Something Inc.LoginSchedule a free consultation
STRATEGY

The Plan You Approve Is The Only Brake You Get

Salesforce put its outbound agent on a runtime built to chase a goal for weeks, and AI SDR agents that run that long turn one approval into six weeks of sending that nobody reads again.

STRATEGYOUTBOUNDNOV 2026

Every autonomous outbound tool before this one worked in task-sized units. You approved a sequence, it sent the sequence, and the blast radius of a bad decision was one send window. AI SDR agents built to pursue a goal for weeks break that arrangement, and the industry is discussing the wrong half of it.

Salesforce announced a long-horizon runtime at Dreamforce 2026 and made Hunter, its outbound sales agent, the first agent to run on it. The runtime has three named parts: memory across sessions, durable execution so a plan survives beyond one conversation, and dynamic steering so the agent updates its own configuration from feedback. Salesforce describes the point of it plainly in its own write-up of long-horizon agents: goals that take days and weeks, not chat turns.

Hunter sits on it as the outbound counterpart to Piper, which already handles inbound. Salesforce describes Hunter as working a pipeline from research through outreach, collaborating with sellers over weeks and months. Its agent portfolio announcement from September 11 lists seven named agents, and Hunter is the one still in pilot while the Agentforce page carries a general availability target of November 2026.

Most of the commentary has gone to whether an agent can write a decent first line. That question was settled months ago and the answer was boring: yes, roughly as well as a tired human, and copy quality was never the binding constraint anyway. We argued that when the data showed cadence outperforming copy on AI-run outbound, and nothing since has changed it.

The new variable is time. An agent with a six week goal, persistent memory and the ability to adjust its own plan is not a faster sequence tool. It is a thing that holds an objective longer than most quarterly reviews, on infrastructure you are responsible for.

TL;DR · 60 SECONDSWhat shipped: Salesforce announced a long-horizon runtime at Dreamforce 2026, built from memory, durable execution and dynamic steering, and Hunter is the first agent on it. Salesforce lists Hunter as pilot with general availability targeted for November 2026, while a Sales Cloud release page describes it as generally available this fall inside Coworker, so treat the exact date as unsettled. What Salesforce got right: there is a real human gate. When the agent wants to deviate from the approved plan, it stops and asks for approval, which is more control than most outbound tooling has ever offered. Our verdict: that design makes plan approval the entire governance decision, because anything inside an approved plan never triggers a prompt. Teams that approve a goal like book meetings with enterprise security buyers have authorized weeks of sending without reviewing any of it. Write the plan as a constraint set with explicit numeric ceilings and stop conditions, decide now who owns the domain the agent sends from, and re-baseline your deliverability reporting before general availability rather than after.

What Changed When AI SDR Agents Got A Long Horizon

AI SDR agents moved from executing a defined sequence to holding a business objective across sessions, which means the unit of human review changed from the message to the mandate. Salesforce names three capabilities that make this possible, and each one removes a natural checkpoint that used to exist by accident.

NOV '26
the general availability target Salesforce lists for Hunter on its Agentforce page, which shows the agent in pilot, while a Sales Cloud release page describes it as generally available this fall inside Coworker
WEEKS
the horizon Salesforce says the new runtime is built for, describing goals that take days and weeks rather than chat turns, with Hunter collaborating with sellers over weeks and months
3 PARTS
the components Salesforce names in the long-horizon runtime: memory across sessions, durable execution of a plan over time, and dynamic steering that updates configuration from feedback
1 OF 7
Hunter's position in the September 11 agent portfolio announcement, the only one of the seven named agents not yet generally available

Memory across sessions is the one that matters most and gets discussed least. A sequence tool forgets. Every run starts from the list and the template, so a mistake in week one does not compound into week five. An agent that remembers what it tried, what bounced and what a prospect said in a reply carries its own assumptions forward, including the wrong ones.

Durable execution removes the second checkpoint. A plan that persists beyond one session does not come back for re-authorization when someone logs in on Monday. It simply continues, which is the entire product promise and also the thing that should make a revenue operations lead sit up.

Dynamic steering is the genuinely new capability, and Salesforce has been careful with it. The agent updates its configuration from feedback, and when the resulting behaviour would deviate from the approved plan it prompts for approval first. In the deal-rescue example Salesforce uses, an outreach email bounces and the agent asks before sending a replacement that departs from what was agreed.

CREDIT WHERE IT IS DUEThat approval-on-deviation design is better than what the outbound tooling market has shipped for years. The standard pattern in cold email automation is an approval at setup and nothing afterwards, which is how teams discover in week three that a sequence has been emailing the same contact at two addresses. A gate that fires when the agent departs from plan is real engineering, not a compliance checkbox.

Why Duration Is The Part Nobody Priced

The approval gate fires on deviation from an approved plan, which means everything the agent does inside that plan proceeds without a prompt, and the specificity of the plan therefore sets the width of the autonomy. This is the inversion most buying committees have not worked through yet.

“An agent that deviates from the approved plan stops and asks. Read that twice, because the sentence is doing far more work than it looks.”

Consider two plans for the same goal. Plan A says: book qualified meetings with security leaders at companies between 500 and 5,000 employees. Plan B says the same thing, then adds a ceiling of two sending domains, a maximum of four touches per contact across 30 days, no new domain introduced without sign-off, hard stop on any list where the bounce rate passes two percent, and a weekly summary of everything sent. Both are approved plans. Plan A authorizes six weeks of unreviewed sending. Plan B authorizes a narrow, measurable campaign that will trip a gate the moment it strays.

Nothing in the product distinguishes them. The gate does its job identically in both cases. The difference is entirely in the quality of the thinking that went into the plan on day one, which is a governance problem dressed as a configuration screen.

AGENT BEHAVIOUR DURING A SIX WEEK GOALINSIDE AN APPROVED PLAN?DOES A GATE FIRE?WHAT YOU SHOULD PRE-DECIDE
Researching 400 more accounts that fit the stated profileYes, if the profile is broadNoA hard account ceiling per period, not just a profile
Adding a third sending domain to raise capacityDepends entirely on plan wordingOnly if the plan names domainsName the domains. Treat a new one as a deviation
Increasing touches from three to six after weak reply ratesUsually yes, cadence is rarely specifiedNoA touch ceiling per contact and per period
Emailing a contact who replied with a soft noOften yes, a soft no is not an opt outNoExplicit suppression rules for ambiguous replies
Rewriting the offer after a bounce, departing from the planNoYes, this is the documented caseWho holds approval authority and their response time
Continuing to send while the bounce rate climbsYes, unless a threshold is written inNoNumeric stop conditions tied to deliverability metrics

Look down the third column. One gate fires out of six plausible behaviours, and it is the one where the agent is behaving most carefully. The five silent rows are not product defects. They are the predictable result of approving an objective instead of a constraint set, and every one of them is fixable in the plan before anybody switches it on.

What AI SDR Agents Do To A Sending Domain Over Six Weeks

A long-horizon outbound agent accumulates sending reputation consequences on domains that outlive the campaign, so the cost of a badly scoped plan is not a wasted quarter of pipeline but a set of burned domains you have to replace and warm again. This is where the duration question stops being philosophical.

The mechanics have not changed just because the sender is now an agent. Volume ramps still matter, bounce rates still compound, and the thresholds teams believe in are mostly folklore, which we went through when we looked at the sending domain thresholds practitioners quote without checking. What changes is who is making the ramp decisions and how often anyone looks.

Domain and mailbox ceilings: no new sending identity without sign-off100%
Numeric stop conditions on bounce rate and complaint rate90%
Touch and frequency caps per contact and per account75%
Suppression rules covering ambiguous and soft negative replies60%
Weekly logged summary of everything sent, reviewed by a named person45%
Copy and offer review20%

Where we would place the gates first on a long-horizon outbound agent, ranked by how much damage the unguarded version does. Our recommended priority, not measured data

Copy review sits at the bottom deliberately. It is the thing teams instinctively want to control and the thing that matters least to whether your infrastructure survives the engagement. We made the same argument about reputation when AI-written outbound started showing up in sender reputation data, and the ranking has only hardened since.

There is an honest counterargument worth stating. An agent with persistent memory could plausibly manage deliverability better than a human sales development representative juggling six other tasks, because it can watch bounce signals continuously rather than at the weekly pipeline meeting. That is a real possibility and the early pilot claims gesture at it. Coverage of Hunter's pilot reports that one customer, Perk, built around 60 percent of its pipeline with the agent, which is a single pilot result and not a benchmark, as the people reporting it have been careful to note.

We looked hard at a similar promise before, when autonomous agents were being sold as self-correcting in the reality check we ran on the autonomous sales development category. The pattern then was that the self-correction worked on the metrics the vendor instrumented and was blind to the ones it did not. Deliverability has historically been in the second group, because the signals live with the mailbox provider rather than in the sending tool.

How To Write An Approval That Actually Constrains

A plan that constrains an agent reads like a contract rather than a brief: numeric ceilings, named resources, explicit stop conditions and a single person who holds approval authority with a stated response time. The objective goes in one line and the constraints take the rest of the page.

1Write the ceiling before the objectiveOpen the plan with the limits: how many accounts per period, how many sending domains by name, how many touches per contact, what time of day. An objective without a ceiling is an open mandate, and an open mandate is what makes the approval gate decorative. The objective is one sentence and it goes second.
2Name the sending identities explicitly, as a listList every domain and mailbox the agent may send from, by name, and state that introducing a new one is a deviation requiring approval. This single clause converts capacity expansion from a silent optimization into a gated decision, and capacity expansion is where domain damage begins in practice.
3Put numbers on every stop conditionBounce rate above a stated percentage, complaint rate above a stated percentage, reply sentiment below a stated threshold: each one needs a figure and a consequence. Pause the goal, do not just flag it. A stop condition that produces a notification nobody owns is not a stop condition.
4Define what a soft no means before the agent decidesAmbiguous replies are where automated outreach does its reputational damage. Write the suppression rule: not interested right now, wrong person, circle back next quarter. Each one needs a defined outcome and a defined re-entry window, or the agent's memory will keep the contact in play indefinitely.
5Assign approval authority to a person with a response clockThe gate only works if somebody answers it. Name the individual, name the backup, and state the window in which a deviation request must be answered before the agent pauses by default. Pausing on silence is the correct default and it needs to be written down.
6Require a logged weekly summary and actually read itAsk for a weekly record of every message sent, every domain used and every threshold approached. The agent is not hiding anything, but durable execution means nothing interrupts you to say what happened. The review has to be scheduled by a human because the software has no reason to schedule it.

None of this is exotic. It is the same discipline a competent agency applies to a managed sending programme, written down in a place the software reads. The reason it feels unfamiliar is that the previous generation of tooling enforced review through its own limitations, and those limitations are what just went away.

What Should You Do Before The GA Switch Flips?

Before general availability reaches your org, decide three things: which domains an agent may ever send from, who holds approval authority on deviations, and what numeric conditions pause a goal outright. Those three decisions take an afternoon and they are considerably harder to make retroactively.

The timing is unusually legible, which is a gift. Salesforce lists Hunter as pilot with a November 2026 general availability target on its Agentforce page, while a Sales Cloud release page describes it as generally available this fall inside Coworker. Those two statements are close but not identical, so the practical read is that access is arriving now in some orgs and shortly in others. Either way the window for deciding policy before somebody enables it is measured in weeks.

This is not an argument against agentic outbound. The runtime design is sound, the approval-on-deviation gate is genuinely more control than the sequence tools it replaces, and for B2B SaaS teams running thin development capacity the economics are obvious. The argument is narrower: the control surface Salesforce built only does work if the plan it compares against is specific, and nothing in the product will tell you that your plan is too vague.

Expect the first wave of bad outcomes to look like mysteries rather than failures. Reply rates that sag in week four for no visible reason. A domain that stops landing. An account that got touched eleven times by an agent faithfully executing an objective nobody re-read after the day it was approved. In each case the logs will show a system working exactly as instructed, which is the least satisfying kind of incident review there is.

The teams who handle the next two quarters well will not be the ones with the best prompts. They will be the ones who treated plan approval as the governance event it actually is, and who wrote down the numbers before the software asked for them.

DO THIS NEXTThis week, write a one page constraint set for any outbound agent your org might switch on: named domains, touch ceilings, numeric stop conditions, a named approver with a response clock, and suppression rules for ambiguous replies. Then re-baseline your current deliverability metrics so you have a before picture. That constraint set is the first artifact we produce on a cold email engagement, and it is the reason a cold email agency earns its keep on infrastructure rather than copy, because the domains outlive every campaign that runs on them.
Is Salesforce Hunter generally available yet?Salesforce's Agentforce page lists Hunter as pilot with general availability targeted for November 2026, while a Sales Cloud release page describes it as generally available this fall inside Coworker. Treat the exact date as unsettled and check access in your own org.
What is a long-horizon runtime?Salesforce's term for infrastructure that lets an agent pursue a goal across days and weeks rather than a single task. It has three named parts: memory across sessions, durable execution so plans persist, and dynamic steering so the agent updates its configuration.
Does the agent send email without approval?Inside an approved plan, yes. The gate fires when the agent wants to deviate from that plan, which Salesforce documents with a bounced-email example. So the breadth of the plan determines how much sending happens unreviewed.
What is the biggest risk with AI SDR agents?Sending domain reputation. A badly scoped plan accumulates bounces and complaints on domains that outlive the campaign, and replacing and warming new domains costs far more time than the pipeline the agent generated was worth.
Should we delay adopting agentic outbound?No. Decide policy first. Name the domains an agent may use, set numeric stop conditions, and assign approval authority with a response clock. That work takes an afternoon and does not require waiting for the product to mature.
How specific should an approved plan be?Specific enough that a reasonable deviation trips the gate. Include account ceilings per period, named sending domains, touch caps per contact, bounce and complaint thresholds with consequences, and defined handling for ambiguous replies.
Is Hunter different from existing AI SDR tools?The duration is different. Earlier tools executed sequences in task-sized units that forgot between runs. An agent with persistent memory and durable execution carries assumptions forward across weeks, including incorrect ones.
Who should own the agent internally?Whoever owns sending infrastructure, not whoever owns quota. The decisions that matter most are domain allocation and stop conditions, and those belong with the function that has to repair deliverability when something goes wrong.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.

TT
Tyler TruffiMANAGING PARTNER, SOMETHING INC.

Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.

Keep reading

All insights →
STRATEGYYour AI SDR Reply Rate Has No DenominatorTwo of the most widely read accounts of AI in sales development landed within days of each other, and they point in opposite directions. Jason Lemkin's six-month write-up at SaaStr reports an outbound agent sending 19,847 messages at a 6.7 percent overall response rate, roughly 3,000 a month against the 75 to 285 a month its human reps managed. Atlassian's Teamwork Lab, publishing November 9, 2026 after interviewing eight sales development reps, reports a blanket internal rule that no AI-written message reaches a customer without human review, and one rep who abandoned AI for outbound email entirely and booked ten meetings in a week writing like a person texting a friend. The reflex is to pick a side. The better move is to notice that neither account defines the denominator under its reply rate, and that SaaStr's own campaign figures do not reconcile against either obvious candidate. Volume changes what a reply rate means. Here is how to read one, which denominator to hold your program to, and when adding sends stops paying.TYLER TRUFFI · 12 MINSTRATEGYYour November Bounce Data Is ContaminatedOn November 5, 2026, senders began getting hard rejections from Outlook.com, Hotmail, Live and MSN addresses, carrying the error 550 5.7.1 Message rejected: Message lacks SPF/DKIM/DMARC Auth Records. Plenty of those senders had passing SPF and DKIM. Plenty were nowhere near the 5,000 messages a day that Microsoft's published bulk sender rule actually covers. On November 10 Microsoft's incident report said a spam enforcement rule update had caused messages to be incorrectly marked and blocked, and by November 11 a spokesperson said the bug was fixed. No postmaster notice went out while it was happening, and the bounce text pointed at SPF when a missing DMARC record was the trigger. For outbound teams the damage is not the week of blocked sends. It is that six days of Microsoft bounce data is now unusable as a reputation signal, and the standard response to a bounce spike, cut volume and rebuild domains, is the wrong move if the spike was never yours. Here is how to read those bounces, what to quarantine, and why the botched rollout still points one direction.TYLER TRUFFI · 11 MINSTRATEGYReverse DNS Is The Cold Email Gate You Do Not ControlSpamhaus updated its Sender Guidelines in October 2026 and says that starting in November 2026 it is gradually rolling out a requirement for every sending IP to have a valid HELO, a reverse DNS (PTR) record and a matching A record. Nothing in that sentence is new advice. What is new is enforcement, and the listing zones behind it are CSS and XBL, which are the ones that take an IP out of the inbox rather than into the spam folder. The awkward part for outbound teams is ownership: SPF, DKIM and DMARC are published in DNS you control, while a PTR record can only be set by whoever controls the IP. Most cold email programmes spent 2026 moving onto private and dedicated sending infrastructure to escape bulk-sender limits, and in doing so handed the one record they cannot publish themselves to a vendor they have never audited. Here is how to find out whose problem it is, by sending setup, and what to fix this week.JOSH BERNSTEIN · 11 MIN
Free consultation

Let us be the last SEO agency you ever work with

A 30 minute call and a free audit of your SEO and GEO position. You keep the findings either way.