Frequently asked questions

An agentic AI makes autonomous decisions within defined parameters — it executes actions in real systems, integrates data from multiple sources, and learns from outcomes. A chatbot only responds. In mid-size companies, agents reduce human intervention on repetitive tasks by 40-60%, while chatbots require constant manual escalation.

Between 6 and 12 weeks for well-defined use cases — customer service, request processing, data automation. We've measured operational cost reductions of 25-35% in those specific areas, with typical payback in 9 months for companies with 200+ employees.

Discovery and design: 2-3 weeks. Development and integration: 4-6 weeks. Testing and refinement: 2-3 weeks. Total: 8-12 weeks to first agent in production. Subsequent projects deploy faster — 60% of time is spent understanding the process, not writing code.

Yes. We work with SAP, Oracle, Dynamics, Salesforce, and local systems. Integration depends on API quality — modern systems take 2-4 weeks, legacy systems require up to 8 weeks. 78% of our mid-size clients operate with two or more systems simultaneously without conflict.

Data is processed under complete company control — on-premises, private cloud, or hybrid. We comply with GDPR, LSSI-CE in Spain, and LATAM standards. Zero data sent to third parties. The agent accesses only authorized information via credentials and granular permissions you define.

A routine agent requires 0.5 to 1 FTE for monthly monitoring and adjustments — can be a junior analyst with 4-5 hours weekly. A complex agent: 1-2 FTE. Contrary to popular belief, you don't need an ML/Data Science team after initial deployment.

Retail, logistics, financial services, and hospitality — 10-15 weeks. These sectors have 3-5x higher transaction volume and repetitive processes. Manufacturing and construction require deeper customization — 16-20 weeks, but annual savings exceed €300K in 300+ person companies.

€25K to €50K for first scoped agent — one specific use case in one process. This includes design, development, integration, and 3 months of support. Larger scales (3-5 coordinated agents): €100K-€200K. Companies investing less than €25K typically see limited results.

Boutique structure — small teams, fast decision-making, obsession with business metrics not pure technology. 100% of our projects have OKRs defined pre-launch. We don't build decorative proofs of concept — only viable deployments with measurable ROI in 12 months.

1) Discovery: map processes, identify friction, estimate savings (2 weeks). 2) Design: agent architecture, integrations, security (1 week). 3) Build: development and integration (4-6 weeks). 4) Pilot: with real data, active metrics (2-3 weeks). 5) Launch and monitoring (ongoing). 35% of value is captured in the pilot.

Off-the-shelf: quick (2-4 weeks), but flexible for only 40-50% of real use cases. Custom: requires 8-12 weeks, but adapts 90%+ of specific workflows. Initial cost similar (€30-60K), but generic tools scale poorly with volume — custom maintains flat or decreasing cost per transaction.

Yes, both regions. Distributed teams in Spain, Colombia, Peru, and Argentina. Local regulatory compliance: GDPR, LSSI-CE in Spain; data protection regulations in LATAM. Timezone aligns with client — less friction than global partners.

1) Transaction volume without human intervention (target: 70-85%). 2) Cost per transaction (pre/post comparison). 3) Resolution time (typical reduction 60-75%). 4) User satisfaction (if customer-facing). 5) Team time freed (hours/week liberated). Companies measuring only 2 of these metrics miss 40% of potential value.

The agent is designed with guardrails — it categorizes incidents and escalates to human automatically. In finance, approvals above threshold go to human. In customer service, complex queries defer. Typically 5-15% of cases require intervention — they're cheaper to resolve because the agent already processed 85-95% of initial work.

Between 3 and 8 coordinated agents is standard in 200-500 person companies. Each agent specialized in one workflow (service, procurement, accounting, HR). Companies with 5+ agents see synergies — they reuse 40-60% of code and shared data, reducing new agent deployment to 3-4 weeks.

Not necessarily. Agents function as a layer on top of existing systems. If infrastructure has robust APIs, integration is clean. If purely legacy, middleware is required — additional cost €10-20K, timeline +2-3 weeks. 70% of our clients preserve their current stack.

Misaligned expectations — executives expect 100% automation, reality is 70-85%. Second risk: incomplete training data — requires 1-2 weeks of audit. Third: lack of operational change in human team. We mitigate with clear governance from week 1 and user training.

Modular architecture — components can be retrained without touching core logic. Minor changes (10-20% of workflow): 1 week. Major pivots: 3-4 weeks. Contracts include SLA for updates — we guarantee agent functionality with normal operational changes remains cost-free for the first 12 months.

Financial services, retail, and logistics — high transaction volume, repetitive processes, clean data. In these, time-to-value is 8-10 weeks and annual ROI exceeds €100K in 300+ person companies. Manufacturing requires more configuration but also generates larger savings — depends on priority: speed or maximum impact.

First 90 days: weekly syncs, performance tuning, team training. Months 4-12: monthly reviews, adjustments as workflows shift, SLA guarantee of 99% uptime. Beyond 12 months: transition to managed support or in-house team with our training. Total support cost: 15-20% of initial project fee annually.