The setting
supplies cloud software that lets developers and businesses build messaging, voice, and email into their applications. Its gathers these tools, along with identity checks and customer data.
A polite conversation ending does not establish that the customer received help.
Selective follow-up protects service continuity, while restraint protects customers who have nothing left to resolve.
Who you’d be doing this for
“Yesterday our agent said goodbye, then the customer opened another chat asking the exact same question.”
Fatemeh Aydin · Application Engineer at an online retailer · a character in this brief
Builds the service agent and the workflow that offers customers additional assistance.
What is at stake
In a sample of 2,000 service conversations, 18% of customers returned within 48 hours, but only some still needed help. You must weigh a useful signal of unfinished service against unnecessary follow-ups and labels that confuse a new question with an old problem.
Why it isn’t already fixed
Every obvious fix costs something else. That’s the part you’d have to decide.
- missed unresolved needs vs unnecessary follow-ups
- predictive coverage vs explicit evidence
- customer-specific patterns vs generalization
- timely predictions vs isolation from live retrieval
Why Twilio
Structured conversation memory supplies context across voice and messaging, but a return visit alone does not establish that the earlier interaction failed.
Written with these in mind
Not your kind of problem? 41 more at Twilio, or browse every organization.
Where the organization context comes fromThe postings this brief was written from, and the reporting read before it.
The organization context comes from public information. The situation and characters are written for the brief.
Job postings it was written from
Public descriptions of the work at Twilio. They are listed as sources, not as openings. The lines under a posting are its own words, where it meets this brief.
- Staff Machine Learning Engineer (opens in a new tab)
- This position is needed to build the foundational AI/ML capabilities and common platform tooling that power intelligent engagement experiences across Twilio.
- You will influence the design and evolution of our company-wide ML/AI Ops foundation, ensuring every AI capability incorporates rigorous offline and online evals to maintain exceptionally high quality.
- Machine Learning Engineer 2 (opens in a new tab)
- Build the extraction pipeline that turns raw data into structured memory, tuning the LLM prompts and schemas that decide what an AI agent remembers about a customer and what it forgets.
- This position is needed to build and scale the pipelines that turn raw conversation data into structured, searchable customer memory, and the evaluation systems that keep that memory accurate as the product evolves.
This is the setup. The work is inside.
Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.
Where similar work appears
Other live briefs connected to this one, and what connects them.
Similar problems at other organizations
Like this one, each of these turns on deciding on evidence that's still thin.
Seven-day repeat contact reaches 24% in SMS long-horizon cases despite similar conversation-end rates.
You must weigh a useful early-risk signal against incomplete links between channels and actual unresolved needs.
Support leaders cannot tell which resolved Messenger conversations lead customers to return with the same issue.
You must weigh a useful quality signal against claims the available evidence cannot support.
Routine order-status conversations still consume human capacity despite approved answers being available.
You have to weigh broader autonomous resolution against the confidence needed to keep replies within approved language.
The same setting, from another role
Each is set around Twilio’s Twilio Customer Engagement Platform, written for a different role. They’re separate briefs, not the same scenario.
Sellers found conflicting buyer assumptions in eight of twelve Messaging deals marked ready for proposals.
You decide how a draft handles a country or volume figure when an Outreach email differs from the Salesforce record, and which differences go to the seller first.
Five software providers choose a supplier in eight weeks, but disagree on who should bill their customers.
You weigh partner incentives and customer ownership against the margin and support obligations each agreement creates.
A software partner’s $9 million Messaging renewal expires tomorrow amid disputed forecasts.
You must weigh retaining its full commitment against signing away the ability to recover rising carrier costs.
Other problems with a similar purpose
In each of these, as in this one, the aim is using evidence and inquiry to uncover the true problem worth solving before committing craft or budget to an answer.
Evaluation customers can spot a journey drop-off but cannot connect it to a responsible technical surface.
You have to weigh a useful cross-product investigation against access, cost, and an understandable first workflow.
Higher-value home carts are leaving after shoppers see that Pay in 4 will not cover enough of the purchase.
You have to weigh incremental checkout completion against early repayment performance and merchant economics.
Hotel-search feedback points to several possible failures, so recurring traveler friction is not reaching one shared problem statement.
You must decide what AI groups, what PMs verify, and what evidence is strong enough to guide the next experiment.
