Ask a new customer what the first week was like, and you often hear the same complaint in different words: "I had to explain everything again." They explained their payroll situation on the sign-up form, to the person who called them, to the salesperson who sold to them and finally to the person who set them up. Each time, a little more was lost.
Most go-to-market teams know this and respond by automating the handoff. The deal closes, a workflow fires, a task appears in onboarding's queue. The transfer is now instant. The context, however, still arrives in pieces, because automation moved the record without anyone deciding what the record must contain.
CloudNala's view is that a handoff is a record with a receiver who accepts it, not a notification. Automate the transfer, the reminders and the routing. Specify the contents, and make the receiver confirm that they have what they need before the sender is done.
This is Part 8 of the series. Part 7 covered the messages a play sends. This part covers what happens when a prospect moves between people inside the company.
Lowveld Ledger is fictional, and its systems and figures are illustrative. It is a Mbombela accounting and payroll software company for SMMEs, with about 2,300 paying customers, a nine-person sales and marketing team, a GTM engineer and a four-person customer success team that includes onboarding.
The promise nobody passed on
The incident that changed Lowveld Ledger's handoffs was small. An account executive closed a 45-employee engineering firm in October. During the final call, the firm's financial manager asked whether their payroll history from their desktop software could be brought across before month-end, because its interim employer reconciliation (EMP501) was due to SARS by 31 October. The AE said yes.
The deal was marked as won. The CRM workflow created an onboarding task with the company name, the plan and the contact's email. Onboarding booked a standard set-up call for the following week and started with the standard questions. The history migration came up only when the customer asked when it would happen. It was not a standard service, it needed a data specialist, and the month-end was nine days away.
Nobody had lied and nobody had been careless by their own standards. The AE assumed the notes on the opportunity would be read. Onboarding worked from the task, which had no notes. The system had moved the deal and dropped the promise.
Four handoffs, four records
Lowveld's journey from first touch to established customer crosses four handoffs. Each one has a sender, a receiver and a record.
Marketing to SDR happens when a lead becomes marketing-qualified. The record is the lead score with its top three reason codes from Part 5, and the source of the lead.
SDR to account executive happens when an SDR has qualified the lead and created an opportunity. The record says why now, who decides, how payroll is run today and what the prospect has already been told.
Account executive to onboarding happens at signature. This is the handoff where the most value is lost, because the sale is where promises are made. The record must list every commitment, the agreed go-live date, the risks and the unknowns.
Onboarding to customer success happens when the customer has run a first live payroll. The record says what was set up, how the first run went and what remains open.
Accountant-partner referrals enter the chain at the AE, with an extra party to keep informed, covered below.
What the record must carry
A record that says only "Company X, Plan Y, closed" is a notification. Lowveld's handoff records have seven fields, adapted for each stage:
| Field | What it holds | Example |
|---|---|---|
| Why now | The event or problem that made them buy | "Interim EMP501 due 31 Oct; desktop payroll licence expiring" |
| Who decides | Names and roles, including who signs off payroll | "Financial manager runs payroll; owner approves" |
| What was promised | Every commitment, in the customer's terms | "History migration before month-end" |
| What is unknown | Questions still open | "Number of employees on weekly pay" |
| Risks | Anything that could derail the next stage | "Tight deadline; migration not a standard service" |
| Source of each fact | Call, form, email, product data | "Final call, 14 Oct, AE" |
| Next step and date | The first thing the receiver must do, and by when | "Data specialist call by 17 Oct" |
The what was promised field is the one that would have prevented the October incident. At Lowveld it is now mandatory on the AE-to-onboarding record, and a deal cannot be marked won while it is empty. "Nothing beyond the standard plan" is an acceptable entry. A blank is not.
The source of each fact field does the same job it does on the account fact sheet in Part 7: it lets the receiver check, and it shows which facts came from the customer and which were assumed.
Accept or reject, never just assign
The second change at Lowveld was to make every handoff a two-sided act. The receiver either accepts the record or rejects it with a reason.
This idea already existed at the first handoff. Part 3 defined a sales-accepted lead as one an AE or SDR has accepted within one working day. Lowveld simply applied the same pattern to every stage. Onboarding now accepts a new customer only when the record is complete enough to plan the set-up. If it is not, onboarding rejects it with a reason, such as "promise unclear" or "no payroll contact named", and the deal goes back to the AE.
Rejections feel like friction, and that is the point. A rejected handoff costs the sender an hour. An unrejected bad handoff costs the customer a week and the company some of the customer's trust. Over time, the rejection reasons also become data: if most rejections at one stage say "decision-maker unknown", the qualification step before it needs to change, not the people receiving it.
Service levels that measure both sides
Lowveld sets two kinds of service level for each handoff. The receiver's level is how quickly the record is accepted or rejected: one working day for every stage. The sender's level is the quality of what was sent: the share of records accepted first time, tracked monthly by sender and by stage.
Measuring only the receiver creates pressure to accept poor records quickly. Measuring only the sender creates pressure to delay sending until everything is perfect. Measuring both creates the conversation the business actually needs, which is about what "complete enough" means at each stage.
The partner referral handoff
Accountant-partner referrals are Lowveld's smallest channel and its best, with the highest win rate of any source. They also had the worst handoff. The partner manager kept referrals in a spreadsheet, passed them to AEs by email and updated partners when someone remembered.
Partners noticed. An accounting practice that refers a client is putting its own reputation on the line, and silence after a referral reads as either failure or ingratitude.
The fix was to treat the partner as a party to the handoff. Each referral now creates a referral record in the CRM, linked to the account and the referring practice. The record carries the partner's notes about the client, which are often the most candid context available. When the AE accepts the opportunity, the partner is told automatically, because the referral record notes that the client agreed to the practice being kept informed. When the client signs, goes live or stalls, the partner manager sees the status and decides what to tell the practice. The spreadsheet has been retired.
What to automate and what to keep with a person
The aim of all this is not to remove automation from handoffs. It is to put automation where it is reliable and people where judgement is needed.
Automation is good at plumbing: creating the record, routing it, prefilling facts already held in the CRM and product data, starting timers and reminders, notifying partners of status changes and logging every change with its source.
AI is useful for one more job: drafting a call summary from notes or a transcript, so the sender does not start the record from a blank page. At Lowveld, an AI summary arrives marked as a draft. The person who was on the call must correct and confirm it before the record can be sent. An unconfirmed AI summary is not a record, because it can confidently omit the one sentence that mattered, and in October that sentence was a promise.
People keep the judgement: confirming what was promised, accepting or rejecting the record, introducing the customer to the next person on a live call, and judging whether the customer is ready for the next stage. The onboarding specialist assigned to the account owns the customer's first week, not the workflow.
Start this week
- Pull your last ten closed deals and compare what the customer was told during the sale with what onboarding knew on day one. Count the promises that did not arrive.
- Add a mandatory "what was promised" field to the sale-to-onboarding record, and block "won" until it is filled in.
- Give every receiver an accept-or-reject step with a short list of rejection reasons, and review the reasons monthly with the senders.
Next: Part 9, Measure learning, not just activity. All parts are on the series page.
How CloudNala can help
CloudNala maps the handoffs in your customer journey, from first contact to an established customer, and designs the records, acceptance steps and service levels that keep context intact between teams. We configure the automation that moves the records and the AI-assisted summaries that help people write them, with confirmation steps so that a draft never becomes the record by default.
Work with CloudNala
CloudNala helps organisations move from technology ambition to practical execution across cloud, AI, data, platform engineering and digital services.
Whether you are exploring AI, modernising your cloud environment, building a public-sector digital service, or turning an idea into a working MVP, we can help you shape the roadmap and deliver the next step.
Book a Go-to-Market Systems Review or write to us at consult@cloudnala.co.za