Skip to main content

Copilotforce: Salesforce Is Now a Plugin Inside Microsoft Copilot [2026]

· 10 min read
Sunder Iyer
Founder, marketbetter.ai

Abstract hero: two large orbital systems converging around a shared interface plane

November 17 was the busiest single day the sales interface war has ever had. In the span of one Monday: OpenAI launched ChatGPT for Sales with 30+ partner plugins. Microsoft GA'd Dynamics 365 in Copilot with 44 prebuilt sales skills. And — the headline nobody had on their bingo card — Microsoft and Salesforce jointly announced Copilotforce, putting Salesforce itself inside Microsoft Copilot as a plugin.

Two companies that spent 2024 and 2025 building rival AI sales agents and openly poaching each other's CRM customers just agreed on one thing: sellers live in chat surfaces now, and the CRM is becoming something those surfaces call into.

Here's what actually shipped at Ignite, what status each piece is really in, what both vendors left out of the announcement, and what it means if you run a sales team on either stack.

What Copilotforce actually is (and isn't)​

First, a definition worth being precise about, because the name suggests a product and the announcements describe something else. Both companies call Copilotforce a partnership, not a product. Salesforce's own framing: "a partnership bringing Salesforce to Microsoft's expansive enterprise productivity ecosystem." Under that umbrella sit three concrete pieces, each at a different maturity stage:

PieceWhat it doesStatus (Nov 17, 2026)
Salesforce in Microsoft CopilotPlugin with 38+ prebuilt sales skills; read CRM context, take supported actions from CopilotPublic preview (beta), live now
Data 360 ↔ Microsoft Fabric Zero CopyQuery Fabric OneLake data inside Data 360 (Data Federation); share Data 360 context into OneLake (Data Sharing)Federation: GA. Sharing: public preview
Sales Cloud in TeamsSalesforce customer context inside Teams conversationsPilot only; open beta "later this year"

Diagram: the three pieces of Copilotforce and their launch stages — plugin in public preview, data integration GA/preview, Teams app pilot only

So the shipping-today reality is narrower than the headlines: one plugin in beta, one data integration direction in GA, and a Teams app that most customers can't touch yet. The pattern matches every major launch we've tracked this fall — 6sense's Breakthrough announcements had the same shape, where the only genuinely GA items were the least flashy ones.

How the plugin works​

The mechanics, assembled from both companies' launch material:

  • It runs on AIforce, which Salesforce describes as "the live interface layer that brings Salesforce capabilities into external AI environments." If that sounds familiar, it's the same layer behind Claudeforce, the Salesforce-in-Claude integration from September. Salesforce has now shipped its CRM into Claude, ChatGPT (via the Nov 17 OpenAI partner list), and Microsoft Copilot in under three months. That's not three integrations; that's a strategy — be present in every chat surface rather than betting on owning the one that wins.
  • Microsoft's Work IQ supplies the other half of the context. Salesforce insights get synthesized with "relevant relationships and content from Teams meetings, Outlook emails, and shared documents across SharePoint and OneDrive." This is the part Microsoft brings that OpenAI structurally can't: the meeting, the email thread, and the deck already live in Microsoft's graph.
  • Access runs over authenticated Salesforce MCP, scoped to each seller's existing Salesforce permissions. The Model Context Protocol keeps winning as the plumbing standard — the same rail Salesloft and Clari exposed their data on, and the same protocol we built MarketBetter's own agent access on.
  • Write-back is confirm-gated. "Supported writeback actions to the CRM require explicit user confirmation and respect existing validation rules." Same design OpenAI chose for its Sales plugin: propose, don't silently write. Nobody wants to be the vendor whose agent corrupted a pipeline report.

The named skills give a feel for the altitude: Pipeline Review, Close Planning, Deal Review, Customer Health, Renewal Radar, Call Follow-Up, plus workflows for customer meeting preparation and deal-health analysis. The full list of 38+ is not published anywhere in the launch material — you get the count, not the catalog.

Distribution is through Salesforce's AgentExchange and the new Microsoft Copilot plugin directory, which Ignite also introduced, with a "unified plugin registry" for admin control.

The frenemy mechanics: Microsoft announced both sides of this trade​

Here's the part of the keynote that deserves more attention than it got. The same Ignite announcements that welcomed Salesforce into Copilot also:

  1. GA'd Dynamics 365 in Copilot — Microsoft's own CRM as a Copilot plugin, with 44 sales skills (versus Salesforce's 38+), covering account plans, meeting prep, pipeline prioritization, decision-maker mapping, and deck building. Microsoft's framing is blunt: "Copilot is the new front door to Dynamics 365."
  2. GA'd Dynamics 365 Activate for CRM — Microsoft's tool for migrating customers off Salesforce and onto Dynamics 365. A Microsoft partner claims one enterprise customer is moving 20,000 Salesforce seats to Microsoft in about five months. This shipped to general availability the same day Microsoft put Salesforce's logo in its keynote.

Read those together and the strategic picture is clean: Microsoft wins either way. If you stay on Salesforce, your sellers now live in Copilot and Microsoft owns the interface, the meeting data, and (eventually) the consumption billing. If you leave Salesforce, Activate moves you to Dynamics 365 — whose sales agents and usage-based Copilot Credit billing we broke down when the Sales Development Agent was announced. Every road runs through the same front door.

Salesforce's calculus is the mirror image. Agentforce's own sales agents are still maturing — Hunter, the outbound agent, remains in pilot with GA promised for this month — and Salesforce has watched sellers drift into chat surfaces it doesn't own. Being the trusted data and action layer inside all of them, powered by AIforce, beats being the tab nobody opens. Miguel Milano, Salesforce's President and COO, said the quiet part at launch: Copilot is "the largest surface in enterprise software today," and Salesforce wants to be embedded in it.

Accenture is the launch proof point for the whole arrangement — design partner and self-described "customer zero," running Copilot across more than 740,000 employees with 55,000 Salesforce users. Worth noting that's about the least representative customer imaginable: a global SI that is simultaneously Partner of the Year for both vendors. Your mileage as a 200-person B2B company will vary.

What the announcement doesn't say​

Our standing practice with launch coverage is to list what's conspicuously absent, because the omissions are usually where the budget surprises live:

QuestionWhat the announcements say
PricingNothing. Only a disclaimer that "pricing and packaging are subject to change"
Required licensesNot stated — neither which Microsoft Copilot license nor which Salesforce edition
GA date for the Copilot pluginNone given
Full 38+ skill listNot published
Regions"Regional availability and customer access are governed by applicable customer agreements"
Sales Cloud in Teams open beta date"Later this year"

The licensing silence matters most. Microsoft said at Ignite that its model is a two-track system — a per-user Copilot subscription plus "a usage-based consumption model" for agent work — and that it will "continue to move more and more capability into the subscription license" over time. Which track Salesforce-in-Copilot queries land on, and whether Data 360 or Fabric entitlements are prerequisites for the Zero Copy pieces, is unstated. If you've followed this year's drift toward AI meters and consumption billing across the sales stack, you already know the default assumption to make: the preview is free, the meter arrives later. Microsoft's own Sales Development Agent followed exactly that arc — free public preview, then usage billing on Copilot Credits starting December 1.

One more flag for the skeptics' file: Microsoft's launch post cites "a greater than 25% reduction in seller effort" and "9% improvement in deal closure velocity" for CRM-in-Copilot workflows. The footnote says those numbers come from internal telemetry in Microsoft's own controlled customer-zero environment, offered with "no warranties." Treat them as marketing, not benchmarks.

What this means for your sales team​

If you're a Salesforce shop on Microsoft 365, Copilotforce is genuinely useful news, and the beta is a reasonable thing to pilot — the meeting-prep and pipeline-review skills attack real seller time sinks, and confirm-gated write-back limits the blast radius. Just go in with eyes open on three things: you're wiring your CRM into an interface whose pricing for this workload doesn't exist yet, the full skill list is unpublished so you can't scope what reps will actually do in it, and the Teams piece most sellers will want first is the piece furthest from shipping.

If you're evaluating the broader pattern, the week of November 17 settled an argument. OpenAI, Microsoft, and Salesforce all now agree the chat surface is the seller's front door, and all three shipped plugin architectures to claim it. The CRM — the system sellers spent two decades being forced to live in — is officially becoming a context provider behind an interface someone else owns. The AI assistants doing that retrieval are already a measurable share of how buyers research too; the interface shift is happening on both sides of the deal.

But notice what none of this does. Every skill named across both launches — pipeline review, deal health, meeting prep, follow-ups, renewal radar — operates on deals and contacts you already have. Same gap we flagged in OpenAI's partner list two days ago: no signal vendors, no visitor identification, no intent data, no outbound execution engine. Copilotforce tells your sellers everything about the accounts in the CRM. It does not put a single new account into the CRM, and it doesn't watch your website, your market, or your buyers' job changes to tell an SDR who to touch today and why.

The interface layer is consolidating fast. The engine layer — the thing that finds in-market buyers and turns signals into executed outreach — is still the part you have to bring yourself. That's the layer MarketBetter lives in: visitor identification, buying signals, and playbooks that tell your team who to act on and what to do, whatever chat window ends up on top of your stack.

See what an engine behind the interface looks like: Book a demo →


Sources: Salesforce and Microsoft Copilotforce announcements (Nov 17, 2026), Microsoft Dynamics 365 and Copilot Ignite 2026 blog posts, Accenture newsroom, CX Today. All availability statuses as stated by the vendors on launch day; pricing unannounced at publication.

ChatGPT for Sales: OpenAI Just Made Chat the Front End of Your GTM Stack [2026]

· 8 min read
Sunder Iyer
Founder, marketbetter.ai

Abstract hero: a central chat surface connected to orbiting tool nodes

On November 17, OpenAI launched ChatGPT for Sales: an OpenAI-built Sales plugin plus more than 30 new and expanded partner plugins that pull your CRM records, call notes, and sales content directly into ChatGPT. Salesforce, HubSpot, Apollo, Clay, Gong, DocuSign — the core of the mid-market sales stack is now installable inside the chat window your reps already have open.

This is the third sales-adjacent release from OpenAI in seven weeks. Dots brought scheduled agents at the end of September. Slides and Space gave sales teams a deck builder in October. Now the Sales plugin ties those pieces to the systems where deals actually live. The pattern is hard to miss: OpenAI isn't building an AI SDR. It's building the interface layer — and betting that the rest of your stack becomes plumbing behind it.

Here's what shipped, what the plugin list tells you about where the market is going, and where the gaps are for an SDR team evaluating it this quarter.

What the Sales plugin actually does​

Per OpenAI's launch material and early coverage, the Sales plugin handles five jobs:

  1. Account research and prioritization — builds prioritized account lists from your connected CRM data plus prospect research.
  2. Meeting prep — assembles briefs and tailored pitches from account history before a call.
  3. Follow-ups — drafts the post-call email, records next steps.
  4. CRM updates — proposes record changes for you to review, rather than writing silently.
  5. Pipeline reviews — flags at-risk deals and suggests next actions.

The more interesting part is what happens when you combine it with Dots, OpenAI's recurring-agent feature. OpenAI's own example: a dot that reviews your pipeline every morning with a connected deal tool and surfaces what changed overnight. That's no longer "ask ChatGPT a question" — that's a standing agent running against live revenue data on a schedule. OpenAI's caveat in its own launch copy is worth repeating verbatim: "consequential work should always be reviewed."

Note the asymmetry, though. Access to the Sales plugin does not automatically grant access to Dots — Dots has its own eligibility and availability requirements. The demo where everything composes into an autonomous morning workflow assumes you've cleared both gates.

The partner list is a map of the 2026 stack​

More than 30 partner plugins launched or expanded alongside the Sales plugin. The names reported across launch coverage, grouped the way OpenAI groups them:

CategoryPartners
CRM and deal dataSalesforce, HubSpot, Attio, Close, monday CRM
ProspectingApollo, Clay, Common Room
Call notes and meetingsGong, Otter.ai, Granola
Quotes and agreementsDocuSign, PandaDoc, Ironclad, ServiceNow CPQ
Content and enablementSeismic, Guru, Confluence, Gamma, Figma, Canva

Three things jump out.

First, the prospecting rails are Apollo and Clay — the two vendors that spent 2026 making themselves machine-callable. Apollo shipped a CLI and MCP server for headless GTM in October. Clay announced Workflows and an API at Sculpt the same month. Both made the same bet: when an AI assistant becomes the interface, the data vendor that survives is the one that's easiest for the agent to call. This plugin list is that bet paying out. The vendors who kept their data locked behind their own UI aren't on it.

Second, the CRM list includes Attio — and Attio is the write-back demo. OpenAI's demo of saving next steps to a CRM ran through Attio, which rebuilt its architecture around agent access precisely for this moment. Meanwhile write capability is explicitly not uniform across CRM plugins — what ChatGPT can change in your CRM depends on each plugin's supported actions and authorization. If you run Salesforce, test the write path before you promise your team "ChatGPT updates the CRM now."

Third, Gong is on the list as a data source. Gong spent this year building its own agent surface; joining ChatGPT's plugin roster means also accepting a future where reps ask ChatGPT about their calls instead of opening Gong. Every vendor on this list made that trade: distribution inside ChatGPT in exchange for giving up the interface. monday.com made the same calculation from the other direction — building voice agents of its own while also plugging its CRM into ChatGPT.

Availability and cost: free plugin, paid everything else​

The plan mechanics, as reported at launch:

  • The Sales plugin itself is available globally across ChatGPT plans — including Free, per OpenAI's help documentation.
  • Admin controls live on Business and Enterprise. Admins can enable the Sales and partner plugins, let members install them, or install them by default from the workspace Admin Console. Some partner plugins require admin approval.
  • Partner plugin availability varies by plan, region, and workspace settings.
  • No separate price was announced for ChatGPT for Sales.

That last point deserves scrutiny, because "free" here is doing careful work. Connecting the Salesforce or HubSpot plugin does not grant you Salesforce or HubSpot — every partner subscription is billed separately, by the partner. The Sales plugin is a free aggregation layer over software you're already paying for. This is the same move we flagged in our October look at Slides and Space: the headline feature is free, and the capability that makes it useful at team scale sits behind a paid tier or a partner invoice. After re-verifying 38 sales-tech pricing pages this fall, our default assumption is that free surfaces exist to meter something later.

What's missing from the list​

Read the partner table again for what's not there.

No intent or signal vendors. No visitor identification, no intent data, no website-session signals. The absence matters because it defines what ChatGPT for Sales is: a layer over data you already have. It can summarize the account history in your CRM brilliantly. It cannot tell you that a target account's VP of Sales visited your pricing page twice yesterday, because nothing in the plugin list captures that signal. 6sense is attacking the same chat surfaces from the opposite direction — embedding its buyer intelligence into assistants — which tells you the signal layer fight is happening outside OpenAI's plugin program for now.

No outbound execution. Nothing in the list sends sequences, places calls, or manages deliverability. ChatGPT drafts the follow-up; a human still sends it through something else. OpenAI stopped cleanly at the research-prep-draft-review loop, which is the defensible place to stop — it's also the part of the job that was already most automatable. The expensive problems in the AI SDR category — sender reputation, sequencing logic, knowing when to reach out — are untouched.

No measured outcomes. Launch coverage noted the announcement reports no independently measured gains in revenue or sales-cycle speed. Standard for a day-one launch, but worth remembering when the LinkedIn posts start claiming otherwise.

What to actually do with this​

If you run an SDR or sales team, the evaluation is cheap and worth doing this week:

  1. Check your admin console. If you're on ChatGPT Business or Enterprise, see which partner plugins are available in your region and workspace before planning anything.
  2. Pilot one read-only workflow first. Meeting prep from CRM history is the lowest-risk, highest-frequency win. Measure prep time per call for two weeks against your current baseline.
  3. Test the CRM write path with one rep, not the team. Write capability differs per plugin. Verify what "proposes changes for review" looks like in your actual CRM before announcing it.
  4. Don't confuse the interface with the engine. ChatGPT for Sales makes your existing data conversational. It doesn't create signal (who to contact, and why now) and it doesn't execute outreach. Teams that treat it as an AI SDR will be disappointed; teams that treat it as a faster window into their stack will probably keep it.

That last distinction is the one we build MarketBetter around: a dashboard — or now, a chat window — that shows you who is only half the job. The other half is a system that turns the signal into the specific next action and runs it. If your team's bottleneck is knowing which accounts are warm right now and what to do about them before the window closes, that's the part no plugin list solves yet — and the part worth seeing in a demo.

The bigger picture​

Seven weeks, three releases: scheduled agents, deck generation, and now the stack plugged into chat. OpenAI is assembling a sales workspace piece by piece, for free, on top of software everyone already pays for. The vendors joined because distribution inside ChatGPT beats being invisible to it. The open question for 2027 is whether the plugin partners are gaining a channel or training their replacement — and the vendors that made themselves agent-callable this year clearly decided they'd rather find out from the inside.

We'll keep tracking what ships next — Dots eligibility expanding, write-back deepening, and whether the signal layer ever makes it onto the partner list.

The 2026 Sales-Tech Repricing: 28 of 38 Pricing Pages Changed [2026]

· 12 min read
Sunder Iyer
Founder, marketbetter.ai

Chart-themed hero: sales software pricing shifting in 2026

Over the past eight weeks we re-verified the pricing of 38 B2B sales tools — CRMs, prospecting databases, sequencers, visitor identification platforms, and AI agent add-ons — directly against their own pricing pages, checkout flows, and official migration docs. Not against aggregator listicles, which we found stale more often than not.

28 of the 38 are selling something materially different than they were a year ago. Plans renamed. Free tiers deleted. Per-seat pricing swapped for pooled credits. Usage meters bolted onto AI features. In several cases the vendor's own meta tags still advertise a free plan that no longer exists.

Last year we published a study of pricing transparency across 70+ sales tools — who publishes pricing and who hides it. This is the sequel question: for the vendors who do publish pricing, how fast does it move? The answer: faster than your renewal cycle.

Here's the full change log, the five patterns behind it, and what to do about it if you're buying.

How we did this (and what counts as "changed")​

  • First-party sources only. Live pricing pages, embedded checkout JSON, official plan-migration help docs, vendor emails, earnings calls, and admin message-center notices. Where a pricing page blocks scrapers, we pulled the structured pricing data embedded in the page itself. Where a vendor's page conflicted with its own FAQ, we flagged it.
  • Each fact carries a verification date (late September through November 16, 2026). Pricing this volatile goes stale; dated claims are the only honest kind.
  • "Changed" means structural: a plan added, removed, renamed, or repriced; a free tier altered; a billing model swapped. Copy tweaks don't count.
  • 38 tools came from re-verifying three of our highest-traffic roundups — free lead generation tools, pipeline management CRMs, and website visitor tracking — plus individual pricing rebuilds we shipped along the way.

Bar chart: what changed across 38 sales tools in 2026 — lineups rebuilt, free tiers cut, prices raised, AI meters added

One tool can appear in more than one bar — Pipedrive, for instance, both renamed its lineup and raised prices inside the rename.

Pattern 1: The free-tier retreat​

The clearest trend in the dataset: free tiers are being deleted or defanged.

  • Freshsales removed its free plan entirely on July 21, 2026. The entry point is now a 21-day trial into Growth at $9/user/month. As of mid-November, Freshworks' own meta description still advertised the free plan.
  • Kaspr ended unlimited free B2B emails — the free tier now includes 15 emails/month, and "unlimited" moved behind paid plans with annual export caps.
  • 6sense discontinued its free Sales Intelligence tier on August 12, 2026.
  • RB2B restructured in January 2026: the free tier lost person-level identification — the feature that made RB2B famous — and now returns company-level data only.
  • Streak stripped pipelines out of its free plan (email tools remain) and killed its Solo plan.
  • Seamless.AI free credits are now a one-time 50, not a monthly allowance.
  • Apollo kept its generous 10,000-credit fair-use allowance on the renamed Starter (free) plan — but gated it behind a corporate email address. Sign up with a personal address and you get 100 credits/month, a 100x difference the pricing page doesn't exactly headline.
  • Clearbit's free tools are simply gone — the standalone product shut down and resurfaced as HubSpot Breeze Intelligence, priced in HubSpot credits.

The notable counter-move: Hunter.io doubled its free tier to 50 credits/month and made sequencing free. When everyone else retreats, generosity is a differentiation strategy.

Why the retreat? Our read: free tiers were designed to convert humans, and in 2026 a growing share of signups are agents and scripts extracting data. When AI agents are doing the searching and the scraping, a free tier's unit economics break.

Pattern 2: Rename the lineup, move the price​

Six vendors raised list prices in our dataset. The interesting part is how. The honest raise — same plan, higher number — is now the minority. The common move is a full lineup rename, with the raise hidden inside the migration map.

  • Pipedrive retired all five of its long-standing plans (Essential, Advanced, Professional, Power, Enterprise) for a new four-plan lineup: Lite $14, Growth $39, Premium $59, Ultimate $79 per user/month annual. Per Pipedrive's own migration doc, Advanced customers map to Growth — a 57% price increase for roughly the same tier. Premium also took a quiet $49 to $59 raise on newer SKUs. Full details in our Pipedrive pricing rebuild.
  • Clay rebuilt its plans in March 2026: Starter $149, Explorer $349, and Pro $800 became Launch $185 and Growth around $495, with credit expansion packs above that. Entry price up 24%, and the real money moved into credit expansions at roughly $880 per 20,000 credits.
  • Attio raised Plus from $29 to $35 and Pro from $69 to $79 (annual) in July 2026 — a straightforward raise, verified against its embedded checkout data.
  • LinkedIn Sales Navigator Core went from $99.99 to $119.99/month.
  • Salesforce raised core Sales Cloud plans in August 2025 ($25/$100/$175/$350) and introduced the $550/user Agentforce 1 Sales bundle above them.
  • Propensity doubled its entry price from roughly $1,000 to $2,000/month.

If you track vendor pricing by plan name — as most procurement teams do — a rename resets your baseline. That's not an accident.

Pattern 3: Seats are losing to pooled usage​

Four vendors in the dataset abandoned classic per-seat pricing this year:

  • Lusha retired per-user plans for account bundles: fixed plans with pooled credits and bundled seats (e.g., Pro at $69.90/month includes 600 pooled credits across 2 seats). Extra seats on Pro cost $49.90/month and add zero credits — the credits are the product now. Full breakdown.
  • Woodpecker killed per-slot pricing for prospect-volume tiers ($35 to $159/month by contacted prospects) with unlimited seats and inboxes — you pay for reach, not people.
  • Mixmax collapsed its plan ladder into a single plan priced by sequence recipients ($29/$89/$129 per user/month by volume) — every feature included, volume is the dial.
  • Clay gives even free workspaces unlimited seats and meters everything through credits and actions.

The logic is sound from the vendor side: when AI does more of the work, seat counts stop growing — so vendors re-anchor on the thing that does grow, which is usage. For buyers it means the quoted price tells you almost nothing until you've modeled your volume.

Pattern 4: AI gets its own meter​

Six vendors attached a new usage meter for AI features on top of existing subscriptions — none of them priced the same way:

VendorAI meterThe math that matters
Salesforce Agentforce$0.10 per action or $2 per conversationBreak-even between the two is about 20 actions per conversation — we ran the numbers
Microsoft Dynamics 365Copilot Credits, $0.01 each (or $200 per 25K)Sales Development Agent burns 400-700 credits per worked lead; usage billing starts Dec 1, 2026 — details
HubSpotBreeze creditsAbout 100 credits per AI-worked lead; a Pro allowance of 3,000 credits is roughly 30 leads/month — analysis
FreshworksFreddy AI Agent sessions$49 per 100 sessions, on top of plan pricing
NutshellMetered "AI outcomes"New plan lineup meters AI usage separately
monday.comAI credits bundled into plansAI ARR hit $36M in Q3 FY2026, but it blends credits, agents, and bundled allowances

The pattern inside the pattern: the subscription gets you the seat; the meter gets you the work. Every one of these meters prices an activity, not an outcome — you pay per action whether or not the lead was worth working. That's precisely backwards from how a sales team thinks, and it's why modeling credit burn before you commit is now a required step in buying, not an optional one.

Pattern 5: The barbell — cheaper self-serve, pricier contracts​

The middle of the market is hollowing out in both directions:

Moving up: Warmly killed its roughly $700/month self-serve plan for annual contracts at $10K, $20K, and $30K per product line. Lead Forensics remains quote-only, with real contracts we found ranging from $6K to $95K/year. Leadfeeder's upper tiers went annual-only.

Moving down: Unify GTM went the other way — it killed its $700/month Growth plan and replaced it with free, $20/seat, and $60/seat self-serve tiers. RocketReach cut Essentials from $53 to $39 and Pro from $179 to $79. Hunter doubled its free tier.

Consolidation is part of the same story: Visitor Queue was acquired by Leadinfo (its site now redirects, with euro-denominated plans), and Clearbit's remnants live inside HubSpot. If your vendor's pricing page starts sprouting "talk to sales" where numbers used to be, that's the barbell pulling them upmarket — and your renewal is where you'll feel it.

The full change log​

Every change we verified, with dates. "Verified" is when we last confirmed it against a first-party source.

ToolWhat changedEffectiveVerified
Pipedrive5 plans retired → Lite/Growth/Premium/Ultimate; Advanced→Growth = 57% raiseRolling, 2025-2026Nov 5
FreshsalesFree plan removed; Freddy AI sessions addedJul 21, 2026Nov 5
CopperStarter plan killed → Basic $23/Pro $59/Business $99May 2, 2026Nov 4
StreakSolo plan dead; free plan lost pipelines2026Nov 4
NutshellRebuilt to 5 plans; metered AI outcomes2026Nov 4
SalesforceCore raise + $550 Agentforce 1 Sales bundleAug 2025Nov 4
Sales NavigatorCore $99.99 → $119.99/mo2026Nov 2
ClayPlan rebuild; entry $149 → $185; credit expansionsMar 11, 2026Nov 2
KasprFree unlimited emails → 15/mo2026Nov 2
Hunter.ioFree doubled to 50 credits/mo + free sequencing2026Nov 2
ApolloFree renamed Starter; corporate-email gate on 10K credits2026Nov 2
LushaPer-user plans → account bundles; phone lookups 10 → 5 creditsFeb-Sep 2026Nov 3
RocketReachEssentials $53 → $39; Pro $179 → $792026Nov 2
Seamless.AIFree credits now one-time 502026Nov 2
AttioPlus $29 → $35; Pro $69 → $79 annualJul 2026Sep 30
Unify GTM$700/mo Growth killed → Free/$20/$60 seats2026Oct 1
WoodpeckerPer-slot plans → prospect tiers, unlimited seats2026Sep 25
MixmaxPlan ladder → one plan priced by sequence recipientsOct 12, 2026Oct 13
WarmlySelf-serve killed → $10K-$30K annual contracts2026Nov 15
RB2BFree tier lost person-level ID; new $79/$149/$199 ladderJan 22, 2026Nov 15
Visitor QueueAcquired by Leadinfo; site redirects; €69-€359 plans2026Nov 15
ClearbitStandalone dead → HubSpot Breeze Intelligence creditsApr 2025Nov 15
6senseFree Sales Intelligence tier discontinuedAug 12, 2026Nov 15
PropensityEntry roughly doubled to $2K/mo2026Sep 28
LeadfeederNew lineup; upper tiers annual-only2026Nov 15
HubSpotBreeze credit meter added for AI leads2026Oct 5
Microsoft D365SDA usage billing via Copilot CreditsDec 1, 2026Nov 16
monday.comAI credits bundled; agents metered2026Nov 10

Stable, for the record: Close, Zoho, Salesflare, Insightly, Wiza, Snitcher, Salespanel, and HubSpot's core seat prices all matched our previous verification. Ten of 38.

What buyers should do with this​

  1. Re-quote before every renewal, by feature list — not plan name. Plan names are now marketing surfaces that get repainted mid-contract. Pipedrive's Advanced→Growth map is the cautionary tale: same name tier, 57% more.
  2. Get grandfathering in writing. Clay grandfathered legacy plans; Pipedrive reprices you to list the moment you switch. Which policy your vendor follows belongs in your contract, not in a support thread.
  3. Model credit burn before buying any AI agent add-on. Divide the bundled allowance by the per-task burn rate. HubSpot's 3,000 Pro credits sound generous until it's 30 leads a month. Microsoft's 400-700 credits per worked lead means the included 1,000 credits cover about two.
  4. Treat free tiers as temporary. Eight of the tools we tracked cut or killed theirs this cycle. If your workflow depends on a free allowance, you're one pricing-page deploy away from a migration project.
  5. Date-stamp your own pricing research. Anything older than a quarter is suspect. Most of the third-party pricing content we cross-checked this fall — including some high-ranking aggregators — was selling retired plans.

Why we track this​

We maintain pricing breakdowns for 70+ sales tools and re-verify the highest-traffic ones continuously — it's how we caught most of these changes within weeks. The same discipline runs our product: MarketBetter watches buying signals so your SDR team acts on what's true today, not what a stale database says.

If your 2027 budget is built on 2025 pricing screenshots, it's wrong. And if your outbound is built on stale signals, it's wrong the same way.

Want to see what signal-driven, up-to-date prospecting looks like? Book a demo →

6sense Goes Agentic: BDR Agent, AI Visibility, and What Breakthrough 2026 Actually Shipped

· 9 min read
Sunder Iyer
Founder, marketbetter.ai

On November 3, at its Breakthrough 2026 conference in Las Vegas, 6sense announced a full rebrand: from "your ABM partner" to "your partner for agentic GTM." Alongside the new positioning came a slate of product announcements — a BDR Agent, an AI Visibility product, explainable account scores, and a generally available MCP server.

The ABM category leader just told the market that account-based marketing "has evolved to include agent-based marketing." That's a big statement from the company that largely defined the ABM category, and it's worth unpacking what was actually announced, what's shipping versus what's in beta, and what it means if you run an SDR team that doesn't have a six-figure martech budget.

Three Tools for 1,100 Endpoints: How We Opened MarketBetter to AI Agents

· 10 min read
Sunder Iyer
Founder, marketbetter.ai

Sometime this year, the "user" on the other end of your product stopped being reliably human. We've written about AI agents replacing search traffic and about vendors coining "business-to-agent" as a category. This post is the practical sequel: what we actually shipped so that an AI agent can operate MarketBetter — not read about it, operate it.

The short version: our entire API surface — over 1,100 operations at the time of writing — is now reachable by agents through exactly three generic tools, on two rails. One rail is a remote MCP connector. The other is a bring-your-own-key skill for an agent tool catalog. Neither rail ships a single hand-written per-feature tool.

They Asked for a Demo, Saw Your Open Times, and Never Booked. Now Someone Notices

· 11 min read
Sunder Iyer
Founder, marketbetter.ai

There's a lead-leak in B2B inbound that almost nobody instruments: the person who fills out your demo form, passes qualification, gets shown a live calendar with open times — and closes the tab.

Your CRM says a lead was created. Your form analytics say the form converted. Marketing counts it. And then nothing happens, because every system involved considers its job done. The one thing that didn't happen is the only thing that matters: no meeting exists.

In late October, one of our customers asked us for exactly this: follow up with people who fill out our demo form but don't book a meeting, using what we know to personalize the outreach. We shipped it over the following week, and it's live now as an opt-in for every workspace running our scheduler. This post is what it does — and, just as deliberately, what it refuses to do.

Congrats on the New Role… You Took in 2018. Champion Alerts Now Fire Only on Fresh Moves

· 9 min read
Sunder Iyer
Founder, marketbetter.ai

A champion changing jobs is the warmest signal in B2B sales. Someone who already bought from you just landed somewhere new, with new problems and new budget. Every champion-tracking vendor — us included — has been telling you some version of that story for two years.

Here's the part of the story nobody leads with: most of the "job changes" these systems detect aren't new. They're old news your CRM was too stale to notice.

We pulled the numbers on one customer's workspace this fall. Of 216 champion job-change signals detected over a 60-day stretch, only 20 involved a role the person had started within the previous 90 days. 98 of them — nearly half — were moves into roles the person had started more than two years earlier. Reps enrolled them in outreach anyway, and real prospects got "congratulations on the new role" emails for jobs they took in 2018.

That's not a warm signal. That's a cold email with a timestamp problem, sent to the exact person whose goodwill you could least afford to burn.

So we changed how the whole pipeline works. As of November 3, a detected champion move becomes a signal — with outreach, owner routing, and email lookup — only when the new role started within the last 60 days. Everything older is still recorded on the champion's history, but nobody gets congratulated for ancient history again.

The Pipeline You Already Paid For: A Nurture Plan for Not-Now and Lost Deals

· 6 min read
MarketBetter Team
Content Team, marketbetter.ai

Think about what it cost to get one buyer to say "not now". Ads, events, content, SDR hours, maybe a demo and a few weeks of an AE's time. Then the deal stalls or the buyer says the timing is wrong, and in most companies that person drops into a CRM view nobody opens again.

That list is the cheapest pipeline you have. These people know your name, have seen the product and have told you the problem is real for them. What they lacked was timing, budget or the right person in the seat. All three change, and when they do, you want to be the vendor they already think of.

This guide is a plan a marketing team can own for three groups: people who said "not now", people who took a demo and never bought, and deals lost on timing, budget or a change of decision maker.

Is Your Website Intent Real? A 30-Minute Audit to Separate Buyers From Bots

· 5 min read
MarketBetter Team
Content Team, marketbetter.ai

Website visitor data is one of the most persuasive reports a marketing team can show: dozens of named companies on your site this week, some of them on the pricing page. It is also one of the easiest reports to get wrong.

A good share of what looks like intent is not a buyer at all. It is a crawler checking your pages, a security scanner, your own team, an existing customer logging in, or a partner looking up a link. Route that to sales and two things happen: reps waste time on accounts that were never in market, and they stop trusting the signal, including the real ones.

This is a 30-minute audit any marketing or RevOps leader can run on last month's visitor data. You do not need a data team. You need an export and a little scepticism.

Your Campaigns Can Now Go Out in 17 Languages — and Every Draft Gets Checked

· 8 min read
Sunder Iyer
Founder, marketbetter.ai

Most of our customers sell globally. IoT connectivity deals in Germany and Japan, EHS compliance deals across five European markets, logistics deals in Brazil. Their buyers read email in their own language — and reply rates show it.

Until this week, running a German campaign in MarketBetter meant writing your templates in German and trusting the AI to take the hint. It usually did. "Usually" is not a word you want anywhere near your outbound.

So we shipped the obvious thing, and then the less obvious thing that makes it actually trustworthy. In the campaign builder's Voice step — right next to tone and AI Context — there's now a Language control. Pick German, and the campaign's emails are generated in German. All of them. Every touch, every follow-up, for every contact on the list.

The pickable set today: Czech, Dutch, French, German, Greek, Hungarian, Italian, Japanese, Korean, Polish, Portuguese (Brazilian and European), Romanian, Spanish (Castilian and Latin American), Turkish, and Ukrainian. English isn't on the list because the default already handles it — more on that below.