Un ingeniero senior de IA en Microsoft acaba de revelar cómo los equipos de Microsoft crean agentes de IA con Anthropic.
34 minutos de taller gratuito, directamente por el equipo de Microsoft.
Opus 4.7 + más de 1 400 herramientas MCP ya listas para usar.
Conectas Claude a un agente → le agregas herramientas → lo despliegas en producción.
Más útil que la mayoría de las formaciones de vibe-coding vendidas por 500 $.
guárdalo y sígueme para más → @karlarboledas
Worried AI customer service will feel cold, generic, or make your customers repeat themselves? 😬 You're not alone.
Watch the full conversation: 🎥 https://t.co/6iB8Rsc6EH
7 Questions to Choose a HubSpot Onboarding Partner - Discover 7 key questions to evaluate HubSpot onboarding partners. Learn what mid-sized B2B teams should ask about experience, integration, and support.
#hubspotpartner#hubspotagency#hubspot#revops#growbetter
Más leads no es la meta. Los leads correctos, sí. No se trata de tener volumen de contactos, se trata de saber quiénes realmente tienen el perfil para convertir, sea una venta, una donación o cualquier acción que busques. #HubSpot#LeadGeneration#RevOps#MarketingAutomation
Ya personalizas tus emails con el nombre del contacto. Es básico, pero funciona.
Ahora imagina llevar eso a tu sitio web, la misma lógica de personalización, pero en cada punto de contacto digital, no solo en el correo.
#HubSpot#Personalización#MarketingDigital#RevOps
Google ya no te manda a un sitio. Te da la respuesta directo. Cada vez más búsquedas se resuelven sin un solo clic. El reto ya no es solo aparecer en resultados: es que TU contenido sea la fuente de esa respuesta.
#SEO#AIO#MarketingDigital#RevOps#ContentMarketing
Gartner expects 40%+ of agentic AI projects to be cancelled by the end of 2027. Their reasons: escalating costs, unclear business value, inadequate risk controls. Notice what is missing. Nobody says the models were not good enough.
#HubSpot #AgentHub#HubSpotPartner #RevOps
Our 90-day rollout, and the part everyone skips.
Days 1 to 30: data and definitions. Nothing live.
Days 31 to 60: two agents, 50-case eval set.
Days 61 to 90: supervised launch, then scale.
Nothing touches a customer before day 61.
The work does not disappear. It changes shape.
Selling time goes up. Research and CRM admin go down.
And a new 12% block appears: supervising agents.
Budget for it or lose the gains to quality drift.
Sequence agents by blast radius, not excitement.
Start here: read-only data agent, KB gap drafting, content repurposing, pre-call briefs.
Not yet: unattended outbound, autonomous discounting.
That last one is the whole game.
Deflection rate AND CSAT.
Meetings booked AND unsubscribe rate.
Records enriched AND records wrongly overwritten.
Measure only the output metric and you will efficiently optimise for a number that hurts you.
Every agent we deploy gets seven fields on one page.
Objective
Scope (the “out” list matters more)
Inputs
Tools, with writes listed separately
Definition of done
Escalation triggers
Two metrics
Three different animals, one budget line, which is why most AI initiatives stall at pilot.
Workflow: you wrote the rules. Fails silently.
Copilot: human decides, model drafts. Fails visibly.
Agent: reasons toward your goal. Fails confidently, at volume, unattended.