At a glance
AI is moving closer to becoming the main interface through which sellers access CRM, enablement content, and workflow support.
Coaching and role-play are starting to converge with broader revenue execution systems.
Enablement roles are being tied more explicitly to pipeline, win rates, retention, and expansion.
AI adoption is growing faster than the evidence around realised business value.
Governance, definitions, metadata, and content quality are becoming more important as AI moves deeper into the workflow.
The most interesting sales technology development this week wasn’t a single product launch.
It was the growing evidence that sellers may interact less directly with the systems underneath their work.
Salesforce is bringing CRM workflows into Claude. Enverus is already letting sellers access Highspot content through ChatGPT and Claude. Outreach is expanding the controls and business definitions its AI can work with. Nooks is bringing sales coaching into a broader revenue agent platform.
There were smaller signals worth watching too. Current enablement roles are being tied more explicitly to pipeline, win rates, retention, and expansion, while practitioner discussions continue to focus on measurement, repeatability, and what happens after training ends.
The technology is changing, but so is the work around it.
Salesforce is bringing CRM work directly into Claude
Salesforce and Anthropic announced Claudeforce on August 26, 2026.
Its first major component, Salesforce in Claude, brings 37 pre-built sales skills into the AI assistant. Sellers can use them for meeting preparation, deal-health reviews, pipeline reviews, CRM updates, and other tasks against Salesforce data.
The more consequential part sits underneath those use cases.
Salesforce is making data, permissions, business logic, workflows, and actions available through its AIforce framework and Model Context Protocol (MCP) infrastructure. The direction it describes is one where AI increasingly becomes the interface through which people interact with the underlying systems.
A recent Highspot discussion with Ryan Costello from Enverus gives us a useful practitioner example of what that can look like.
Enverus has connected Highspot to the AI environments its go-to-market teams already use. Sellers can combine call transcripts with enablement content, identify resources that fit a buyer’s situation, and create pitch links from within the AI conversation.
Costello reported that ChatGPT-side connector usage grew from a few hundred uses during a soft rollout to roughly 1,000 in the previous month following wider adoption. Those numbers come from a practitioner interview hosted by Highspot, so I wouldn’t treat them as independently verified performance evidence.
The implementation approach is more useful anyway. Enverus tested scenarios before rollout, spent roughly six weeks gathering early examples, and then incorporated the use cases that worked into broader sales training.
For enablement, this starts to change what in the flow of work means.
You may no longer need to persuade the seller to visit another destination. The enablement platform can remain the governed source while the content, guidance, and workflow appear somewhere else entirely.
This puts more pressure on the architecture which sits under it. Your methodology, messaging, content, definitions, permissions, and knowledge need to be structured well enough for an agent to find and use them correctly.
Make better enablement decisions. Execute faster. Deliver greater business impact.
Agentforce growth highlights the difference between AI adoption and AI value
Salesforce’s second-quarter fiscal 2027 results reported Agentforce annual recurring revenue above $1.5 billion, up more than 240% year on year.
Agentforce and Data 360 together reached nearly $3.9 billion in annual recurring revenue, while Salesforce reported 3.2 billion Agentic Work Units during the quarter.
Those are strong commercial and consumption signals.
A TD Cowen survey of Salesforce implementation partners, covered by Salesforce Ben, showed a more cautious picture. No surveyed partner identified Agentforce as a driver of recent bookings, while 33% reported strong buying or trial interest and 56% expected demand to develop as initiatives matured.
The survey sample size wasn’t disclosed in the coverage, so I would treat those findings as directional rather than definitive.
Still, the contrast surfaces an important measurement problem.
Usage tells you that people are using something. It doesn’t tell you whether the behaviour or business outcome you care about has changed.
For enablement teams rolling out AI, the measurement chain needs to continue beyond licences, interactions, or hours saved.
Did the seller make a better decision? Did the workflow actually become easier? Was capacity created? Where did that capacity go? Did execution improve?
The AI adoption conversation becomes much more useful when those questions sit alongside the usage numbers.
Nooks is pulling sales coaching into the revenue agent stack
Nooks acquired FullyRamped on August 25, 2026.
FullyRamped specialises in AI-powered voice role-play and sales coaching. Founder Aaron Marks is joining Nooks as the company moves beyond outbound prospecting towards what it describes as a broader Revenue Agent Platform.
That brings capabilities together that have traditionally sat in different systems.
A seller might practise in a role-play platform, prospect somewhere else, receive conversation feedback from another tool, and work opportunities in the CRM.
Nooks is starting to connect some of those jobs.
I think this becomes more interesting when you look beyond the acquisition itself. If practice, live call behaviour, coaching, prospecting, and workflow assistance begin to share the same context, the question for enablement changes.
It becomes less about whether the organisation has an AI role-play tool and more about how practice connects with what the seller actually does in the field.
The strongest system would be able to identify a behaviour in live work, create an opportunity to practise it, reinforce the standard, and then see whether execution changes.
There is still a lot to solve before that loop works reliably, but the architecture is starting to move in that direction.
The Seismic–Highspot merger creates an interesting technology choice
The Seismic–Highspot merger completed on August 18, 2026, just outside this edition’s main coverage window. Its commercial consequences became more visible this week.
The combined company now operates under Seismic and reports serving 2,500 customers and 3.5 million users. Both businesses have also been framing the category more broadly around go-to-market performance rather than traditional sales enablement alone.
Competitors are already responding.
On August 25, Allego published an article aimed at organisations approaching enablement renewals, using the merger as a reason to reconsider platform return on investment, adoption, migration risk, and consolidation.
That is competitor positioning, so it should be read in that context. But the response illustrates a bigger tension.
Large enablement platforms are consolidating at the same time that MCP and AI interfaces are making more composable technology environments possible.
One direction says: put more capabilities into a single platform.
The other says: specialist systems may increasingly work together behind an AI layer without the seller needing to navigate each one.
I wouldn’t assume one model wins. The more useful procurement question may become which architecture gives the organisation the right combination of governance, specialisation, integration, and seller experience.
Outreach is showing what AI governance looks like in practice
Outreach’s August product release contains a couple of developments that are easy to overlook because they sound operational.
Its AI Control Hub lets administrators decide which AI capabilities are enabled and whether future features should switch on automatically.
AI descriptions for custom fields go a level deeper. Administrators can explain in plain English what Salesforce and Outreach fields actually mean so that AI systems can reason over them more accurately.
Outreach has also expanded its MCP server with additional tools while retaining existing role-based permissions.
This is where AI readiness starts to become a knowledge and governance issue.
Someone needs to define what a field means. Someone needs to decide which source is trusted. Someone needs to determine what the agent can do, where human judgement is required, and what business rules it should follow.
Enablement won’t own all of that.
But when those definitions involve qualification, methodology, content, coaching standards, or expected seller behaviour, enablement needs to be involved in the operating decisions behind the AI, not simply the rollout afterwards.
Signals worth watching
Enablement roles are being tied more explicitly to commercial outcomes
A few current organisational and hiring examples stood out this week.
On August 25, 2026, Envision Pharma created an SVP of Commercial Growth role combining sales enablement, lead generation, marketing, client implementation, client experience, and customer success within one commercial structure.
Current Hootsuite enablement roles also reference measures such as quota attainment, pipeline, win rates, retention, and expansion. A current insightsoftware Director of Global Sales Enablement brief asks for someone who thinks more like a revenue leader, builds systems, drives adoption, and measures field outcomes.
Three examples don’t establish an organisational trend. But they are worth tracking because the public definition of enablement accountability appears to be stretching further into commercial performance and across the customer lifecycle.
Practitioner discussions keep coming back to measurement and repeatability
Current LinkedIn discussion captured in this week’s research repeatedly returned to two issues: how enablement proves impact and how organisations create repeatable behaviour.
Sales Enablement Collective has been discussing the shift in leadership questions from onboarding completion towards revenue impact. Another current discussion challenged the idea of asking sellers to copy how the best rep runs discovery, arguing instead for clearly defined and coachable standards.
These are individual discussions rather than market evidence. Still, they connect with the hiring signals above.
The expectation is moving towards demonstrating what changed in the field, not simply showing that an enablement activity happened.
Repetition is becoming part of the coaching conversation
A couple of smaller community signals pointed towards the same issue.
A current Reddit discussion argued that the gap between training and enablement is partly repetition: teaching a behaviour once is different from creating the conditions for it to survive. Another current discussion around AI role-play suggested repeated objection practice had been useful, although the sample is far too small to draw a broader conclusion.
A SalesHood session on August 27 with Neo4j also focused on manager reinforcement, coaching, behaviour change, executive sponsorship, and adoption.
Individually, these are weak signals. Together with the Nooks acquisition and wider movement towards continuous AI coaching, they make repeated practice and reinforcement an area worth watching.
What I’m paying attention to
The enablement destination may matter less
Salesforce in Claude and the Enverus example are coming from very different places, but they point in a similar direction.
Sellers can increasingly ask an AI what to send, how to prepare, what happened in a call, or what they should do next. The AI retrieves information and interacts with the systems underneath.
If that continues, adoption may need to be measured differently.
Direct platform visits could become less meaningful if governed enablement knowledge is being used constantly through another interface.
Enablement’s remit is getting wider while its proof standard is getting harder
The hiring signals, practitioner conversations, and technology developments connect here.
Enablement is increasingly being asked to operate around systems, workflow, management, adoption, and commercial outcomes. At the same time, leadership wants stronger evidence that the work changes something beyond participation.
This puts more emphasis on diagnosis and measurement.
If the team can’t define the behaviour, workflow, or performance condition it is trying to change, proving impact will remain difficult regardless of how sophisticated the technology becomes.
AI makes the operating layer underneath enablement more important
A lot of attention still goes to models, agents, and interfaces.
What stood out this week was how often the less visible layer appeared: permissions, metadata, business definitions, governed content, workflow rules, reinforcement, and measurement.
Those decisions determine what an AI system can reliably do.
The technology may be new. Much of the work required to make it useful is familiar.
Make better enablement decisions. Execute faster. Deliver greater business impact.
Worth reading / watching
Salesforce and Anthropic announce Claudeforce – Useful for understanding how Salesforce is thinking about AI as an interface into CRM data and workflows.
Embedding GTM Enablement Into Every AI Tool – The strongest practitioner example this week because it covers the rollout, adoption, training, content retrieval, and workflow behind the technology.
Why Nooks Acquired FullyRamped – Worth reading for the argument that coaching and practice may become part of broader revenue execution infrastructure.
Outreach August 2026 Release Notes – A useful look at how permissions, field definitions, AI controls, and MCP are becoming part of day-to-day revenue operations.
Salesforce Partners Are Not Seeing Agentforce ROI – A useful counterpoint to Salesforce’s strong growth figures, with the caveat that the underlying partner survey has limited disclosed methodology.
The bigger picture
The clearest shift this week is that AI is moving closer to the point where revenue work actually happens.
There is a great opportunity to remove friction. Sellers can access content without searching for it, prepare for meetings without manually stitching information together, and receive support without moving between several applications.
But the smaller signals are important too.
Enablement roles are being described in broader commercial terms. Practitioner discussions are putting more pressure on measurement and repeatability. Coaching is moving towards continuous reinforcement rather than isolated events.
All of this raises the bar on the system that sit under the work.
If content is poorly governed, definitions vary across teams, managers reinforce different standards, or nobody can explain what good execution looks like, AI won’t remove those inconsistencies. It may simply make them easier to reproduce.
For enablement teams, that makes the operating model around the technology every bit as important as the technology itself.