. Afterwards we stay on board as advisor if you wish.
You get: a sustainable AI capability of your own instead of dependence on consultants.
From “we should do something with AI” to a prioritized, fundable roadmap.
Independent review of architecture, vendors and build vs. buy, before budget is committed.
External technical leadership of agreed AI initiatives. We develop the roadmap and budget planning, coordinate technical workstreams and are accountable for the agreed reviews and decision papers. Budget approvals, personnel decisions and operational responsibility remain with you.
From prototype to production pilot: with your team, your vendors or the Zapote team.
You have alternatives: large consultancies, AI agencies, freelancers or an internal hire. These criteria help you assess any offer, including ours.
| Criterion | Zapote | Check with other providers |
|---|---|---|
| Senior involvement | Technically led by Oscar | Who actually works on the project? |
| Strategy and architecture | From a single source, from vision to code review | Who owns strategy and technical delivery, and who connects the two? |
| Transparent vendor selection | No vendor commissions; own delivery interests disclosed | Which partnerships and compensation arrangements exist? |
| Time to start | Project start typically 2–4 weeks after commissioning, depending on scope and current capacity | When and with whom does the project actually start? |
| Cost & flexibility | Agreed per project | What costs and commitments arise over the term? |
| Knowledge transfer | Part of the agreed scope, with the goal of building internal capability | Which handover and enablement is included? |
| Hiring a Head of AI internally: when does it make sense? | Can make sense when permanent strategic and operational leadership tasks arise. Budget and time to fill the role need to be planned individually. | Which role, budget and timeline fit your initiative? |
Often the best solution is a combination: Zapote supports architecture and quality, a delivery partner implements, your internal leadership decides on priorities and resources, and your future Head of AI is onboarded.
Examples of technical work from Oscar’s professional career in leadership, CTO and engineering roles, including before his work at Zapote. These are not exclusively Zapote projects. Scope and goals for your project are agreed individually. We only name clients with their approval.
Work on AI strategy, governance, prioritization and an operating model. Supported teams through an AI champion program.
Work on data quality computation and an architecture using semantic search.
Development of an LLM-based classification system to automate manual workflows.
Whether assistant, process automation or autonomous agent: we anchor security, quality and compliance in the architecture from the start. This gives you technical foundations for your approval processes and traceable operation. A particular focus is governed agentic AI: agents in which identity, permissions, policies and human approval are part of the architecture from the start.
Identity, permissions and policies decide what AI may do. The model only proposes.
Evaluation, tests and monitoring instead of gut feeling: quality gates before every release.
We incorporate requirements under the EU AI Act, GDPR and, where relevant, DORA defined by your legal and data protection advisors into architecture, controls and documentation. Legal assessment and clearance remain with your legal and data protection advisors.
“For AI to work in practice, technology, priorities and technical leadership need to fit together.”
Oscar combines 12+ years of experience in software engineering and AI with executive strategy and production engineering. As Head of Business Unit Data & AI, as CTO of an AI consultancy and as co-founder & CTO of a digital health startup, he has built and led teams across AI, data, software and hardware.
He knows what a board meeting on AI strategy feels like, and what a deployment at two in the morning feels like. That is why he stays technically hands-on as an advisor: in architecture, prototyping, code reviews, evaluation and quality gates. At Zapote you work with him from the first call to the handover, with no junior layer in between.
Zapote Ventures is an independent AI advisory based in Berlin. We work on the client side: from AI vision, priorities and roadmap through solution design, AI architecture and build-vs-buy decisions to steering delivery, with your team, your vendors or ours. The offer on this website covers technical AI advisory and implementation for businesses.
Neither. We combine strategy with technical depth: the same person who shapes the vision with you also owns architecture, vendor selection and quality. Delivery can be done by your team, your vendors or us. What matters is what works best for you.
With a free 30-minute intro call. This is typically followed by a clearly scoped entry of 2 to 4 weeks, such as an AI Vision & Roadmap Sprint or an Architecture & Vendor Review. If it fits, we then support delivery for 3 to 12 months, for example as Fractional AI Lead.
We receive no commissions, partner bonuses or referral fees from vendors, platforms or agencies. We justify build vs. buy vs. partner decisions transparently with cost, risk, dependencies and time to value. If we recommend delivery by our own team, we disclose the reasons. Our own delivery creates a commercial interest; you make the sourcing decision.
Mid-sized companies and corporates with complex processes or digital products that want to use AI seriously but do not yet have established AI leadership or need an independent second opinion. Typical contacts are managing directors, CIO, CTO, COO, CDO or Head of Digital.
We take on agreed advisory and project services: roadmap, architecture, reviews and technical coordination, over an agreed period. Personnel, budget and operational decisions remain with you. Scope and working arrangements are agreed for each engagement.
Not every process needs an AI agent. We assess where classic, rule-based automation is enough, where AI should assist people and where controlled autonomy creates real value. If we do not recommend an agent, we say so openly.
Yes, it is a focus area. We design agents so that identity, permissions, policies, human oversight and audit are anchored in the architecture outside the model. The typical entry is a Governed Agentic AI Blueprint of 2 to 4 weeks.
Both. We are based in Berlin and work in Germany, Austria and Switzerland. Workshops and key milestones on site where useful, ongoing collaboration mostly remote.
30 minutes with Oscar Morillo Victoria: we discuss your starting point, your goals and the next sensible decision. You leave with an honest assessment, even if we don’t end up working together.