HP's MySales360 shows how a conversational agentic environment can eliminate fragmented systems, saving frontline staff time and improving the consistency of outputs across complex workflows.

Fragmented Systems and Lost Productivity

Across large organizations, frontline staff often lose significant time navigating fragmented systems to find accurate, up-to-date information. In HP’s sales organization, sellers had to search multiple tools to answer basic questions (product specs, competitive comparisons, availability/backlog). This slowed proposals and increased the risk of using outdated or incorrect data. The resulting inefficiency reduced time spent on higher-value work and created inconsistencies in downstream outputs (e.g., proposals drafted with the wrong information). The challenge is common in government as well, where information and services are often siloed across departments and systems, creating delays, duplication, and avoidable administrative burden for staff. 

How MySales360 Works

MySales360 is a persona-based, conversational agentic environment designed around the end-to-end workflow of a seller, consolidating tasks that previously required navigating many disconnected systems. Through a single front end, users can ask questions, request analyses, or initiate actions. An agent orchestrator interprets each request, determines what information or outputs are needed, and activates specialized back-end agents (e.g., product knowledge, competitive intelligence, inventory and backlog systems). These agents can run in parallel, with the orchestrator synthesizing results into a single, coherent response. Critically, the system routes requests only to authoritative data sources, improving accuracy and consistency across downstream outputs such as proposals. The architecture is designed to scale by adding new agents over time. 

Measuring Impact

Trust is maintained through extensive pre-release testing, continuous monitoring for accuracy and drift, and structured user feedback incorporated into rapid release cycles. 

HP tracks time saved explicitly (in hours) and converts this into an estimate of margin/revenue uplift from shifting seller time from searching to customer engagement. They also track accuracy with multiple sub-metrics, and user adoption/continued usage as a proxy for user satisfaction. Key challenges included data readiness (identifying authoritative sources and cleaning and preparing data, which consistently extended project timelines), and change management, marked by initial scepticism about accuracy and value, as well as natural resistance to automation.

Key Lessons

A key lesson was to start with an end-to-end process mindset: solving one pain point in a multi-step workflow can limit uptake; greater adoption comes from redesigning and automating the full process, supported by continuous user-driven iteration. In government, a similar agentic pattern can support procurement and acquisition teams by orchestrating data across siloed systems to draft RFPs, compare bids, and synthesize information for decision-making, reducing manual workload while preserving human oversight over final approvals.