← All speakers

Bio, Work & Ideas

Justin Joyce

Conference affiliation: Cloudflare · 2026

On this page

Justin Joyce brings experience in sales operations and machine learning to building agent-assisted systems for go-to-market teams. At the time of his World’s Fair 2026 appearance, he was principal sales operations and strategy manager at Cloudflare, working within revenue operations to support lead-generation teams and customer experience after the sale. His work combines reusable business knowledge, analytical tools, and reviewed reporting workflows to help employees understand performance and prepare for customer conversations.

From sales operations to machine learning

Joyce began in sales and sales operations before spending approximately seven years working in machine learning at Grainger. He pursued that transition to develop prescriptive analysis: using predictions to help a business decide what action to take next. By 2021, his authored-writing profile identified him as a senior data scientist at Grainger. His weighted ensemble forecasting tutorial implemented a method developed by other researchers that uses time-series characteristics to determine how to combine forecasts. He worked with commodity-price data and artificial training series, reported limited success on the small test set, and linked supporting material to his Learnings repository.

Closing the context and expert gaps

Returning to sales operations at Cloudflare gave Joyce a setting in which to combine those analytical methods with his commercial experience. He identifies two recurring obstacles to effective customer work. The context gap arises when employees must gather different information for successive prospecting, adoption, and customer-satisfaction conversations. The expert gap separates an experienced colleague’s judgment from that of someone still learning how to assess a customer’s needs or handle an objection. Better access to data helps with the first problem; reusable guidance about how to approach the situation helps with the second.

In his account of Cloudflare’s go-to-market workflows, Joyce explains how his team connects analytical assistance, proactively delivered insights, and self-service tools. These address different ways employees seek help: asking operations a question, receiving a regular performance summary, or preparing independently for a particular customer.

Role-specific skill files connect business definitions and data relationships to recurring analytical questions. For example, they include guidance for investigating changes in an opportunity’s closing date or amount. Employees who cannot write SQL can ask those questions without routing every request through a specialist. Operations staff also reuse the business knowledge and table information in the skills when building applications. Joyce’s aim is to reduce routine analytical work so the team can spend more time on strategy and enablement.

Delivering and checking weekly performance reports

His reporting approach starts with preparing the data. The team organizes it by time period, management group, region, and metric, engineering the relevant filters and aggregations before agents analyze it. Weekly summaries then explain progress against goals, trends, strong performance, and areas requiring attention. Joyce argues that this information should reach employees directly because dashboard adoption varies: some people regularly investigate reports, while others rarely open them.

The reporting workflow separates drafting, checking, and presentation. A first agent retrieves data through Model Context Protocol connections and drafts the analysis; a second checks factual consistency; a third uses example messages to balance the treatment of risks and opportunities. The team can inspect the inputs and responses for each model call. Joyce describes reviewing individual runs over two to three months while developing the workflow, making inspection and iteration part of the implementation rather than assuming that multiple agents guarantee accuracy.

Cloudflare OS and curated business skills

Cloudflare OS, the company’s internal agent workspace, gives go-to-market employees a place to use curated skills and connected data for forecast briefs, quarterly business-review decks, account planning, and renewal preparation. A renewal workflow, for example, can examine customer usage to inform an adoption or upsell conversation. The workspace runs on Cloudflare Workers and Durable Objects and combines skills, Model Context Protocol connections, and an AI gateway. Joyce describes these capabilities as team-built systems; the supplied evidence does not establish him as the workspace’s sole creator.

He places particular weight on skill curation. Operations and go-to-market colleagues submit and review skills centrally so that useful expert knowledge can be shared without an uncontrolled proliferation of overlapping instructions. Feedback from internal users tests whether the resulting workflows actually help them do their jobs and identifies what needs improvement.

Extending agents into customer meetings and system updates

In his 2026 presentation, deeper meeting integrations, quoting, approvals, and Salesforce updates remained under development. He envisaged attaching prepared artifacts to customer meetings and applying checked workflows to consequential system updates. As experimentation expands, he favors a more deliberate approach that keeps business systems aligned while preserving the flexibility employees need for ad hoc work.

1 conference talk

Key ideas

Scroll to read ↓

Justin Joyce explains how Cloudflare combines reusable business skills, automated weekly summaries, and Cloudflare OS to give go-to-market teams timely data and expert guidance—and why preparing the data and curating the skills matter as much as the agents.

  • Reusable skills connect business questions to data meaning and query logic, allowing users without SQL knowledge to obtain familiar analyses without waiting for a specialist.
    6:41 ↗
  • Weekly reporting combines prepared filters and aggregations with sequential drafting, factual review, and tone adjustment. Cloudflare inspected every run for two to three months with visibility into individual LLM inputs and responses.
    9:57 ↗
  • Pushed summaries provide shared performance context; Cloudflare OS supplies data and curated expertise for immediate customer tasks. Feedback and central skill review support both forms of delivery.
    8:57 ↗
  • Quoting, approvals, and Salesforce updates remain further work. Deeper integration makes alignment between skills, business systems, and sources of truth increasingly important.
    17:34 ↗

References