Startup 2 Standup 
5 Takeaways
- AI can take repetitive prospecting work away from salespeople, but the human conversation still matters.
- AI will not fix an unclear ideal customer profile or weak value proposition.
- Start with one narrow AI task and keep a human checking the output.
- AI-assisted outbound needs testing and refinement rather than a set-and-forget approach.
- The right level of automation depends on the customer, deal and sales process.
Summary
During our SU2SU session, St John Dalgleish, Founder and CEO of Perlon AI, explored how AI agents are changing outbound sales and go-to-market.
His message was not that AI should replace salespeople. AI can handle more research, prospecting, personalisation and follow-up, leaving people with more time for calls, meetings, relationships and closing. But founders still need to understand who they are targeting and why customers should care before asking AI to scale the process.
Introduction
During the session, St John described AI moving beyond email writing into systems that can coordinate research, messaging and follow-up.
He described outbound moving from field and phone selling, through CRM-led sales and early automation, into what he called the “agentic” era.
But the discussion kept returning to a more useful founder question:
If AI can do more of the work around selling, where should people still be involved?
Takeaway 1: Outbound has moved beyond basic automation
St John explained that older personalisation often meant inserting a first name or company name into an otherwise standard sequence.
Buyers have become used to that.
He described the newer opportunity as using AI to handle more of the research and create genuinely personalised outreach at scale.
Are you improving the relevance of your outreach, or simply automating more of it?
Takeaway 2: Automate the repetitive work, not the relationship
St John saw AI supporting prospect research, contact sourcing, personalised messaging and follow-ups.
That gives salespeople more time for the phone, face-to-face meetings, relationship-building and closing.
One founder in the room shared his own recent experience of outbound calling into Silicon Valley and described the phone as a useful “pattern interrupt”. That was his experience, rather than St John’s example.
Efficiency should create more room for conversation, not less.
Takeaway 3: Get your go-to-market thinking clear before automating it
This was one of the clearest warnings from the session.
St John said founders need to understand their value proposition and ideal customer profile before going to an AI agent.
He said Perlon had previously worked with very early-stage startups with no customers that
expected an agent to “fix” the business.
His point was simple: the system cannot work out fundamentals the founder has not
established.
Takeaway 4: Start with one narrow job
St John suggested beginning with one agent and one outcome.
At the start, a human should remain involved to check the messaging, content and brand alignment.
His recommendation was to begin with a tightly defined use case rather than handing too much to the system at once.
What is one repetitive GTM task you could test first?
Takeaway 5: Treat the agent like a junior employee
St John encouraged founders to treat outbound as an experiment, with metrics, testing and iteration.
His analogy was that an AI agent is closer to a junior employee: it needs training, attention and tweaking before it can perform consistently.
He shared one example from Perlon’s experience where a videography business took around five or six months of refinement before, according to St John, closing a roughly £500,000 deal with Google DeepMind through the platform.
The point was not the size of the deal. It was how much refinement came first.
Takeaway 6: AI may change sales roles as well as sales tools
St John said salespeople are becoming more “full cycle” because more outbound work can now be supported by automation.
He shared a Perlon client example where a PE-backed B2B software company moved from five SDRs and ten account executives to one SDR and ten account executives while retaining the same quota.
He was also clear that removing SDR teams is not the objective for every business.
The better question is: If AI removes repetitive work, where can your people add more value?
Takeaway 7: The right automation depends on the sale
The founder questions after St John’s presentation showed why there is no single model.
One founder asked whether a lower-ticket product could move from outbound email directly into purchase without salesperson involvement. St John explained that Perlon generally works with more sales-led, higher-ticket outreach, although other approaches can work when the targeting and message are right.
Another founder asked about finding buyers for corporate retreats. St John discussed people teams and, for smaller businesses, direct outreach to senior decision-makers based on Perlon’s experience.
The technology may scale. The sales motion still needs to fit.
Where does AI help, and where do people still matter?
| Sales activity | What we heard in the session |
| Prospect research | Strong opportunity for AI support |
| Personalised email | AI-supported, with good inputs |
| Follow-ups and admin | AI-supported |
| Phone conversations | Human-led |
| Face-to-face meetings | Human-led |
| Relationship-building and closing | Human-led |
St John’s closing view was clear: email, LinkedIn, the phone and physical events are not going away. AI can support more of the process around them, but sales remains a people’s game.
How can founders continue the conversation?
This is why we bring specialists into the SU2SU room.
Our role is not to replace people such as St John. It is to bring founders and specialists together, create space for real questions and help those conversations continue.
Founders can bring live challenges into our Founder Surgery sessions, explore previous guest-speaker sessions, watch recorded discussions on the Startup 2 Standup YouTube channel.
Members also have access to the SU2SU Knowledge Bank and wider founder community.
Founders who want to continue those conversations can explore Become a Member, while businesses interested in supporting the founder community, guest-speaker activity and interactive workshops can explore our sponsorship opportunities.
Conclusion
The biggest takeaway from St John’s session was not that AI makes people less important.
It changes where their time can be most valuable.
Founders still need a clear ideal customer profile and value proposition before expecting an AI agent to improve outbound. Start narrow. Keep humans involved. Test what works. And leave the parts of sales built on trust and relationships to people.
FAQs
Can AI replace a startup’s outbound sales team?
St John did not frame AI that way. His focus was on augmentation: using AI for more of the repetitive outbound work while people remain central to calls, relationships and closing.
What should founders know before using an AI sales agent?
St John said founders should understand their ideal customer profile and value proposition first. The agent cannot reliably establish those fundamentals for the business.
Where should founders start with AI in outbound?
St John suggested one narrow job and one outcome, with a human reviewing the work initially.
Does AI-powered outbound work immediately?
Not necessarily. St John encouraged founders to treat outbound as an experiment that needs testing, measurement and refinement.
Which parts of sales still need people?
St John placed particular importance on calls, face-to-face meetings, relationship-building and closing.
If this is a question you are working through now, bring it into a SU2SU Founder Surgery, explore our expert sessions or join the Startup 2 Standup community to continue the conversation with founders and specialists who understand the challenge.
Recent Comments