Recommendation · MoneyIn a low-stakes internal chat agent environment, start with tracing and monitors for user frustration and go to production directly without heavy offline evals. For customer-facing or regulated use cases, build ground truth datasets before production instead.
AI Engineer · Evals for High-Stakes AI Agents — Doneyli De Jesus & Annabell Schäfer, Langfuse
Recommendation · MoneyDoneyli De Jesus advises asking domain experts for unfiltered raw commentary on traces rather than asking them to invent failure categories themselves, which overcomplicates annotation. Suggested framing: 'Give me your unfiltered raw commentary on these traces.'
AI Engineer · Evals for High-Stakes AI Agents — Doneyli De Jesus & Annabell Schäfer, Langfuse
Recommendation · MoneyDoneyli De Jesus advises that when running an LLM judge on production data, starting too strict means missing real failures. Instead, start with a few too many false positives and tune down over time, so you catch a little more than necessary but don't miss anything.
AI Engineer · Evals for High-Stakes AI Agents — Doneyli De Jesus & Annabell Schäfer, Langfuse
Recommendation · MoneyDoneyli recommends taking a sample of 10% or 20% of activity and filtering to focus on specific traces — for example, ones where the customer talks about disputes or credit card information — to create a manageable, representative review set.
AI Engineer · Evals for High-Stakes AI Agents — Doneyli De Jesus & Annabell Schäfer, Langfuse
Recommendation · MoneyImpersonate a few different types of clients — e.g., everyday customers vs ultra high net worth individuals — to test whether the agent respects guardrails for those personas.
AI Engineer · Evals for High-Stakes AI Agents — Doneyli De Jesus & Annabell Schäfer, Langfuse
Recommendation · MoneyKormos recommends injecting the DAG state into your agent so it knows where it fits into the bigger picture.
AI Engineer · How to Know When an AI Agent Can Work Unsupervised — Evan Kormos, Coinbase
Recommendation · MoneyKade Hinkle advises that if a prospect you pitched posts on LinkedIn afterward, comment on that post — relevant comments make them check your profile and often respond to your DM. Alex Murphy shares a similar follow-up method: when a pitch is ignored, he likes and comments on the prospect's post to put his name back in front of them. Alex also checks the prospect's page first when a LinkedIn touch is due in a multi-channel sequence, and if they just posted, uses that post as the follow-up instead of re-pinging an ignored DM.
30 Minutes to President's Club | No-Nonsense Sales · #615: How to Master Social Selling on LinkedIn to Book Meetings | Kade Hinkle | 30MPC Hall of Fame
Recommendation · MoneyKade Hinkle recommends including two specific, useful signals in your outreach (e.g., titles complaining about competitors, new job changers asking their community what tools to adopt) so the prospect gains value even if they never respond. When pitching, he always gives the prospect two concrete ideas for how they could use his tool — leading with usable value rather than features.
30 Minutes to President's Club | No-Nonsense Sales · #615: How to Master Social Selling on LinkedIn to Book Meetings | Kade Hinkle | 30MPC Hall of Fame
Recommendation · MoneyAlex describes citing specific hiring signals in a video message (e.g., 'you're hiring 12 new reps, added four last quarter, had a fundraise') to hypothesize the prospect's growth-pressure problem, and naming the common struggle the prospect likely faces (e.g., 'most teams are struggling to get ramp time where it needs to be') to build relevance. Kade Hinkle uses hiring activity and recent fundraising as trigger signals to hypothesize a prospect's growth-pressure problem in cold outreach, pairing high growth targets with struggling new-hire ramp times.
30 Minutes to President's Club | No-Nonsense Sales · #615: How to Master Social Selling on LinkedIn to Book Meetings | Kade Hinkle | 30MPC Hall of Fame
Recommendation · MoneyIncluding a video of yourself walking through the proposal makes it easy for stakeholders beyond your champion to digest.
30 Minutes to President's Club | No-Nonsense Sales · #617: Three Steps to Use LinkedIn to Find More Prospects | Charlotte Johnson | 30MPC Hall of Fame
Recommendation · MoneyThe host recommends connecting with every C-suite or VP persona that follows your company, even if you won't prospect them, to build mutual connections that raise acceptance rates on future requests.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host recommends filtering prospects by 'following your company' because followers took a deliberate step toward your brand and are warmer than cold prospects.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host recommends saving Boolean searches in Sales Navigator (e.g., a 'function' filter) and renaming them (e.g., 'sales following your company') so new people matching the criteria automatically appear in the saved search.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host recommends filtering by 'years in current company' less than one year, since these people still want to make an impact and are looking at new products — and to save this filter for every enterprise account on the target list.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host discourages filtering by 'changed jobs in past 90 days' because new joiners are flooded with outreach and haven't finished onboarding — everyone else is probably reaching out to them too.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host discourages leading your LinkedIn headline with just 'SDR' or 'AE at company' because it lowers connection acceptance and response rates.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host's favorite discovery question is the multiple-choice format: 'Typically when I'm talking with folks like you, they're focused on A, B, or C — which is it for you?' — where A, B, and C must be problems you can solve, steering the conversation toward where you can help.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host recommends a three-part LinkedIn headline formula: a value framework (what you help people do), companies you've helped, and a call to action — noting a good headline increases connection rates.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
Recommendation · MoneyThe host recommends asking past customers 'who are two or three people like you that I should be connected with?' — especially when they change jobs — and limiting referral requests to a maximum of three people per message.
30 Minutes to President's Club | No-Nonsense Sales · #616: How I Use LinkedIn Sales Navigator to Find Warm Leads (Crash Course)
RecommendationPamela Fox recommends that when using RAG, you force the model to prove every piece of information in its answer came from the provided sources, associating each claim with a source. She notes this can make the model realize when information was not actually in the sources, though a model can know information is not source-provided and still give it anyway.
AI Engineer · The Model Swap Workshop: GPT-5.5 vs Kimi vs DeepSeek vs Mistral — Pamela Fox, Microsoft