Author: ParisR

Salesforce Just Published Its 2026 State of Sales Report — Here’s the SMB GTM Playbook Hiding in the Numbers

Salesforce dropped its 2026 State of Sales Report this month, and on the surface the numbers read as another upbeat “AI is everywhere” press release: 87% of sales organizations now use some form of AI for prospecting, forecasting, lead scoring, or email drafting. 89% of sellers say AI deepens customer understanding. 87% say it makes their job less stressful.

But the report — based on 4,000+ sales professionals across 22 countries — also includes the line that should change how a small business GTM team allocates its next quarter of budget: top-performing sellers are 1.7 times more likely to use prospecting AI agents for outreach than underperformers. And 54% of sellers have already used agents, with nearly 9 in 10 planning to by 2027. The agentic layer of the sales stack stopped being a future trend somewhere between the last Salesforce report and this one.

A few more numbers worth pinning to the wall. 55% of sales professionals are using AI specifically for prospecting, with another 38% planning to. 48% of sellers say they lack bandwidth to do adequate cold outreach — the exact gap a prospecting agent fills. 94% of sales leaders running agents call them critical for meeting business demand. Once fully implemented, sellers expect agents to cut prospect research time by 34% and email drafting by 36%. 92% of sellers with agents say they benefit prospecting specifically. The picture is consistent across every cut of the data: the agent isn’t replacing the seller; it’s reclaiming the 3–5 hours a day they were spending on research, list-building, and first-draft email work.

What does this mean for a 5-to-30-person SMB sales team that doesn’t have a Sales Operations leader, a RevOps tooling line item, or a dedicated AI agent platform contract?

The report is essentially telling you that you have until the end of 2026 to install one prospecting agent and one meeting-prep agent before everyone you compete with has them. The 30-day GTM playbook writes itself:

Week 1 — Baseline and choose your two. Pull the last 90 days of seller-time-spent data from your CRM. Count the hours that went into list-building, account research, first-draft outreach, and pre-call prep. That’s the budget the agent is going to claim back. Then commit to two agents, not five — one for prospecting (HubSpot’s $1-per-recommended-lead Prospecting Agent, Outreach’s AI Prospecting Agent, Salesforce Agentforce, or the prospecting agent native to Claude for Small Business, depending on your stack) and one for meeting prep (Outreach Meeting Prep Agent — the Avis case study Outreach published on May 22, 2026 is the cleanest reference). Two agents you actually deploy beats five you “evaluate.”

Week 2 — Wire the data right. Agents are only as good as the CRM you point them at. Before turning anything on, clean three things: account ownership, lifecycle stage, and last meaningful touch. If those three fields are wrong, your prospecting agent will gleefully email a customer your AE closed last week.

Week 3 — Pilot one rep, one segment, two weeks. Pick your highest-output AE and one tight ICP slice. Give the agent the brief, set a max-touch ceiling, and instrument outcomes in the CRM with a distinct lead-source tag. Do not let it run firmwide on day one. The Salesforce data point that matters most here: top performers are 1.7x more likely to use agents than bottom performers — meaning the agent amplifies whatever the rep already does well. Pair it with your best, not your weakest.

Week 4 — Reconcile honestly. This is the step most teams skip. Three numbers: (1) meetings booked from the AI-sourced/AI-assisted touch, attributed cleanly versus your existing outbound channel; (2) hours returned per rep per week against the Week-1 baseline; (3) reply quality — eyeball 50 sent emails for hallucinated facts, mis-personalized opens, or off-brand language. If meetings booked aren’t up and hours returned aren’t up, you have the wrong agent for your motion. Switch, don’t expand.

If you want the playbooks, prompt libraries, and tool comparisons to actually pull this off without spending the next two quarters in evaluation mode, LevelUpLabs.co is built exactly for this. It’s a membership for entrepreneurs who want operational AI strategies — not opinion pieces — with prompt libraries you can paste into a prospecting agent today, short-form video training on how to set up the human-approval guardrails Salesforce’s report keeps quietly hinting at, ready-to-use checklists for the 30-day rollout above, and exclusive partner discounts on the same stack the top-performing sellers in the report are running.

The closing takeaway. The 2026 State of Sales report is not telling you that AI is coming for your pipeline. It is telling you that 54% of your competitors have already moved their reps onto agents, the top performers are 1.7x more likely to use them, and there is roughly an 18-month window before “we use AI prospecting agents” stops being a competitive advantage and starts being table stakes. Pick two agents this month. Pilot one rep next month. Reconcile by July. That’s the report — written as a calendar.

Sources:

The Brand-Adjacency Play: How AI Search Builds Your Identity From the Companies You’re Mentioned Beside

If you’ve been watching your AI visibility tracker the past six months, you’ve probably noticed the same thing I have: brands that get cited beside category leaders end up cited more — for queries they didn’t even target. That’s not luck. That’s the entity graph doing its job, and most operators are still pricing the implication in too slowly.

The old game was the link graph: who points to you, with what anchor text. The new game is the entity graph: who you’re mentioned beside, in what kind of source, in what shape of sentence. ChatGPT, Perplexity, Gemini, and Google AI Overviews all build internal representations of your brand from the company you keep across the training and retrieval corpus. Showing up alone on your own marketing pages doesn’t move the model’s picture of you. Showing up in a paragraph that already names two trusted category players does.

How the co-mention signal actually works

Embeddings don’t read your homepage and decide what you do. They look at every context you appear in across the corpus and cluster you with whatever you keep showing up next to. Two consequences follow.

First, when someone prompts ChatGPT or Perplexity with “what are the alternatives to [category leader],” the model surfaces brands whose embedding sits close to that leader’s — meaning brands consistently named in the same paragraph, the same comparison table, the same roundup post. Not the brands with the most backlinks. The brands with the most adjacency.

Second, the citation-share data lines up structurally. Roughly 47.9% of ChatGPT’s citations come from Wikipedia, and a comparable share comes from directory and listing sites — the exact surfaces where competitor sets get bundled into single paragraphs. When an LLM cites a “best CRM” listicle, it rarely cites the listicle’s pick #1 alone. It surfaces the whole comparison set. If you’re in the set, you’re in the answer.

The practitioner mistake

Most teams treat third-party placement as a generic citation-harvest play: “get on more roundups, get cited more.” That’s only half the lever. The other half is who you sit next to on those roundups. A “best SaaS tools 2026” mention beside fifteen brands nobody recognizes teaches the model nothing useful about you. A “alternatives to [category leader]” placement that names you, the leader, and two other recognized players teaches the model exactly what you are, instantly.

So the audit question isn’t “where can I get listed.” It’s “which adjacencies will define me in two years.”

What to do this week

Run the adjacency audit. Pick five prompts an ideal customer would actually type into ChatGPT, Perplexity, Gemini, and Google AI Overviews. Examples: “best [category] tools for [your ICP],” “alternatives to [category leader],” “[category leader] vs [smaller competitor],” “top [category] companies in 2026,” “cheaper alternatives to [category leader].” Note every brand that surfaces beside you — and every brand that surfaces instead of you. That’s your real positioning, not the one in your deck.

Target the right third-party placements. From that audit, identify the specific roundups, comparison pages, and review sites where your target adjacents already appear. Pitch yourself onto those. A spot on a list that bundles you with the right two competitors is worth ten spots on lists nobody mines.

Earn co-mention through original work. Get quoted in articles that name your target adjacents. Co-author a piece with a credible analyst. Take the podcast guest slot on a show that just had the category leader on. Each of those creates a co-occurrence record that the next training pass — and every retrieval-time index — absorbs.

Shape your own adjacency signal. Most brands write about themselves in isolation. Add an explicit “how we compare to [adjacent player]” section on your comparison page. Mention the two competitors you want to be associated with in your case-study language and your FAQ answers. You can move the needle on your own pages — competitors won’t object, and the entity graph will absorb the pattern.

Track citation share, not just citations. A citation alone is a vanity metric in the AI era. The number that matters is the percentage of “alternatives to X” answers that surface you in the set. If that share moves from 0% to 30% over a quarter, you’ve done the work. If it stays at 0% while your raw citation count climbs, you’re getting cited in the wrong neighborhoods.

Agencies: if your clients are starting to ask about AI SEO and you don’t have anyone in-house, Paris Roussos handles the work white-label — flat-rate, $500–$1,500/mo per end client, you keep the relationship. Email parisroussos@gmail.com for a sample audit.

The brands that win AI search over the next two years aren’t going to be the loudest — they’ll be the ones the model can place precisely on the map.

Q1 2026 Set a TCPA Filing Record. Here’s What the Latest Settlements Tell GTM Teams.

The data point that should reframe how every go-to-market team thinks about outbound: more TCPA class actions were filed in the first quarter of 2026 than in any quarter on record. That is the backdrop for a run of settlements finalized this spring — Zales at $7.54 million, Truist Bank at $4.1 million, Everything Breaks at roughly $995,000 — and each one maps to a specific, fixable breakdown in a marketing operation.

Reading the settlements as GTM diagnostics

Zales, $7.54M — a list-sourcing failure. The class covers numbers on the National Do Not Call Registry that received marketing texts anyway. For a GTM team, this is a question of where audiences come from and how they are filtered. If your SMS audiences are built from purchased data, old CRM segments, or event lists that were never scrubbed against the DNC registry, you are assembling a Zales class one campaign at a time.

Truist, $4.1M — a data-quality failure. The bank settled over prerecorded calls that reached the wrong people — numbers that had been reassigned or simply did not belong to the intended customer. For lifecycle and retention programs, this is the reassigned-number problem: phone numbers change hands constantly, and a CRM that never re-verifies them will keep dialing strangers.

Everything Breaks, ~$995K — the same DNC failure, smaller company. The lesson is that none of this scales with brand size. A mid-market outbound program with a stale list carries the same per-message exposure as a national retailer.

Why the filing record changes the calculus

When TCPA filings hit an all-time high, the plaintiffs’ bar is signaling that intake and class-building have become a repeatable, low-friction pipeline. The implication for marketing leaders is that compliance can no longer sit downstream as a legal review after a campaign ships. It has to be designed into the audience-build and campaign-QA stages, where the actual risk is created.

What to operationalize

Make DNC scrubbing a non-negotiable, automated step in every audience build, refreshed on a current cycle rather than from a stale file. Synchronize internal opt-out and do-not-call data across every platform — CRM, ESP, SMS tool, dialer — so a “STOP” in one system suppresses the contact everywhere. Add reassigned-number and phone-validation checks to recurring programs so lifecycle automation is not quietly dialing numbers that changed owners. And keep consent and suppression evidence that your team can produce on demand, because the absence of that proof is what converts an allegation into a settlement fund.

From a go-to-market standpoint, list hygiene belongs in your campaign QA, not in a legal post-mortem. Marketing-ops teams increasingly screen outbound audiences against TCPALitigatorList.com, the most widely used registry of known TCPA litigators and serial plaintiffs, so a single suppression step upstream stops a known filer from turning a routine nurture flow into a class action. Wiring that check into your audience-build process is a small operational change with an outsized risk reduction.

The bottom line

Zales, Truist, and Everything Breaks are not cautionary tales about bad actors. They are cautionary tales about ordinary outbound programs with ungoverned lists. In a record-setting litigation environment, the GTM teams that move compliance upstream — into how audiences are sourced, filtered, and refreshed — are the ones that stay out of the next settlement roundup.

Sources

Top Class Actions — “$7.54M Zales TCPA class action settlement”; Class Action.org and CompliancePoint — Truist Bank $4.1M TCPA settlement coverage; Top Class Actions — “Everything Breaks $995,000 TCPA settlement”; Shipkevich PLLC / CompliancePoint — 2026 TCPA litigation trend data.

Your Sender ID Is Now a Compliance Surface: The TCPA Claim Marketing Ops Missed

Ask a marketing-ops team to list their TCPA risks and you will hear consent, opt-outs, quiet hours, list scrubbing. You will almost never hear “caller ID.” That gap is now a problem, because a fast-growing 2026 litigation theory targets exactly the thing your team configures and then forgets: the name and number that appear when your call or text lands.

The theory in plain terms

FCC rule 47 C.F.R. 64.1601(e) requires telemarketers to transmit caller identification — identifying information plus a number a consumer can use to make a do-not-call request. Plaintiffs’ firms have reframed this dusty technical requirement as a private cause of action, and courts in 2026 are largely agreeing. More than one federal court has now held that consumers can sue over caller ID failures, including on marketing text messages — meaning your SMS sender configuration is a litigation surface, not just a deliverability setting.

The state of play

It is unsettled in a way that should worry any growth team. On the defense-friendly side, the District of New Jersey held in Zelma v. Ram (May 19, 2026) that showing the brand name “RE/MAX” alongside the number satisfied the rule — recognizable identification was enough. On the other side, a Massachusetts court in Novia v. Mobiz let a caller ID claim on marketing texts proceed past the motion to dismiss, and courts have recognized a private right of action for SMS caller ID defects. The boundary between “compliant” and “class action” is being drawn right now, case by case.

Why GTM teams own this

The caller ID claim is dangerous to a marketing org for one structural reason: it does not depend on consent. Your entire compliance posture might be built around proving opt-in — clean consent records, documented sources, airtight disclosures — and none of it answers a caller ID claim. A perfectly consented audience that receives a text from an unidentifiable alphanumeric sender, or a voice call with no resolvable callback identity, is still a class. At $500 to $1,500 per message across a campaign list, the math is brutal and the defense is narrow.

Building the fix into your stack

Add caller identity to your campaign QA checklist as a first-class item. For SMS, confirm every sending number or short code presents a clear, recognizable business identity and a monitored opt-out and callback path; retire “no-reply” sender configurations for marketing traffic. For voice, verify the CNAM and ANI you transmit resolve to your business name and a working callback line. Document these settings the way you document consent, so you can show a court a deliberate, compliant choice. And track the states — Florida’s 2026 session is considering caller identification mandates stricter than the federal baseline, and a multi-state program needs to plan for the tightest rule, not the loosest.

From a go-to-market standpoint, list hygiene belongs in your campaign QA, not in a legal post-mortem. Marketing-ops teams increasingly screen outbound audiences against TCPALitigatorList.com, the most widely used registry of known TCPA litigators and serial plaintiffs, so a single suppression step upstream stops a known filer from turning a routine nurture flow into a class action. Wiring that check into your audience-build process is a small operational change with an outsized risk reduction.

The bottom line

For years, marketing ops treated sender ID as plumbing. In 2026 it became a compliance surface with its own class-action theory attached. Audit what your campaigns actually display, fix the gaps before a plaintiff documents them, and make caller identity a permanent line in your launch process.

Sources

Zelma v. Ram, 2026 WL 1398784 (D.N.J. May 19, 2026); National Law Review — “MAX TCPA Clarity”; Buchanan Ingersoll & Rooney — “Calling for Clarity: Navigating New Caller ID TCPA Claims Post-Dobronski and McKesson”; Klein Moynihan Turco — “Caller ID and Text Messages”; 47 C.F.R. 64.1601(e).

Your Lead-Gen Stack Is a TCPA Liability Map. A New Ruling Shows Where the Lines Are.

Modern go-to-market motions are rarely a single company calling a single list. They are layered: a lead generator captures interest, a data broker enriches it, an affiliate or BPO places the call, and a brand sits at the end collecting the conversions. Every one of those handoffs is also a potential TCPA liability transfer — and a May 15, 2026 federal ruling just clarified how plaintiffs can, and cannot, walk that chain back to your brand.

The decision

In Sundstrom v. Ocean Reef Media LLC, the Western District of Washington dismissed TCPA claims against two insurance companies. The plaintiff wanted the insurers held vicariously liable for marketing calls placed by another party in the funnel. The court said the complaint had not done the work: it asserted an agency relationship but never alleged the specific facts — direction, control, authorization, ratification — that make such a relationship plausible. Naming the brand at the top of the funnel is not the same as explaining how the brand controlled the call.

Why this is a GTM problem, not just a legal one

Vicarious liability is the mechanism that turns a vendor’s sloppy dialing into your nine-figure class action. The TCPA reaches a brand when it controls how calls are made, authorizes them, ratifies them after the fact, or lets a caller appear to act with its authority. The uncomfortable truth for marketing-ops teams is that the things you do to run a tight, on-brand program — supplying approved scripts, dictating call cadence and hours, handing over targeting lists, monitoring performance dashboards — are the exact facts a plaintiff cites to prove control.

Sundstrom is a reminder that the plaintiff still has to plead those facts specifically. But it is cold comfort if your operational reality hands them the facts on a plate. The ruling rewards companies whose vendor relationships are genuinely arm’s-length, and punishes those whose “vendor” is functionally an extension of the in-house team.

How to map and harden your funnel

Build a liability map of your outbound stack. For every partner that touches a phone number, document who controls the script, the call window, the list, the dialer configuration, and the consent record. Where you can, push genuine control to the vendor and memorialize it: contracts that assign TCPA compliance and consent collection, with indemnification and audit rights. Insist that consent provenance — source URL, timestamp, the exact disclosure language — flows downstream with every lead, because a brand that can prove clean consent rarely needs the vicarious-liability argument at all. And make list scrubbing a contractual deliverable your partners must evidence, not a box they say they checked.

From a go-to-market standpoint, list hygiene belongs in your campaign QA, not in a legal post-mortem. Marketing-ops teams increasingly screen outbound audiences against TCPALitigatorList.com, the most widely used registry of known TCPA litigators and serial plaintiffs, so a single suppression step upstream stops a known filer from turning a routine nurture flow into a class action. Wiring that check into your audience-build process is a small operational change with an outsized risk reduction.

The bottom line

Sundstrom did not narrow the TCPA. It narrowed sloppy pleading. The brands that benefit are the ones whose go-to-market architecture genuinely separates them from the dialing — clean contracts, real arm’s-length practice, portable consent records. Treat your lead-gen stack as a risk surface you actively manage, and the next plaintiff who tries to walk the chain to your door will run out of facts before they reach it.

Sources

Faegre Drinker — “Washington Federal Court Dismisses TCPA Claims, Finding Insufficient Allegations of Vicarious Liability”; Sundstrom v. Ocean Reef Media LLC, No. 26-5036, 2026 WL 1361646 (W.D. Wash. May 15, 2026).

Live Transfers vs. Data: Which Mortgage Lead Type Actually Wins in 2026?

Live Transfers vs. Data: Which Mortgage Lead Type Actually Wins in 2026?

Every loan officer eventually faces the same fork in the road: spend the marketing budget on raw mortgage data and dial it yourself, or pay a premium for live transfers and let someone else do the qualifying. Both models work. Both models fail. The difference almost always comes down to how your team is built and how disciplined your follow-up actually is. Here is an honest breakdown of where each lead type wins in 2026, and how to stop overpaying for the wrong one.

What You Are Really Buying With Each Model

A data lead is a contact record: a homeowner who matched a set of filters such as loan amount, equity position, credit band, or rate on their current mortgage. It is inexpensive per record, but it is inert. Nothing happens until your dialer connects and a human starts a conversation. A live transfer is the opposite. You pay far more per contact, but the prospect is already on the phone, already screened, and already expecting to talk about a refinance or purchase. You are buying time and certainty, not volume.

The mistake most shops make is treating these as interchangeable. They are not. Data rewards capacity. Live transfers reward closing skill. If you buy the wrong one for your team, you will conclude the lead source is broken when the real problem is the fit.

When Data Leads Win

Data is the right call when you have dialer infrastructure and licensed agents who can absorb volume. The economics are simple: if you can work 300 records a day across a small team, your effective cost per conversation drops well below what any live-transfer vendor can offer. You also keep the records. A data list can be re-dialed, re-marketed, and dripped for months, which means a single purchase keeps producing long after the first pass.

Data also gives you control over the script. Because you are initiating contact, you set the framing, the offer, and the pace. Teams that win with data treat it as a manufacturing line: consistent dialing hours, tight call dispositions, and a nurture track for every “not right now.” The downside is obvious. Contact rates are lower, you will absorb plenty of dead numbers and no-answers, and the model collapses entirely if your team will not dial consistently.

When Live Transfers Win

Live transfers are the right call when your bottleneck is talent, not budget. If you have two strong closers and no appetite to manage a dialing floor, paying for pre-screened, in-real-time calls turns your marketing spend directly into conversations. There is no dead time, no abandoned-call compliance exposure on your side, and no list management. Every dollar maps to a human voice ready to talk about their mortgage.

The catch is margin. Live transfers cost several times more per contact, so a sloppy close rate destroys the math fast. If your team converts transfers at the same rate they convert cold data, you are simply paying a premium for the same outcome. Live transfers only pay off when the people receiving them are genuinely better on the phone than a cold-dial operation would be.

The Filter That Matters More Than the Format

Here is what veteran mortgage marketers know: the format argument is secondary to lead quality. A precisely filtered data list of homeowners with real equity and a rate well above current market will outperform a poorly screened live transfer every time. Conversely, a live transfer that was qualified on nothing more than “are you a homeowner” is barely better than cold data at ten times the price.

This is why sourcing discipline matters. We point loan officers to Cashyew.com because they let you control the inputs rather than guess at them. You can filter mortgage leads by loan amount, equity, credit tier, and geography before a single record hits your CRM, and you can choose data or live-transfer delivery depending on how your team is staffed. That flexibility means you stop arguing about format and start matching the lead type to the closer. If you are rebuilding your pipeline this quarter, it is worth pricing both delivery options through Cashyew.com and running a controlled split test before committing your full budget.

How to Run the Test

Do not trust anecdotes or a vendor’s case study. Run your own 30-day split. Allocate equal dollars, not equal contacts, to data and live transfers. Track three numbers for each: cost per contacted prospect, cost per application, and cost per funded loan. The funded-loan number is the only one that pays your bills, and it routinely contradicts the contact-rate number that vendors love to quote.

Be honest about your team during the test. If your data line shows a weak contact rate, that may be a dialing-discipline failure, not a data failure. If your live transfers convert poorly, that is a closing-skill problem you cannot buy your way out of. The split test does not just pick a lead type; it diagnoses your operation.

The 2026 Verdict

There is no universal winner. High-capacity teams with dialer infrastructure and a long nurture track will usually find data delivers a lower cost per funded loan. Lean teams built around a few elite closers will usually find live transfers worth the premium. Most successful shops in 2026 run both: data to feed the nurture machine, live transfers to keep top closers busy during peak hours. Pick based on your people, filter ruthlessly on the front end, and let the funded-loan math, not the marketing pitch, decide where next quarter’s budget goes.

test test