whynotroman

News · Nov 03, 2026 · 14 min

What sales blogs wrote, November 3. An agent checks every email before it goes out

SaaStr on an email about 2024 goals and an agent that replaced a vendor in one day, Close on ten AI prompts for sales leaders, Outreach on four questions for an AI agent, HubSpot on permission-first prospecting, Sales Gravy on a Q4 plan. Short facts, my opinion, a link to the article.

I went through what the sales blogs published over the last few days. The main idea this week is easy to see. A person who sends a lot of emails stops reading them, and an agent reads every one. There is also a lot about how a sales leader keeps AI under control and how to close the year. I added what I think about each piece and what a dealership can take from it.

Main in this issue

A seller asked about 2024 goals, and Lemkin wants an agent to check every email

Jason Lemkin, founder of SaaStr, wrote on November 2 about two cold emails from a seller at a large public software company. Both came within 48 hours. The second one talked about "the remainder of 2024" and asked about his "end of 2024 goals". Lemkin says the year was written into a sequence step and nobody checked it through two yearly planning cycles.

His point is this. A person who sends at high volume stops reading what goes out, while an agent can check each record before sending. He also warns that an AI SDR with only a name and a job title writes the same generic email, only 50 times more often. SaaStr gives its own numbers, and the article has no outside check.

Lemkin suggests four changes. An agent reviews every outbound email, including the ones people write. It stops emails with old dates, long product lists or no facts about the prospect. The agent gets the public facts about the prospect.

Instead of a menu, the email has one pain, one number and one ask. If you cannot name the pain, the email waits. People work on large deals and key relationships, not on cold sequence steps.

My opinion. Dealership email templates have the same problem. An email about winter tires goes out in May, and a customer who sold the car gets a service invitation.

I would have AI check every campaign and every salesperson email before it goes out. Does it have the customer's model, the right date and one reason to write? No reason, no email. AI does not replace the salesperson, it shows who did not do the work.

Source SaaStr

A vendor went silent for a day, and a SaaStr agent replaced it the same day

On November 1 Lemkin described how a core product from a vendor he does not name was down for a full day. Every screen and every API call showed an error, and all agent workflows stopped. The vendor did not update its status page, did not email the customer and gave no time for a fix. When asked, it only confirmed that the SaaStr account was fully broken.

SaaStr has an agent called 10K, which it calls its AI VP of Marketing. 10K built a replacement the same day and advised never to use the vendor again. SaaStr moved at least part of the work that day, even though the year was already paid. Lemkin says a team of people would have waited, but an agent does not hold on to money already spent.

His lessons are these. Silence during an outage is worse than the outage. The renewal is decided on the day of the outage, not on the renewal date. SaaStr now looks first at two things in a vendor. How its API works under heavy agent use, and how it talks to customers during a failure.

My opinion. This is a story about buyers who more and more choose through an agent. An agent does not wait and does not forgive silence.

For a dealership the lesson is direct. The website is down, service booking does not work, or a customer's car is stuck in the workshop. I would write to the customer at once, saying what happened and when they will hear back. Silence on a day like that can cost a customer for years.

Source SaaStr

Ten prompts a sales leader can use to have AI review the team's work

Liz Stephany, VP of Customer Experience at Close, shared ten prompts for Chloe Chat, the AI assistant inside the Close CRM, on November 2. This is a vendor article about its own product. The value is in the prompts, and you can ask them of any AI that can read your CRM.

The three most useful ones are these. The first is speed to lead for the last 30 days. It counts the time from a new lead to the first action by a real person, without automatic emails or AI agents. It reports the median and the 90th percentile by rep and by hour, and shows leads with no human contact separately.

The forecast meeting prompt finds deals for this quarter with a past close date or no contact from the rep in 14 days. It also finds deals where the conversation does not match the stage in the CRM.

The other prompts cover a week of a rep's calls and the reasons for wins over 90 days. One compares two weeks of activity with the six-week average and marks changes over 25%. One prepares a one-on-one with deals that had no activity for 10 days and three questions. Another writes handoff notes when a rep is away or leaves. The author advises naming a specific pipeline or specific reps, not "my team".

My opinion. I would take the speed to lead prompt into a dealership today. Leave out auto-replies and count the time to the first answer from a real salesperson, by person and by hour.

Dealer data from different countries shows that 40% of inquiries come outside working hours. This report shows at once how many of those leads wait until morning. I would run the forecast prompt every week before the sales meeting. No record of every contact, no sales team.

Source Close Blog

Before you let an AI agent into the CRM, ask the vendor four questions

Maria Akhter wrote in the Outreach blog on November 2 about why sales AI pilots get stuck in security review. This is a vendor article. Agents now change records, add prospects to sequences and start workflows, sometimes without approval at each step. The author says a vendor can pass a security test and still not know which CRM fields its agent writes to.

The four questions for the vendor are these. Who does the AI act for, and does it have the rights of a named user or role, not a shared account? What customer data does it touch, with a written list of fields, call recordings and mailboxes? What goes into the log, and can you check it? Who controls the agent, and who can switch it off?

The author cites IDC. It expects that by the end of the year 40% of roles at Global 2000 companies will work directly with an AI agent. The author advises starting with lower-risk tasks, such as call summaries and account research. A person should approve every CRM change and every customer email, and "the model decides" counts as an open question.

My opinion. Before connecting any AI to a dealership CRM, I would ask these four questions in writing. The last one matters most. Who switches the agent off, and how fast?

I would start with call summaries and meeting prep, where the agent only reads. The right to write to the database and email customers comes later, with approval from a salesperson. AI prepares, a person decides and is responsible.

Source Outreach Blog

Prospecting in five steps, with every contact logged and every customer request respected

Kiran Shahid wrote in the HubSpot blog on October 29 about prospecting built on the customer's permission. She cites the HubSpot 2026 State of Sales report. In it, 38% of sales people say buyers come to them better informed. Video makes up 42% of sales calls, in-person meetings 33% and phone calls 20%.

The steps are these. Describe your ideal customer and the rules for choosing prospects. Research each prospect using what they shared in public, and save it in the CRM. Rank prospects by fit and signal strength.

Then choose a channel, and do not send the same pitch in every channel. Log every contact and review the results.

On persistence, she suggests six to eight attempts over two to three weeks, then a pause or a longer nurture track. You can come back when there is a new reason. If a prospect says email is fine but asks not to be called, you save that in the CRM. Her sample call opener starts with a reason, not a job title. "I saw your team recently opened three new locations".

My opinion. I would bring the "do not call" rule into a dealership as it is. If a customer says "don't call, text me", that is a CRM field, and every salesperson and every agent sees it.

Six to eight attempts over three weeks is a fair limit for a car lead. After that, pause until a new reason, such as the end of a loan or a new model.

Source HubSpot Sales Blog

A plan for the fourth quarter starts with calls to your biggest clients

On October 29 Sales Gravy published a four-step plan for the end of the year. The author box shows Mike Brooks, founder of Mr. Inside Sales. The article has no numbers or research.

The steps are these. Write on a card how much you will earn by December 31 and how much you need to close for that. Call your biggest clients first, for new business and referrals, and add 10% to the amount you expect from them. Then call your "A" leads, meaning deals that almost closed, people who asked you to call back, and regular buyers. The fourth step is to picture yourself reaching the goal.

For sales leaders, the author suggests meeting each rep, adding their targets into a team target and checking that every large opportunity has an owner. Hold a short weekly check-in on calls and deals. Listen to calls, join key meetings and role-play closing talks.

My opinion. For a dealership this is a week of calls through the database. Who looked at a car and did not buy, who asked for a call back, whose loan or warranty is ending.

I would have AI build this list from the CRM in one evening, and the salespeople would call in the morning. Each salesperson names their own goal for December 31, and the head of sales adds them up and checks them every week.

Source Sales Gravy

Also published

The Salesloft Blog did not open. The list of posts did not load on the blog page or in the feed. The Apollo.io and Pipedrive pages opened but showed no dates, so I could not confirm any post from this period there.

Six blogs had no new posts in these days. They are Gong, LinkedIn Sales Blog, RAIN Group, Harvard Business Review, The Bridge Group and The Sales Blog by Anthony Iannarino. From Salesforce I took only the sales posts, seven more posts in the feed were about other topics.

Next note · 3 minSales blogs, November 17. How to check your team's results →Four reads on marketing targets, deal stages, sales metrics and AI value. The main ideas and practical checks for a sales team.

Roman Spitsyn. I build sales teams where AI agents do the work and people close the deals.

Read next

← All notes