AI Engineer World's Fair 2026
How AI Agents Let GTM Teams Scale — Justin Joyce, Cloudflare
Slide chapters
17 slide and demo groups · Timestamps are approximate.
Read the talk
How AI Agents Let GTM Teams Scale
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.
From a talk by Justin Joyce
At a glance
Ideas worth remembering
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.
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.
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.
Quoting, approvals, and Salesforce updates remain further work. Deeper integration makes alignment between skills, business systems, and sources of truth increasingly important.
Repeated analysis, changing context, uneven expertise
Every week, operations teams rebuild analyses in Excel or Sheets. More projects mean more recurring work. Dashboards make information easier to distribute, but a standard view still leaves some teams waiting for answers to questions it does not cover. This is the practical problem behind Justin Joyce’s diagnosis that traditional go-to-market operations do not scale.
Joyce is Cloudflare’s principal sales operations and strategy manager, supporting lead-producing teams and the customer experience after a sale. After seven years working on machine learning, he returned to sales operations to combine business knowledge with prescriptive analysis: helping teams decide what to do next. His opening joke about the conference’s AGI pills gives way to a concrete concern—making the whole operation more efficient, from back-office analysis to customer conversations.
Customer-facing work adds two distinct gaps:
- Context gap: A representative moves from a prospect call to an existing-customer conversation and then to an adoption discussion. Each requires different information. The preparation is necessary; repeatedly gathering it between calls consumes time.
- Expert gap: An experienced representative and someone still ramping up may handle the same rejection, adoption problem, or customer satisfaction issue differently. Shared information must also help people assess the situation and choose a useful response.
These problems compound. An overloaded analytical team cannot answer every request, while representatives spend time assembling context and still lack the judgment embedded in an expert’s approach. Improving only the dashboard would leave much of that chain intact.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Three ways to make information available
The response has three pillars, each addressing a different moment in the work:
- Scale analysis: Help operations answer data questions and build applications using business context.
- Scale insight: Deliver recurring performance stories across management levels, with customer-level insight as a further ambition.
- Provide self-service: Give representatives data and expert guidance when a particular customer situation arises.
The first pillar changes what operations can spend its time on. Joyce frames the ambition as reducing two hours of work to five minutes. After building skills in January, he could ask questions directly of the data and receive answers while doing other tasks. Faster answers create room for the strategy part of sales operations: deciding what the business should do with the information.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Put business meaning and recurring questions into skills
Role-specific skill files connect business context to the underlying data. A technical user may already know SQL and data engineering; someone closer to sales may understand the business question without knowing how to write its query. The skill files carry the logic those users would otherwise have to obtain from a data specialist.
Opportunity changes provide a concrete example. The business wants to examine changes in close dates or opportunity amounts. Previously, a user who did not know SQL would send that request to someone who understood the data and could write a complex query. Through testing, the team added these familiar question types and their relevant logic to the skills. The user can then ask the question directly and receive an answer rather than wait in the specialist’s queue. The observable change is who can obtain the analysis; the enabling step is packaging the business-to-data knowledge beforehand.
Joyce estimates that the skills answer 80% or more of these questions, with the remaining 20% potentially involving more complex strategic analysis. That estimate concerns coverage of familiar business requests, rather than an accuracy score or a claim that every strategic question has become routine.
The same knowledge also supports application building. Business semantics and information about the data columns give the team reusable context for creating tools. Joyce reports that his team built multiple applications this way, reducing dependence on an IT queue. One investment in explaining the data can therefore support direct questions, new applications, and the customer-facing skills used later in Cloudflare OS.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Prepare the data, then bring the story to the team
The second pillar produces a weekly summary of business performance: progress toward goals, trends, standouts, and areas to watch. The presentation’s example uses synthetic data. The intended experience is a briefing people can receive and act on during their day, with reports and dashboards available when they want to investigate further. 8:57
Pushing the summary addresses uneven adoption of key performance indicators. Some people love dashboards; others will rarely open them. A common performance story can reach both groups, while the dashboard remains useful for drilling into details. Generating an answer and getting someone to encounter it are separate jobs.
Consistent reporting starts before the language model sees the data. The team organizes inputs by time, business slices such as manager and theater, and the metric being examined. It engineers the filtering rules and the aggregations the business wants upfront. This prevents each reporting run from having to reconstruct those same analytical choices.
The preparation uses both wide and long data layouts. Wide data spreads values across columns; long data organizes repeated observations into rows. Cloudflare’s trend information uses the long form, followed by preprocessing to highlight trends. The important decision is to give the agents a consistent business view suited to the question, with recurring calculations already prepared.
Joyce separately estimates that these prepared performance views handle 80% or more of requests because so many questions ask how teams are doing. Standardization makes common requests easier to serve, but it also selects the slices and calculations available in the summary. Deeper investigations can still go down to raw data through self-service.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Separate the draft, factual review, and tone
How does prepared data become a weekly message, and where do its checks happen? The reporting workflow runs three agents in sequence. Data retrieval and the first-pass draft involve calls to the team’s MCP connections. A reviewer then checks the factual basis of the draft. Finally, a tone agent uses a multi-shot prompt—examples of the desired message—to give risks and opportunities equal weight. 11:05
The sequence below makes the division visible: data preparation determines the business view, the drafter turns it into prose, the reviewer checks its factual basis, and the tone agent shapes the final emphasis. Factual review and editorial balance have their own stages, rather than depending entirely on the first draft.
Every run exposes the inputs and responses of individual LLM calls. The team inspected every run for two to three months to find what went wrong before settling on this architecture. That visibility makes failures easier to locate: a questionable result can be traced back through the calls that produced it. The workflow provides repeated checks and an operational basis for confidence, without establishing a numerical error rate or guaranteeing that review catches every mistake. Extending the summaries to more teams and individual customers remains a next step.
From prepared performance data to a balanced weekly summary
Read the diagram as text
- Prepared business data. Time, business slices, metrics, filters, aggregations, and trend preprocessing.
- Drafter. agents — Retrieves data through MCP connections and produces a first-pass draft.
- Reviewer. agents — Checks the factual basis of the reporting.
- Tone agent. agents — Uses a multi-shot prompt to give risks and opportunities equal weight.
- Weekly summary. Brings performance, goal pacing, trends, standouts, and watches to the business.
- Prepared business data → Drafter: Supplies consistent analytical inputs.
- Drafter → Reviewer: First-pass draft.
- Reviewer → Tone agent: Factual review precedes tone.
- Tone agent → Weekly summary: Crafts the message for distribution.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Cloudflare OS puts curated expertise within reach
Cloudflare OS is the internal agentic workspace that supplies the third pillar. Joyce describes each user getting compute and a persistent environment using Cloudflare Workers and Durable Objects. Representatives can enter the workspace when they need information for a customer task, rather than wait for the next scheduled summary. 12:03
The workspace combines expert skills, an MCP connection, and AI Gateway. Conversations retrieve data and apply curated guidance to a job. Access to customer information addresses the context gap; guidance about how to use that information addresses the expert gap. Both are needed to support a useful customer conversation.
The tasks span several kinds of preparation:
- Business and account planning: Forecast briefs, account plans, and general data queries.
- Customer-call artifacts: QBR decks and purchase decks describing what an onboarding customer bought.
- Renewal preparation: Examining customer usage to consider an upsell or help the customer adopt more of the product.
In the daily-work example, a user asks a question and the agent begins pulling information through MCP to assemble a prescriptive plan. Another example generates a custom slide deck for a customer call. These outputs put retrieved information into a form the representative can use—a plan or deck—rather than leaving the user to assemble the artifact from separate answers.
Self-service depends on a maintained skill repository. Skills are submitted through a central alias and curated by go-to-market and operations teams. Review helps prevent uncontrolled proliferation and keeps expert knowledge available for different customer situations. Users choose when to ask for help, while the organization reviews the guidance shaping that help.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Curation and feedback make the three layers useful
The three pillars serve complementary needs. Self-service supplies data and expertise for the situation at hand. Pushed insights establish standardized performance context. Greater analytical capacity lets operations answer questions and build applications that would otherwise remain unmet needs. The cost of an overloaded operations team reaches beyond its own backlog: customer-facing teams lose help they could have used.
Skill curation is the shared foundation. Business knowledge guides analysis and application building, while customer-task expertise guides work inside Cloudflare OS. The intended result is more predictable, consistent execution across users; curated skills do not by themselves guarantee deterministic model behavior.
Internal tools also need a product feedback loop. Teams must find out whether the outputs are useful, where people encounter problems, and what would make the work more efficient. Layering the delivery modes respects how users already work: some prefer asking operations a question, some benefit from a pushed briefing, and some want to retrieve information themselves.
Joyce reports 2× efficiency from the combined approach, alongside better access to information needed for the job. The talk does not define the efficiency measure or its baseline, so the figure is an operational result he reports rather than a reproducible benchmark.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
The next step moves from preparation toward system updates
The next integration goal removes another retrieval step. Instead of requiring a representative to pull a QBR deck or renewal-preparation artifact, the system could set up the meeting and attach the material there. The self-service portal would remain available for ad hoc needs. Capturing meeting notes across calls is another proposed extension. Both require additional system integration and security setup. 17:08
Quoting, approvals, and CRM updates are harder problems. Cloudflare uses Salesforce, and the team is building the connections that would let agentic systems update it. Joyce expects multi-agent workflows resembling the reporting pipeline to help check that the work is done correctly. This is work in progress: the talk does not establish a completed Salesforce write workflow or explain its approval controls.
That shift also changes the organizational problem. Joyce describes agent adoption as a Cambrian stage: excitement produces an explosion of skills and attempts to solve almost anything with AI. As those experiments become integrated workflows, each team needs a deliberate approach to using them so that the systems and their sources of truth stay aligned. The closing challenge is keeping shared business meaning intact as agents become involved in more of the work.
Suggest correction
This note stays in this page until you copy or download it. Nothing is submitted; reloading clears the draft.
Resources
Related talks
- Combine Skills and MCP to Close the Context Gap
A related next talk for the combination at the center of Cloudflare OS: data connectivity and reusable task expertise.
- Don't Build Agents, Build Skills Instead
Explains reusable procedural knowledge in more detail, including skill packaging and the distinction between connectivity and expertise.
Read the complete timestamped transcript
15 slide groups · 2 demo milestones. Extracted from the recording; expand builds to follow changes. Timestamps are approximate.
- 0:01
[music]
- 0:12
>> Well, thank you everyone for joining.
- 0:13
I'm hoping that you guys had a great
- 0:15
time so far at this conference and you
- 0:17
guys have a lot of takeaways back, you
- 0:19
know, to your company.
- 0:20
Um, I don't know if any of you guys saw
- 0:22
the uh AGI pills downstairs.
- 0:24
Yeah, well, I just took some. So, if I
- 0:26
say any phrases about that's the uh
- 0:30
the
- 0:30
the gun What's that What's that the
- 0:31
phrase? That's the uh burning gun or the
- 0:34
smoking gun or I start hallucinating in
- 0:36
general, well, please snap me back.
- 0:38
That's probably just the AGI pills.
- 0:41
All right. So, without further ado,
- 0:43
let's get started.
- 0:45
Uh my name is Justin Joyce. I'm a
- 0:46
principal sales operations and strategy
- 0:48
manager at Cloudflare.
- 0:50
Um and I work with the go-to-market team
How AI Agents Let GTM Teams Scale
Cloudflare’s Justin Joyce shares lessons from deploying AI for GTM teams. His work spans lead generation and post-sale customer experience.
Text in image
How AI Agents Let GTM Teams Scale Learnings from deploying AI for GTM teams Justin Joyce Principal Sales Operations and Strategy Manager Cloudflare SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 0:53
as part of the revenue operations
- 0:55
organization, specifically on the teams
- 0:58
that uh produce leads for the sales
- 1:00
teams, as well as the customer
- 1:02
experience team, which works on uh the
- 1:05
customer experience after the sale the
- 1:07
sales have been done.
- 1:09
And uh
- 1:11
just a little background about me.
Background
Justin Joyce moved from sales operations into machine learning for 7 years, then returned to GTM to apply those skills.
Text in image
01 // SPEAKER PROFILE Background Started in Sales Operations before transitioning into Machine Learning for the last 7 years. Returned to the GTM side to bring high-tech leverage to a traditionally less tech-heavy function. Edmond Fallot LA MOUTARDERIE SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 1:13
As Moda said, I started in sales
- 1:15
operations uh
- 1:17
and sales and then I moved to sales uh
- 1:20
to the machine learning side about the
- 1:21
last 7 years at Grainger. And I really
- 1:24
wanted to do that to be able to learn
- 1:25
how to
- 1:27
uh
- 1:27
have prescriptive analysis and
- 1:29
prescriptive uh prediction so I can help
- 1:31
the business better to make decisions
- 1:35
and to understand what's the next step
- 1:37
next best step.
- 1:39
So, uh
- 1:41
6 months ago, I had an opportunity to
- 1:43
move back into sales operations
- 1:46
uh because I really wanted to use all
- 1:47
the skills that I'd been learning from
- 1:50
machine learning as well as from sales
- 1:52
operations in general.
- 1:57
So, what's the general problem? The
- 1:59
general problem is that traditional
- 2:01
go-to-market does not scale.
- 2:04
There's a few um
- 2:06
facets to this.
- 2:07
The first facet is that usually teams on
- 2:10
the back office side, uh they're either
- 2:12
doing work in Excel or sheets at worst,
- 2:16
um building analysis each week, multiple
- 2:19
hours a week, and as they take on
- 2:21
multiple projects, it gets exponentially
- 2:23
long with how many of those analysis
- 2:25
that they're doing.
- 2:26
Uh at best, they are producing
- 2:28
dashboards,
- 2:29
um providing information to the uh
- 2:32
leadership and executive team, um which,
- 2:35
you know, meets the needs of most teams,
- 2:37
um but not all of them. And that needs
- 2:40
meets
- 2:41
that needs uh
- 2:43
means that in general, not all the
- 2:44
requirements of the go-to-market teams
- 2:46
are met. They're not able to really
- 2:48
provide all the information that the
- 2:50
teams need when um they need it.
- 2:55
The second problem, which is more on the
- 2:58
go-to-market side with the sales and uh
- 3:01
the sales teams and the other teams that
- 3:03
I mentioned that I support, is they have
- 3:06
two gaps. Essentially, the first gap is
- 3:08
the context gap, meaning when a
- 3:11
salesperson is
- 3:12
uh talking to a prospect um one call and
- 3:15
talking to a current customer in the
- 3:17
next call, or um talking at an adoption
- 3:20
conversation the call after that, they
- 3:22
have to constantly switch contexts and
- 3:24
they have to gather information for
- 3:26
those specific calls,
- 3:28
which is good. They really need to get
- 3:29
that information to have those calls and
- 3:32
understand how to approach the
- 3:33
situation, but they have to do all that
- 3:35
work in between.
- 3:37
Um the second one what is like what I
- 3:40
like to call is the expert gap, which is
- 3:42
the gap between how your expert
- 3:44
salesperson or expert go-to-market um
- 3:47
sales individual, how he would approach
- 3:50
a situation, how he would talk to a
The Problem: Traditional GTM Does Not Scale
Slow manual analysis leaves sales reps waiting for insights. Joyce also describes an expert gap: newer reps must gather information and learn to approach customer situations with the judgment of experienced sellers.
Text in image
The Problem: Traditional GTM Does Not Scale The Current State: Inefficiency > Manual processes create a bottleneck: analysts produce results slowly, delaying insights across the entire chain. Sales reps are left waiting, losing momentum. The Gaps > Constant context-switching between interactions with different customers. Teams struggle to bridge the "Expert Gap"—capturing how a senior sales executive would interpret a customer's unique situation. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 3:51
prospect, how he would uh work on
- 3:54
adoption call, how he would handle a
- 3:56
customer satisfaction issue, and
- 3:59
between also the
- 4:01
a new salesperson or someone that's just
- 4:03
ramping up. So, ideally like everyone
- 4:05
working at the same operational level,
- 4:07
so you have consistency in execution,
- 4:08
consistency in messaging of how to
- 4:11
assess a
- 4:13
a customer's problems and how your
- 4:15
product can help fit that portfolio. So,
- 4:18
with these two problems with manual work
- 4:21
as well as
- 4:22
uh
- 4:23
salespeople not having enough
- 4:24
information and having the gap of having
- 4:27
to get all the information they need
- 4:29
and gather that as well as not be able
- 4:31
to execute the same level, it really
- 4:33
creates an inefficiency in the
- 4:34
go-to-market organization.
- 4:37
And so, for the last 6 months or so
- 4:39
since I've
- 4:40
joined Cloudflare, I've been really
- 4:42
focusing on how can I make this
- 4:43
operation efficient from back to front
- 4:46
and there's a lot of great things that
- 4:48
we've been doing at our company in
- 4:49
general that's helped me enable that and
- 4:51
I really want to share some of those
The Problem: Traditional GTM Does Not Scale
Slow manual analysis leaves sales reps waiting for insights. Joyce also describes an expert gap: newer reps must gather information and learn to approach customer situations with the judgment of experienced sellers.
Text in image
The Problem: Traditional GTM Does Not Scale The Current State: Inefficiency > Manual processes create a bottleneck: analysts produce results slowly, delaying insights across the entire chain. Sales reps are left waiting, losing momentum. The Gaps > Constant context-switching between interactions with different customers. Teams struggle to bridge the "Expert Gap"—capturing how a senior sales executive would interpret a customer's unique situation. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 4:53
findings with you.
- 4:56
So, the framework that I'm proposing
- 4:59
here, which I think is is
- 5:02
it's going to really be effective in as
- 5:03
we flesh this out in the future is a
- 5:05
three-pillar approach.
- 5:07
The first pillar approach is how can we
- 5:09
scale analysis and the ability of
- 5:12
operations team to meet the data needs
- 5:15
of executives, leadership,
- 5:17
as well as be able to build applications
- 5:20
using business context. How can they
- 5:22
take things that would take 2 hours to
- 5:23
do down to 5 minutes? Um back in well,
- 5:27
not back in January of this year,
- 5:30
I after I joined I joined the company,
- 5:33
I had built these skills and I'd started
- 5:35
asking questions of the data directly
- 5:37
and I was able to get answers
- 5:38
immediately while doing other things and
- 5:40
I saw the huge power of how if we can
- 5:43
scale the analysis and the operations of
- 5:45
the teams, we can actually focus on the
- 5:47
second part of my job, which is
- 5:48
strategy.
- 5:50
The second pillar is how can is to scale
Three Aspect to Scaling with AI
Joyce’s three pillars scale analysis for complex requests, embed insight in daily workflows, and give sales teams self-service access to expertise for customer conversations.
Text in image
Three Aspect to Scaling with AI Core Philosophy: Automate the analysis, Augment the humans. Pillar 1: Scale Analysis Handle higher-order, ad-hoc, director and VP-level inquiries seamlessly. Build applications with business logic. Pillar 2: Scale Insight Embed deep analysis directly into weekly and daily operational workflows. Pillar 3: Self-Service Pull Model: Provide sales reps and leaders with AI leverage, ensuring augmentation rather than replacement. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 5:53
insight.
- 5:54
There's
- 5:55
story in the data, and how can we
- 5:56
provide that to the team? How can we
- 5:58
provide that team at, you know, the
- 6:00
weekly level, at the different levels of
- 6:02
management? How can we provide that
- 6:03
information, that insight, that story to
- 6:05
every customer that the sales teams are
- 6:08
talking to?
- 6:11
And the third one, which is arguably the
- 6:13
biggest one, is
- 6:15
uh to provide self-service capabilities
- 6:17
to the go-to-market team.
- 6:19
When these sales individuals, when
- 6:22
they're talking to a customer, how can
- 6:24
they get the expert-level information
- 6:26
that they need to interact with that
- 6:28
customer and to best assess, you know,
- 6:31
how they should approach the situation,
- 6:33
how to handle rejections, how to upsell
- 6:35
them, and how to handle customer
- 6:37
satisfaction issues.
- 6:39
Uh this is a huge part of what I've seen
- 6:41
we've done at Cloudflare, and I'll share
- 6:43
a little bit what that looks like.
- 6:47
So, as it relates to scaling the
- 6:48
analytical capability, the back-office
- 6:50
operations,
- 6:51
what we have done as we have built
- 6:53
role-specific skill files, which have
Scaling Analytical Capability Through AI
Cloudflare built role-specific skill files that tie business context to data for technical and non-technical users. The example defines Salesforce opportunity fields to help users ask questions of the data.
Text in image
Pillar 1: Scaling Analytical Capability Through AI THE BUILD > Develop specific AI skills for every team role Column Definitions Types below reflect field usage for the referenced columns. Tables and columns not listed have no upstream definition CSV; consult the Table Catalog above. salesforce.opportunity — referenced columns Column | Type | Description opportunity_id_18_char__c | STRING | The unique 18-character Salesforce ID of the opportunity record, useful for external integrations. id | STRING | The unique 15-character Salesforce ID of the opportunity record. accountid | STRING | The unique Salesforce ID of the customer account associated with the opportunity. name | STRING | The name or title of the Salesforce opportunity. type | STRING | The classification of the opportunity (e.g., 'New', 'Upsell - Insertion Order', 'Renew', 'Extension'). stagename | STRING | The current stage of the opportunity in the sales pipeline (e.g., 'Prospecting', 'Closed Won', 'Closed Lost'). closedate | DATE | The expected or actual close date of the opportunity, indicating when the deal is anticipated to be won or lost. createddate | TIMESTAMP | The timestamp when the opportunity record was created in Salesforce. createdbyid | STRING | The unique Salesforce ID of the user who created the opportunity record. [SCALE IMPACT] One analyst can now serve exponentially more requests by managing the agent rather than writing the code. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 6:56
the context of the
- 6:58
business information tying it to the
- 7:00
data. This is for both technical and
- 7:02
non-technical users. Technical users,
- 7:05
you could say the ones who are building
- 7:06
SQL and being able to data engineer a
- 7:09
lot of solutions. And then the
- 7:10
non-technical people would be more
- 7:13
individuals who are closer to the
- 7:15
business with the sales people who may
- 7:17
not know how to write SQL.
- 7:18
And so, we have skill files that they're
- 7:21
able to use to ask questions of the data
- 7:23
to get answers fairly quickly while
- 7:25
doing other tasks.
- 7:27
And
Applying AI Skills to Complex Questions
Skill files embed SQL for opportunity close-date and ACV changes, helping users who do not know SQL ask questions of their data. Joyce says the same files also help his team build applications.
Text in image
Pillar 1: Scaling Analytical Capability Through AI THE BUILD > Develop specific AI skills for every team role THE APPLICATION > Designed specifically for harder, complex questions and applications. ## Block: Field History — Close Date Changes SELECT opportunityid, CAST(createddate AS DATE) AS change_date, oldvalue AS old_close_date, newvalue AS new_close_date FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'CloseDate' ## Block: Field History — ACV Changes SELECT opportunityid, createddate AS change_timestamp, SAFE_CAST(oldvalue AS FLOAT64) AS old_acv, SAFE_CAST(newvalue AS FLOAT64) AS new_acv, SAFE_CAST(newvalue AS FLOAT64) - SAFE_CAST(oldvalue AS FLOAT64) AS acv_delta FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'ACV__c' [SCALE IMPACT] One analyst can now serve exponentially more requests by managing the agent rather than writing the code. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 7:29
one example here is in those skill
- 7:31
files, we also,
- 7:32
through testing, we've included the
- 7:35
types of questions that the business
- 7:36
would ask of the data. In this case,
- 7:39
looking at
- 7:40
closed date changes and opportunities,
- 7:41
as well as uh changes in the amount of
- 7:43
the opportunities, so that we can answer
- 7:45
essentially 80% or or more of the
- 7:48
questions, where the other 20% might be
- 7:50
more uh complex strategic questions.
Applying AI Skills to Complex Questions
Skill files embed SQL for opportunity close-date and ACV changes, helping users who do not know SQL ask questions of their data. Joyce says the same files also help his team build applications.
Text in image
Pillar 1: Scaling Analytical Capability Through AI THE BUILD > Develop specific AI skills for every team role THE APPLICATION > Designed specifically for harder, complex questions and applications. ## Block: Field History — Close Date Changes SELECT opportunityid, CAST(createddate AS DATE) AS change_date, oldvalue AS old_close_date, newvalue AS new_close_date FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'CloseDate' ## Block: Field History — ACV Changes SELECT opportunityid, createddate AS change_timestamp, SAFE_CAST(oldvalue AS FLOAT64) AS old_acv, SAFE_CAST(newvalue AS FLOAT64) AS new_acv, SAFE_CAST(newvalue AS FLOAT64) - SAFE_CAST(oldvalue AS FLOAT64) AS acv_delta FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'ACV__c' [SCALE IMPACT] One analyst can now serve exponentially more requests by managing the agent rather than writing the code. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 7:54
And so, overall, this allows the teams
- 7:56
to be able to embed all of the logic
- 7:59
into the skill files and get answers uh
- 8:02
fairly quickly. So, I've seen users who
- 8:04
do not know any SQL, and the essentially
- 8:06
their request in the past we bottleneck
- 8:08
to someone who knows data and can write
- 8:10
SQL for complex queries, be able to just
- 8:12
ask questions of the data and get
- 8:14
answers.
- 8:15
And this is very useful also for what I
- 8:18
show later on on the third pillar is for
- 8:21
building skills for these go-to-market
- 8:23
teams, so that they um can and ask
- 8:27
questions of their data and and get
- 8:29
answers. Also, uh in our team we've used
- 8:31
these same skill files to build multiple
- 8:33
applications. Uh when usually, you know,
- 8:36
that is uh done in IT and bottlenecked
- 8:40
in those areas, we're able to use the
- 8:41
semantic information about the business
- 8:44
um knowledge, as well as the columns
- 8:46
table to build these applications rather
- 8:48
quickly. So, this allows us to free up
- 8:50
our time, so that we can focus on
Applying AI Skills to Complex Questions
Skill files embed SQL for opportunity close-date and ACV changes, helping users who do not know SQL ask questions of their data. Joyce says the same files also help his team build applications.
Text in image
Pillar 1: Scaling Analytical Capability Through AI THE BUILD > Develop specific AI skills for every team role THE APPLICATION > Designed specifically for harder, complex questions and applications. ## Block: Field History — Close Date Changes SELECT opportunityid, CAST(createddate AS DATE) AS change_date, oldvalue AS old_close_date, newvalue AS new_close_date FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'CloseDate' ## Block: Field History — ACV Changes SELECT opportunityid, createddate AS change_timestamp, SAFE_CAST(oldvalue AS FLOAT64) AS old_acv, SAFE_CAST(newvalue AS FLOAT64) AS new_acv, SAFE_CAST(newvalue AS FLOAT64) - SAFE_CAST(oldvalue AS FLOAT64) AS acv_delta FROM `analytics.salesforce_raw.opportunityfieldhistory` WHERE field = 'ACV__c' [SCALE IMPACT] One analyst can now serve exponentially more requests by managing the agent rather than writing the code. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 8:52
strategy and enablement.
Automating Proactive Analysis
Pre-processing segments by KPI and detecting trends supports a weekly summary for the go-to-market team. The synthetic example highlights progress toward goals, standouts and declines.
Text in image
Pillar 2: Automating Proactive Analysis Data Strategy > Pre-process every segment by KPI. Advanced Detection > Continuous trend/outlier detection to highlight winners and anomalies. Automated Deliverables > Weekly call-outs: quarter progress, identified risks, and key metrics. Summary Zone Alpha continues to carry the load at 114% QO pace and 111% Bookings pace, though QO volume dipped to 221 from 248. Zone Beta delivered the week's standout Bookings number at $21.3M (up from $15.2M), pacing 116% on Bookings but only 97% on QO — they need 224 QOs/week over the final 3 weeks to close the gap. Zone Gamma bounced back to 99 QOs (from 71) but remains under pace at 90% QO and 96% Bookings. Zone Delta is the clearest concern at 61% QO pace and 59% Bookings pace. Win rate is a broad headwind — Global at 16% QTD against a 23% target, with both Alpha and Beta sliding week-over-week. Zone Gamma is the bright spot at 31% this week. Twenty-one manager activity alerts fired this week as we enter the stretch run. Notable Trends — REGION SCORECARD Records — strongest QTD starts Day 76 · vs. next-best prior quarter Zone Alpha · Core — $116.8M (+15% vs next-best Q1) Zone Alpha · East — $4.7M (+11% vs next-best Mar) Zone Alpha · South — $15.2M (+8% vs next-best Q4) Zone Beta · Central — $10.6M (+2% vs next-best Q4) 6 segments at record highs (top: Other $63.5M, Software $15.1M, Financial Services $9.4M) Standouts — accelerating momentum Multi-period growth signals Zone Beta · South COMM (QO ↑4q +880%) Zone Beta · East TERR (Bookings ↑4q +312% cooling) Tech & Services (Bookings ↑3q +176% cooling) Zone Alpha · NMD (Bookings ↑3q +68%) Zone Alpha · COMM (Bookings ↑4mo +47%) Public Sector (Bookings +12% QTD vs LQ) Watch — declines Reversals worth monitoring Zone Beta · South FLSM (QO -100%) Zone Alpha · TERR (QO -9%) Zone Gamma · MGD (Bookings -6% WoW) SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 2: AUTOMATION
- 8:57
All right, for the second pillar, uh
- 8:59
what I mentioned earlier, there's a
- 9:00
story in the data and they really
- 9:02
shouldn't have to search for it. Uh what
- 9:04
I'm showing you here is synthetic data
- 9:06
on the right. Uh we have a weekly
- 9:08
summary that goes out, which highlights
- 9:10
how the business is doing, how they're
- 9:12
pacing to their goals, and then
- 9:14
highlighting um trends, uh standouts, as
- 9:16
well as watches.
- 9:18
So,
- 9:20
this uh we provide this information to
- 9:22
the business, so they can, as you just
- 9:24
like you can open your phone now uh on
- 9:26
Gemini, if you have it, and you can see
- 9:27
your notes for the day or the things
- 9:28
that you need to do,
- 9:30
giving that same level of information to
- 9:32
the go-to-market team, so they can just
- 9:34
go along their day, and if they do need
- 9:36
to look at some of the reports or
- 9:37
dashboards, they can to drill in, um but
- 9:40
we bring the story to them.
- 9:42
And I'll pull this together why I think
- 9:44
this is really important. Um you know,
- 9:46
of course there's a place for dashboards
- 9:48
and standard information, but there's
- 9:50
different level of adoption of the KPI
Automating Proactive Analysis
Pre-processing segments by KPI and detecting trends supports a weekly summary for the go-to-market team. The synthetic example highlights progress toward goals, standouts and declines.
Text in image
Pillar 2: Automating Proactive Analysis Data Strategy > Pre-process every segment by KPI. Advanced Detection > Continuous trend/outlier detection to highlight winners and anomalies. Automated Deliverables > Weekly call-outs: quarter progress, identified risks, and key metrics. Summary Zone Alpha continues to carry the load at 114% QO pace and 111% Bookings pace, though QO volume dipped to 221 from 248. Zone Beta delivered the week's standout Bookings number at $21.3M (up from $15.2M), pacing 116% on Bookings but only 97% on QO — they need 224 QOs/week over the final 3 weeks to close the gap. Zone Gamma bounced back to 99 QOs (from 71) but remains under pace at 90% QO and 96% Bookings. Zone Delta is the clearest concern at 61% QO pace and 59% Bookings pace. Win rate is a broad headwind — Global at 16% QTD against a 23% target, with both Alpha and Beta sliding week-over-week. Zone Gamma is the bright spot at 31% this week. Twenty-one manager activity alerts fired this week as we enter the stretch run. Notable Trends — REGION SCORECARD Records — strongest QTD starts Day 76 · vs. next-best prior quarter Zone Alpha · Core — $116.8M (+15% vs next-best Q1) Zone Alpha · East — $4.7M (+11% vs next-best Mar) Zone Alpha · South — $15.2M (+8% vs next-best Q4) Zone Beta · Central — $10.6M (+2% vs next-best Q4) 6 segments at record highs (top: Other $63.5M, Software $15.1M, Financial Services $9.4M) Standouts — accelerating momentum Multi-period growth signals Zone Beta · South COMM (QO ↑4q +880%) Zone Beta · East TERR (Bookings ↑4q +312% cooling) Tech & Services (Bookings ↑3q +176% cooling) Zone Alpha · NMD (Bookings ↑3q +68%) Zone Alpha · COMM (Bookings ↑4mo +47%) Public Sector (Bookings +12% QTD vs LQ) Watch — declines Reversals worth monitoring Zone Beta · South FLSM (QO -100%) Zone Alpha · TERR (QO -9%) Zone Gamma · MGD (Bookings -6% WoW) SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 2: AUTOMATION
- 9:52
metrics at any given company. You're
- 9:54
going to have people who are love
- 9:56
dashboards, people are never going to
- 9:57
look at them. So, I think you really
- 9:59
need to have a way to scaffold that
- 10:00
across
- 10:01
um
- 10:02
the business. So, how do we do this
- 10:04
automated analysis? So, a big part of
- 10:06
this is simplifying the data so that the
- 10:08
AI agents can actually um analyze the
- 10:11
data in a very consistent and clean way.
- 10:14
Here what we do is we transform the data
- 10:16
by the
- 10:17
dimension of time, also slice of the
- 10:20
logical part of the business, which is
- 10:21
manager, theater, and finally the
- 10:24
metric. Here we have data uh that is
- 10:26
wide. You could also go um from wide to
- 10:28
long. Uh our trend information that I
- 10:30
showed you, um
- 10:32
uh that data is long, and then we do
- 10:34
some pre-processing on that data um to
- 10:37
then highlight trends. So, the the uh
- 10:40
embedding of the logic of how you would
- 10:41
filter this data to even analyze it, um
- 10:44
as well as the logical um aggregations
- 10:47
the business want to see is all
- 10:49
engineered up front. This
- 10:51
from my experience, this handles uh 80
- 10:54
or plus percent of the requests is just
- 10:56
getting information about the
- 10:58
performance of the teams and how they're
- 10:59
doing. You can always go down to the raw
- 11:01
data, but this last uh pillar, which
- 11:03
I'll go over in a minute, um allows them
A Robust Multi-Agent Workflow
A draft agent pulls data through MCPs, a reviewer checks it against the raw source, and a tone agent balances risks and opportunities. Each LLM call is observable, including its input and response.
Text in image
Infrastructure: A Robust Multi-Agent Workflow MCPs 1. Draft Agent • MCP Integration / Structured Analysis • Calls MCPs, pulls pre-transformed data • Processes it further, and drafts daily structures. 2. Reviewer Agent • Veracity Check / Anti-Hallucination • Cross-checks veracity against raw data source to aggressively catch and eliminate hallucinations. RAW DATA SOURCE 3. Tone Agent • Multi-shot Prompting / Risk Balancer • Perfectly balances positive highlights with potential risks using multi-shot methodology. Pros Cons Robust Outputs THE SCALE ANGLE Trust + Repeatability = The ability to run the system entirely unattended. Observability SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // AI IN GTM
- 11:05
to do that. To be able to orchestrate
- 11:07
this uh effectively and be able to rely
- 11:09
on it, we have a multi-agent workflow
- 11:11
where we first get the data, and then we
- 11:13
do a first pass draft on the data
- 11:16
calling our MCPs,
- 11:18
um and then this we have a second
- 11:19
reviewer agent who checks the veracity
- 11:21
of the data, and then we have a third um
- 11:24
agent, which is a tone agent, who using
- 11:26
a multi-shot prompt um is able to just
- 11:28
craft the message and highlight the
- 11:30
risks and opportunities
- 11:32
um equally. And with every run, we have
- 11:34
observability into each of the LLM
- 11:35
calls, so we can see what is passed and
- 11:38
what is the response that is going on
- 11:41
there.
- 11:41
And so, this architecture we tested for
- 11:45
about 2-3 months and, you know, looking
- 11:47
every single run to see what is going
- 11:49
wrong. And this is the the model that we
- 11:52
had set up that is really working for
- 11:54
us. And we really hope to expand this
- 11:56
beyond just what I've shown you for
- 11:58
multiple teams, but also down to the
- 12:00
customer level like I was just talking
Pillar 3: The Pull Model — Creating an Agentic Workspace
Cloudflare OS gives go-to-market teams a self-service workspace to request data when they need it. Joyce says each workspace gets its own compute and persistent environment.
Text in image
Pillar 3: The Pull Model — Creating an Agentic Workspace SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE Interface text: Cloudflare OS Send feedback Home Workspaces Workspaces Outputs Shared with me Scheduled Tasks Context Skills Search... New workspace FAVORITES 2 RECENT 7 SCHEDULED TASKS 1 Task: Sunday America... What would you like to do? Start a new workspace for your task Ask Seal to research, write, analyze, or create something... Claude 4.6 Sonnet
- 12:02
to you about.
- 12:03
The third part is the self-service
- 12:05
model. And what I'm showing you here is
- 12:08
our internal tool called Cloudflare OS,
- 12:11
which is an agentic workspace that is
- 12:13
running on Cloudflare where the
- 12:15
go-to-market team can come in here. It
- 12:17
spins up their own compute and their own
- 12:20
persistent environment using Cloudflare
- 12:22
workers as well as durable objects,
- 12:25
which is basically a storage
- 12:27
sort of like S3.
- 12:29
And so, the sales people can come in
- 12:30
here and get the data they need it when
- 12:33
they need it.
List of Use Cases
Sales teams use the workspace for forecast briefs, decks, account plans, CRM questions, and renewal prep. Joyce describes reviewing customer usage to identify upsell and adoption opportunities.
Text in image
List of Use Cases 01 Forecast Brief Automated generation of concise forecast summaries for sales leadership teams. 02 QBR Deck Generation Creating data-driven Quarterly Business Review presentations at scale. 03 Purchase Deck Generation Generating tailored pitch decks to facilitate the final purchasing decision process. 04 Account Plan Generate a strategic account plan slide deck for a named Cloudflare customer. 05 General CRM Queries Natural language interface for asking general questions on complex CRM data. 06 Renewal Prep Close plan for deal given CRM data. SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE
- 12:36
So, some use cases that these teams are
- 12:38
using it for is doing a forecast brief,
- 12:40
building QBR decks,
- 12:42
building a purchase deck on what the
- 12:44
customer that they're onboarding has
- 12:46
purchased, doing account planning,
- 12:47
general queries of the data,
- 12:50
as well as renewal preparation. How are
- 12:52
they going to look at what the customer
- 12:53
has used and
- 12:55
either upsell them or figure out how
- 12:57
they can get them adopting their product
- 12:59
more.
Pillar 3: The Setup
Cloudflare’s agent workspace connects AI inference through a gateway, loads curated skills and context, and accesses systems of record through an MCP portal. It also leads to deployed apps and agents.
Text in image
Pillar 3: The Setup Cloudflare’s agentic workspace CLOUDFLARE AI Inference Model optionality AI Gateway AI Agent Workspace Deployed or local harness Deployed Apps & Agents One-click environments Loads skills + context Content Library Curated context + skills Skills Content MCP Portal Systems of Record Transactional data Salesforce, Jira, Email, Radar Enriched by SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE
- 13:01
So, just a little bit more into that
- 13:03
Cloudflare S setup that I just showed
- 13:05
you. The AI agent workspace is where
- 13:07
that screen I was I was showing you. And
- 13:10
through the the three-part
- 13:12
piece of the skills in the lower left,
- 13:14
which is our like expert
- 13:16
level information,
- 13:18
as well as the MCP connection and the AI
- 13:21
gateway, they're able to have
- 13:23
conversations in this agentic workspace
- 13:25
to pull data they need and using
- 13:27
expert-level skills, which is curated,
- 13:30
um so that they're able to execute the
- 13:32
jobs that they need to do when they need
Skill Repository
Requests from an email alias and expert workflows pass through triage into centrally managed skills for the agent workspace. Teams review them to limit skill proliferation.
Text in image
Skill Repository Investing in a central context layer Email Alias Magic inbox for work requests Expert Tasks Documented expert workflows Triage & Audit AI agent operators curate the work Skill Repo Curated skill files + context, centrally managed AI Agent Workspace Surfaced and ready-to-use in Cloudflare OS Defining the jobs-to-be-done as centrally managed skill files: 1. Tap experts to document workflows and create skill files. 2. Create a “magic inbox” email alias where everyone else can send work. 3. Staff the inbox with AI agent operators who run it as a service. 4. Collect and curate the use cases, convert to skill files. 5. Everything is surfaced and ready-to-use in Cloudflare OS. CLOUDFLARE SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE
- 13:34
to do it.
- 13:36
And so, a little more information about
- 13:37
the skill repository, uh we have a
- 13:39
central alias where skills are presented
- 13:42
uh to uh the central team, curated by
- 13:45
the uh go-to-market team, as well as by
- 13:47
operations team, and they're reviewed,
- 13:49
so we can make sure that we're not
- 13:51
having a proliferation of skills, and we
- 13:53
have an expert-level knowledge skill at
- 13:55
every level, so that they can really get
- 13:57
all the information they need for how to
- 13:59
approach uh any customer situation.
Use Cases: Prescriptive Seller Plan
The agent pulls data through MCP and reads context files while preparing a seller’s daily plan. The document pane is still empty.
Text in image
Use Cases Every seller starts their day with a prescriptive plan Sellers Interface text: Cloudflare OS Send feedback [illegible] SELECT Name, Amount, ACV__c, CFI_ACV__c, CloseDate, Stage... [illegible] SELECT Name, Amount, StageName, CloseDate FROM Op... [illegible] SELECT Name, StageName, CloseDate FROM Opportunity WHERE... [illegible] SELECT CaseNumber, Subject, Priority, Status, CreatedDate... [illegible] SELECT CaseNumber, Subject, Status, ClosedDate, JIRA__c... tasks:update [illegible] status: completed [illegible] text = 'Target Corporation' AND type = 'page' [illegible] fs:list cloudflare/context fs:write memory/[illegible] fs:list cloudflare/context/facts fs:list cloudflare/context/proof fs:list cloudflare/context/mappings fs:list cloudflare/context/verticals fs:read cloudflare/context/facts/competitive.json fs:read cloudflare/context/proof/case-studies.json fs:read cloudflare/context/mappings/product-mapping.md I have enough to draft. Writing the prep doc now. tasks:update update research-target +2 more status: completed Thinking... Send a message... Claude 4.7 Opus Stop Send No documents yet Ask the agent to draft a document, or click + above to start. Documents appear here from this workspace. Files 68
14:03Demo milestone
- 14:04
And so, just a few more uh um images
- 14:06
here of uh them using it. Here, uh we
- 14:08
have them building a prescriptive plan
- 14:11
for their daily work. They're asking a
- 14:13
question, it's their um you can see the
- 14:15
agent is responding by looking into the
- 14:17
MCP and starting to pull the data
- 14:18
together.
Use Cases: Custom Slide Decks
For a customer call, a seller requests a deck for a retailer with physical locations and a large digital presence. The workspace shows a visual-style choice alongside a Cloudflare presentation cover.
Text in image
Use Cases Generate custom slide decks for each conversation Sellers Cloudflare Developer Platform for Retail I need a slide deck for a retailer who operates physical locations and has a large digital presence about how they should adopt Cloudflare's developer platform I need to ask you which visual style you'd like before I start creating slides. Questions 1 answered Which visual style would you like for this retail developer platform deck? Cloudflare Corporate Template (Recomme[illegible]) CLOUDFLARE Accelerate retail innovation with Cloudflare Unify your physical and digital experiences on one global network 1/8 SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE Interface text: Projects Retailer Deck for DevPlat Export Session Changes Ask anything... Build Claude Opus 4.6 Default Images Slides 8 slides
14:19Demo milestone
- 14:20
And related to the QBR deck, uh here's a
- 14:23
slide of generating a custom slide deck
- 14:26
for a customer call.
- 14:29
And so, this has really, I think,
- 14:30
unlocked the ability of the go-to-market
- 14:32
teams to be able to really have all
- 14:35
their information uh serviced to them.
Bringing it together
Self-service lets reps pull data, pushed insights reach daily workflows, and operations scales ad-hoc analysis. Joyce argues that missing layers leave information stranded or create an operations bottleneck.
Text in image
Bringing it together Three Aspects to Scaling with AI Automate the analysis. Augment the humans. FOUNDATION Self-Service Reps query data & use AI skills to serve customers directly Pull Model SALES REPS DISTRIBUTION Scale Insight Analysis embedded into daily workflows Push Model OPS → REPS ACCELERATION Scale Analysis Handle ad-hoc, leadership-level inquiries at speed Ops Efficiency OPERATIONS WITHOUT SELF-SERVICE Ops – overloaded Insight Self-Service × Every question funnels up → bottleneck at Ops WITHOUT INSIGHT PUSH Ops Insight × Self-Service Analysis exists but doesn’t reach reps → disparate information FULL COVERAGE Ops Insight Self-Service Every layer serves its purpose → compounding leverage
- 14:39
And so, bring it together with the three
- 14:40
pillars that I talked about.
- 14:42
Um
- 14:43
if you really don't have all these, I
- 14:44
think you have issues with serving the
- 14:46
go-to-market needs uh in terms of uh
- 14:48
using
- 14:49
optimizing the use of Agentyc uh
- 14:51
systems.
- 14:53
With self-service, you allow them to be
- 14:54
able to pull data when they need it for
- 14:56
whatever situation they need it with the
- 14:57
expert-level information.
- 14:59
The the second is by pushing the uh the
- 15:02
insights to the business, you're able to
- 15:04
surface the generalized standardized
- 15:06
information of how these teams are
- 15:08
doing, and also um
- 15:11
yeah, so so that there's no and also uh
- 15:13
so they're aligning with source of truth
- 15:14
on performance.
- 15:15
Um and then, the the third one is the
- 15:18
scaling of the analytical team for them
- 15:19
to be able to answer queries and build
- 15:21
applications for the teams, which really
- 15:23
unlocks a lot um because the opportunity
- 15:26
cost of that team being overloaded and
- 15:27
being able to help is that the meet the
- 15:30
needs of the go-to-market team um is not
- 15:32
met.
- 15:34
So, some findings um and the future.
- 15:37
So,
- 15:38
uh the first thing is skill curation is
- 15:40
the basis for all of this agentic
- 15:42
workforce. If you're able to embed the
- 15:44
knowledge of the business into the skill
- 15:47
files as well as the skills uh uh for
- 15:50
the an analyst to be able to build uh
- 15:53
answer questions or build applications
- 15:55
as well as the skills that I showed you
- 15:56
in the Cloudflare OS,
- 15:58
you're really able to uh give them the
- 16:01
ability to use the agentic systems in a
- 16:03
more predictable and deterministic way
- 16:05
so that they can execute um evenly
- 16:08
across the board.
- 16:09
The second thing is through this whole
- 16:11
process, the feedback loop is very
- 16:13
important. Just like uh a company would
- 16:16
try to sell a product externally and get
- 16:18
feedback, with these internal teams, uh
- 16:20
the feedback loop is very important to
- 16:22
be able to see is is what you're
- 16:24
building is it actually useful? What are
- 16:25
some issues that they're having? And how
- 16:27
can you make this uh work more
- 16:29
efficiently? And the third thing is the
- 16:31
layering of those uh three pillars that
- 16:34
I talked about. Being able to answer
- 16:36
questions where the team comes to you,
- 16:38
some of the go-to-market team, that's
- 16:40
how they like to interface with the
- 16:41
operations team is to be able to ask
- 16:42
questions. Um and then the pushing of
- 16:45
information and then
- 16:46
self-serviceability. Through that,
- 16:47
you're able to uh interweave all the
- 16:50
needs of the team to be able to be met
- 16:53
by this uh agentic run uh team. So,
- 16:56
through all these uh
- 16:58
different pillars that I talked about,
- 17:00
we've really been able to 2x our
- 17:01
efficiency and be able to serve the
- 17:03
teams as well as allowing them to be
- 17:05
able to get the information that they
- 17:07
need to do their job.
Findings and the future
Joyce lists skill curation, feedback and layering as findings. He looks ahead to embedding QBR and renewal artifacts in meetings, tackling quoting and CRM updates, and aligning sources of truth across teams.
Text in image
Findings and the future... FINDINGS 01. Skill curation is the basis 02. Feedback Loop 03. Layering THE FUTURE... Deeper Integration Harder Problems Standardization SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE
- 17:08
And some some things um that I see going
- 17:10
into the future. Number one is a deeper
- 17:12
integration with uh
- 17:14
our systems that we work in. So, for
- 17:16
example, those QBR decks and um
- 17:20
renewal call skills,
- 17:22
uh how can we set up meetings for the
- 17:25
go-to-market team and embed those
- 17:26
artifacts in those meetings so they
- 17:29
don't have to actually pull them. We can
- 17:31
allow them to that self-service portal
- 17:33
to be more ad hoc in what they need.
- 17:35
But, that requires some information or
- 17:37
some security setup and how can we do
- 17:39
that? And then also, another example is
- 17:42
getting meeting notes from those calls,
- 17:45
which you have to set up that across the
- 17:47
board. So, there's some system side
- 17:49
thing that we have to work on there. The
- 17:51
second thing is harder problems
Findings and the future
Joyce lists skill curation, feedback and layering as findings. He looks ahead to embedding QBR and renewal artifacts in meetings, tackling quoting and CRM updates, and aligning sources of truth across teams.
Text in image
Findings and the future... FINDINGS 01. Skill curation is the basis 02. Feedback Loop 03. Layering THE FUTURE... Deeper Integration Harder Problems Standardization SYSTEM.GTM.AI // JUSTIN JOYCE // 2026-06-20 // PILLAR 3 // AGENTIC WORKSPACE
- 17:54
around quoting and approvals and
- 17:57
updating the CRM itself. Uh we use
- 18:00
Salesforce and we're just in the midst
- 18:02
of
- 18:04
building the connections and the ability
- 18:05
for us to update Salesforce with these
- 18:07
Agenty systems. And I see that being um
- 18:11
set up in a way that I set up with that
- 18:13
automated analysis where you have
- 18:15
workflows uh to just make sure that
- 18:17
everything is getting um done right. And
- 18:20
the second thing is we're sort of
- 18:21
reached the Cambrian stage of using
- 18:23
Agenty systems, which means there's an
- 18:25
explosion of excitement and skills and
- 18:29
finding out ways to solve anything with
- 18:31
AI. But, I see as we get to this
- 18:35
uh fuller integration and
- 18:36
standardization, we're going to
- 18:38
uh want to come back and and and not
- 18:41
really limit, but just figure out a
- 18:43
really strategic approach for allowing
- 18:45
each team to use the Agenty system so
- 18:47
that the source of truth in all the
- 18:48
systems are aligning.
- 18:50
All right. Well, thank you for joining
- 18:52
this talk and I appreciate you,
- 18:54
um you know, coming here. Hope you have
- 18:56
a great conference.
- 18:57
>> [applause]
AI Engineer World's Fair
AI Engineer World's Fair logo in white on a black background.
Text in image
AI Engineer World’s Fair HTTPS://AI.ENGINEER
- 19:12
[music]