Agentic AI Experience Strategy

Defining the agentic AI experience strategy for McKinsey's client development platform — a fundamental shift from passive CRM to proactive relationship intelligence. The system, named Activate, surfaces who to contact, when, and why, with AI-drafted outreach ready to send. Primary value is delivered via push channels (email briefings and notifications) before users ever open an app.

Partners cross-reference 5+ disconnected systems — Salesforce, Excel, PowerPoint, email threads, and personal notes — to piece together a coherent picture of any client relationship. Every insight requires manual effort to surface and manual entry to record.

The firm's existing Salesforce platform delivers a tool-first, pull-based system that's misaligned with how Partners actually work. CRM adoption failure in professional services is widely cited at ~70%. The root cause: these systems are built for management visibility, not Partner productivity.

When a Partner is busy, relationships go cold silently. When a Partner leaves, institutional knowledge walks out the door. There is zero systematic relationship monitoring — no proactive nudges, no automated follow-through, no intelligence delivery.

Mapped the client development workflow end-to-end to identify high-value intervention points. Established six design principles: Zero-Effort Value (every interaction delivers more than it asks), Push Beats Pull (go to the Partner, don't wait for them), Insights Over Data (stories not tables), One Contact One Truth (canonical Contact 360), The System Thinks The Partner Decides (AI recommends, human controls), and Proper Visibility (firm access boundaries).

Developed the interaction framework with three trust-calibrated tiers: AI suggestions (user decides), AI-assisted actions (user confirms), and autonomous actions (user monitors) — each with clear visual language communicating the AI's confidence and role.

Prototyped and tested a 9-signal nudge engine (stale contact, missed meeting, news trigger, job change, sentiment shift, engagement drop, opportunity signal, relationship decay, custom) with configurable thresholds and consolidated card model. Designed multi-channel delivery: morning briefing email, real-time notifications, desktop dashboard, mobile conversational interface, and voice briefings.

Push-first intelligence delivery: AI-generated morning briefing email weaving nudges, meetings, news, and relationship context into a 2-minute daily read. Pre-drafted outreach ready to review and send. Value delivered before Partners open any app.

Contact 360 and Ask Anything: a 7-section AI dossier combining CRM data with live web intelligence — profile, relationship history, communication timeline, firm connections, news and signals, action items, and talking points. Natural language conversational AI with RAG-powered retrieval across all client data.

Cadence engine for automated follow-through: multi-step relationship sequences with context-aware email drafting, angle adjustment on no-response, and escalation paths. The system follows up so Partners don't have to.

Established the foundational AI experience strategy and interaction patterns guiding the integration of agentic AI into McKinsey's client development platform, serving 450+ client service teams.

Created a reusable framework for trust-calibrated AI interactions applicable across enterprise product surfaces. Built and validated the Activate prototype — a working platform demonstrating proactive relationship intelligence with dashboard, nudge engine, Contact 360 dossiers, conversational AI, and cadence-driven outreach.