The Only Sales Role That Will Exist in 2030 w/ Christopher O'Donnell (Founder, Day AI)

He traces sales specialization to Aaron Ross and Predictable Revenue at Salesforce, and says many companies over-specialized last decade because it was in vogue rather than right for their context. He argues that handoffs are inefficient for both customer and organization, and that AI lets one person handle all the roles. Consolidating the roles does not mean getting rid of the people.

“So I've been talking and inspiring a lot of companies to get rid of the concept of an sdr, an AM and a cs. It doesn't mean you're getting rid of those people. It just means we're going to a single role.”

Listen to the episode Sales team, hiring & comp Link to this

He says 2.5 PRs per week held at HubSpot and at Stripe, and that about 80% of engineers are still at that baseline despite using Cursor and Claude Code. In a board meeting he would ask for last quarter's pull requests per week, looking for 5 (a 2x) up to 20. He adds that faking AI usage is more work than actually using it.

“What was your pull request per week in the last quarter? And I'm looking for five, which would be a 2x up to 20”

Listen to the episode Metrics & finance Link to this

Selling time, the share of the week a rep spends in front of buyers or customers, goes from 25% to 75% once admin, CRM work, prep and pipeline reviews are removed. The rep-to-manager ratio goes from about 7:1 to 14:1. Reps who have averaged $250K a quarter would average $500K a quarter.

“Number one, selling time goes from 25% to 75%.”

Listen to the episode Sales team, hiring & comp Link to this

Christopher argues that humans no longer judge a tool by its interface or feature set, because the AI is the one using it. He suggests connecting a set of tools to Claude, asking what it could use those MCP servers for, and having it try them and report back. Some excellent products 'won't really give Claude what Claude really wants.'

“You can connect a bunch of different tools to something like Claude and say, what do you think of all of these? Like, what could I use these MCP servers for? Try some of it.”

Listen to the episode AI Link to this

Christopher describes using the full customer history to find which customers closed fastest, what they said they came looking for, when the solution clicked for them, what words the seller used and how the prospect reacted. The AI then finds lookalikes. Outreach draws on real customer language and feelings rather than 'a bunch of us sitting in a room coming up with marketing slogans.'

“Okay, so imagine taking the entire history of everything and just extracting from that history just those bits. Who are these people? What are the patterns? And then being able to say, get me 100 more of these.”

Listen to the episode Pipeline & demand generation Link to this

He calls it a hypothesis that one of the strongest exploitable advantages is how you apply AI to your team and organization. An incumbent with 1,000 go-to-market people built on SDR/AE/CSM specs has few people who fit a full-cycle, AI-enabled profile and must retrain or replace them. A startup can hire into that profile from day one.

“As a startup I can hire into that profile right from the get go. That is a massive innovator's dilemma that product founders can exploit.”

Listen to the episode Strategy & market Link to this

Christopher says that in 2011-2012 he was PM, support expert, user researcher, go-to-market lead, front-end developer and designer on his products at HubSpot. By 2016 those were separate teams, and they kept splitting into roles like UX copy. Teams meant to be three people became seven, eight or nine, 'like a Quaker meeting', and lost effectiveness. Now PMs are doing sophisticated coding.

“And these small autonomous teams that we wanted to have all of a sudden were seven, eight, nine people.”

Listen to the episode Product Link to this

Christopher says that around six months earlier he realized PMs, especially but not only at early stage, need much more detail upfront, which he did not see coming. He had always worked by shipping small and iterating with A/B tests and gating. Now every error case and all the polish can go into the spec, which becomes like 'an iron on decal' applied to the codebase.

“And the cost of writing the code from the right spec is asymptotically approaching zero.”

Listen to the episode Product Link to this

On his first Monday the marketer installed cmux (a terminal setup for Claude Code that even engineers find extreme), authenticated Claude Code, and learned agents, skills and MCP while building his rig. By Thursday afternoon he was deploying to production on Day AI's commercial web app. He had never written code and still had not opened a code editor.

“By Thursday afternoon, he was deploying to production crazy meaning on our commercial web app.”

Listen to the episode Hiring & team building Link to this

Christopher describes it as part post-sales CSM, part sales engineering. Agent engineers are personable, have business acumen and carry quota, but spend most of their time using the customer memory layer across Day AI's roughly 500 customers. They find pockets of opportunity, such as getting marketing departments into the product. They pursue these autonomously, for example building a new notification, and report A/B test results in Slack.

“we have this agent engineer role that is coming from a sales background, super comfortable being on the phone, super personable, polished business acumen, quota, et cetera. But 80% of their time is in Claude code or in day AI”

Listen to the episode Sales team, hiring & comp Link to this

He dismisses headlines claiming AI makes people 15% more or 15% less productive. He says it is 'safe to say' that people still working in three to five years will produce 2, 5 or 10x the work product. He expects this to reach sales, beyond engineering where most progress has been so far.

“Anybody who is still working three years a year, five years from now, I think it's safe to say that they will be producing 2 to 5 to 10x as much work product”

Listen to the episode AI Link to this

Christopher says everyone should write job descriptions as if hiring a virtual worker, especially if they have no headcount and have never written one. For sellers, the job worth keeping is being on camera or in person with prospects and customers, and listening.

“you should still take your current role and cut it up into 1, 3, 5 job descriptions of work that you don't want to be doing and focus on your new job description, which is work you do want to be doing.”

Listen to the episode AI Link to this

Christopher says that with AI note-takers the seller can pay full attention, build personal rapport and go deep in discovery. If something is missed it will 'pop up and remind us', and if another channel has already answered a question it doesn't need to be asked again. The seller can focus on what is still unknown about how this prospect could succeed.

“We can get very deep into discovery because we know that if we miss something in discovery, it's gonna pop up and remind us.”

Listen to the episode Sales process & deals Link to this

Christopher says renewals, training, selling and discovery are different specializations, which is why customers deal with many people. AI onboarding and implementation coaching do not remove the need for someone who owns the customer relationship. He suggests, tentatively, that it may fold sales into account management.

“Why do you have six touch points with a given vendor? It's not because the customer wants it, the business model requires it.”

Listen to the episode Sales team, hiring & comp Link to this

He recalls a HubSpot rep who sold to sign shops for close to 20 years and 'sold like every sign shop in America'. He speculates that broader seller roles could own one slice deeply, with AI helping produce battle cards, white papers, testimonials and solutions-page content. Today that knowledge tends to disappear into a 'black hole' because sellers don't share it with customer marketing.

“we wonder if sellers will become masters of their own domain.”

Listen to the episode Sales team, hiring & comp Link to this

He notes that around 1995 there was one role, the salesperson, and that over 20 years it split into SDR, AE, AM, CSM, support and RevOps. Specialization matches talent to role but creates local-maximum optimization. He says AI will 'probably' bring back the generalist athlete, and that in a post-AI world the full-cycle salesperson will be the optimal design.

“Specialization is great. It allows you to align the talent with a role and use the hardest to find skills in that most important part of the funnel. But it comes with a cost, a cost of local maximum optimization. In a post AI world, the full cycle salesperson will be the optimal design.”

Listen to the episode Sales team, hiring & comp Link to this

Christopher describes skills as small training manuals or slash commands that let the AI do something perfectly every time, and calls them settled. Agents have job descriptions, a view of the world and opinions about what matters, and you need more than one. How skills get automated (loops, schedules, triggers) is still undecided and will have many vendors. The fourth layer is the data source, which he says should be the actual customer truth down to the word said, who said it and when.

“I think there are a couple parts of it that are really decided and that is the idea of an agent and the idea of skills.”

Listen to the episode AI Link to this

He describes two groups: people who are 2-5x or more productive, and a large group still at baseline. The 15% figure would describe someone using autocomplete while still coding by hand, which he says is not what anyone is really doing.

“My take on all that noise is it's going to be pretty bimodal already.”

Listen to the episode AI Link to this

Coaching is one of the most time-consuming parts of a sales manager's job. He argues AI can review 70 calls rather than two, and can learn from past coaching sessions which kind of coaching each rep responds to best.

“But today's AI is a better coach than humans. They can look across 70 calls, not just two.”

Listen to the episode Sales team, hiring & comp Link to this

Christopher says CRM keeps its role for opportunities and comp, but agents need call, Slack and email data from across the company, including the CEO's inbox. That quickly requires controls so people understand what is shared and visible to others. A customer-facing agent also has to be prevented from doing things like deleting meetings from someone's calendar just because it was asked to.

“if you have a customer facing agent, boy, you better make sure that they can't delete all the meetings from Mark's calendar just by asking it to.”

Listen to the episode AI Link to this

Christopher says a single-user CRM needs 45 minutes, not a weekend; once permissions over who sees whose leads come in, people buy HubSpot or Salesforce. Customer memory is a superset of CRM and gets 'really intense really quickly': it has to populate almost instantly and be built for the AI. Teams building it themselves reach for vector databases and property graphs rather than relational databases, and he thinks they will ultimately decide to use a vendor.

“if you're just trying to build yourself a CRM, you don't need a weekend. You need 45 minutes. The minute you go multiplayer.”

Listen to the episode Product Link to this

He created user researcher, data analyst and marketing copywriter agents, then a VP agent to manage them, which immediately began asking for more data the way a real VP would. He then built a CEO agent. He still says he can't hand over company strategy or rely on AI to inspire employees and answer their questions. For sellers, he says the human role is the human connection, trust and credibility.

“I can't count on the AI to inspire the employees and to answer their questions and to meet them in their train of thought”

Listen to the episode Leadership & culture Link to this

Christopher says the AI SDR emails he gets mostly just pull from LinkedIn; one confused his Winchester, Mass. location with a cathedral. A caller from a well-known company had shared investors, past demos and a lost deal with him to work with, yet sent generic outreach. He says the models are very good at writing and personal relationships, but they lack the starting material and can't find patterns in everything the company has ever said to anyone.

“If you give them the right starting information, they will come up with something to say.”

Listen to the episode Pipeline & demand generation Link to this

Speaking as a buyer who enjoys buying software, he says that as he moves from BDR to AE to renewal manager, his vendors have no idea who he is. In his view, everyone at a vendor should now know every detail that matters about the customer.

“it's like the goldfish with the 5 second memory. They just have absolutely no idea who the hell I am. That is inexcusable.”

Listen to the episode Retention & customer success Link to this

He compares it to Open Evidence, which instantly pulls the exact relevant data for doctors, and says every domain needs instant recall of the most relevant facts. Agents, skills and eventually automation can come from any vendor (Claude, ChatGPT, Cursor, Perplexity), but memory is the hard part. CRM data is valuable if the team keeps it clean, but it is 'not enough'. He cites Jack Dorsey's piece on the world model of a business.

“So the memory layer, no matter what you're doing, becomes the single most important thing you can get agents, you can get skills, you can get eventually automation from any vendor.”

Listen to the episode AI Link to this

Asked for one tactic for young engineering founders building in a vacuum, Christopher (acknowledging he is promoting Day AI) suggests recording meetings and connecting Slack into a customer memory layer. From there feedback can flow automatically to GitHub, or be worked through Claude Code. He says the expectation should be past tickets: automated action, rolled-up strategic insights, and answers to questions like who a feature would affect or what the biggest gaps are for a new persona.

“I think we're definitely past the kind of ticket concept, you know, level and you should be able to have some sort of automated action off of that”

Listen to the episode Product Link to this

He says it is 'probably an AE', especially one who is very tech-enabled. He also says he could see himself promoting an SDR and still has work to do on what the optimal go-to-market hire is.

“I will say on the go to market side, if I had to promote like an sdr, an AE or CSM or whatever, it's probably an AE that's the best.”

Listen to the episode Sales team, hiring & comp Link to this

He frames the startup-versus-incumbent contest as fair: startups must get distribution before incumbents innovate. To make it less fair, he would let people in any department write code. His answer to the nervousness is process: have Claude triple-check the code, put it in front of a human, and create a release engineering role for ideas that come from other departments.

“Why not have people in whatever department able to write code, you know. Well, we're nervous about. We're nervous about what? Deal with your nervousness. Put the right process in place.”

Listen to the episode Hiring & team building Link to this

Alongside the rising emphasis on taste, he says design is converging back to a set of clean base defaults. The goal is doing the thing people expect most, not inventing something novel.

“It's not about doing something new. It's about doing the thing that people really expect the most.”

Listen to the episode Product Link to this

Christopher says there is no need for an analyst to quantify the problem or a designer to mock up options. Someone on the phones and someone in product sit together over coffee, find something, verify its impact and ship it. In his office, four or five people may talk for three hours around couches over data and customer feedback, and by the end the software is written.

“You need somebody on the phone and somebody in product to sit together over a coffee and get a lead on something you could do, verify that it's going to have this impact and ship it 45 minutes later.”

Listen to the episode Product Link to this

He says there are many skeptics and you have to find people who are genuinely, deeply curious. This applies across the company, not only to engineering; everyone at Day AI uses Claude Code for most of their work. He says this holds whatever your business is, not just because Day AI sells an AI product.

“The people who think that stuff is unbelievably cool and are already playing with it nights and weekends and then figure it out from there.”

Listen to the episode Hiring & team building Link to this

He argues the functional divides exist because of human limits; few people study both finance and code. Those divides create inefficiencies, and he says finance being closer to sales, product closer to support and HR closer to product would help. He offers this as a 'perhaps' for a later phase, after a first phase in which AI mainly raises selling time.

“We would benefit from finance being closer to sales, product being closer to support, HR being closer to product. Perhaps over time those functions will blur in the optimal organizational design.”

Listen to the episode Strategy & market Link to this

He agrees with Christopher that taste, which he calls creativity, will remain human, citing songwriters, film, books, art and sports. He has recently developed conviction about human-in-the-loop roles, comparing them to pilots who remain even though planes can fly themselves. On a million-dollar software purchase, he wants to shake a human's hand to confirm his needs are understood and will be delivered.

“And if I buy a million dollars of software from someone, I'd like to shake a human's hand to make sure that they have blessed my needs and the fact that they will actually deliver.”

Listen to the episode AI Link to this