Resource · template

What Should a Good Client Follow-Up Email After a Meeting Include?

Key takeaways

  • The systemPrompt field in a Dave by voolama (hellodave.ai) agent's Config JSON is the primary control for AI behavior in a client follow-up workflow. The agents help documentation (DEV v0.11.0, confirmed 2026-08-30) states temperature controls randomness in responses (0.0 = deterministic, 1.0 = creative). For client follow-up emails, set temperature to 0.3 to produce consistent, professional output with low variation across different meeting types.
  • The Human Review task in Dave by voolama (hellodave.ai) pauses the workflow after the AI produces a draft and creates a task in the Task Inbox for a named reviewer. The reviewer reads the draft in the Payload section, runs the 10-point checklist on this page, and clicks Approve, Reject, or Request Changes. The decision is recorded in the audit log with the reviewer's user ID and a timestamp. Nothing is sent until a person decides. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.
  • The Comment field on the Human Review task is the reviewer's record of which checks they ran. The tasks help documentation (DEV v0.11.0, confirmed 2026-08-30) states: 'Use comments to document your decisions. Even a brief note explaining why you rejected or requested changes helps workflow owners and other reviewers understand the history.' A comment such as 'Checked all action items against meeting notes. Removed one commitment not discussed. Approved.' creates an audit trail for client disputes.
  • Dave by voolama (hellodave.ai) has six default roles: Admin, Create, Curate, Use, Reporting, and Financial. Roles are editable. Only Admin and Use roles can view the Task Inbox and complete Human Review tasks. A staff member who only reviews and approves follow-up email drafts needs the Use role. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.
  • Prompt locking in Dave by voolama (hellodave.ai) prevents any Curate-role team member from overriding the follow-up email template when building a workflow. The agents help documentation (DEV v0.11.0, confirmed 2026-08-30) states: 'When locked, workflow designers cannot override it, the agent always uses its locked prompt.' Navigate to Agents in the sidebar, switch to the Prompt Library tab, click New Prompt, paste the completed template, and save. Then open the agent, select the library prompt, and toggle the Lock Prompt switch.

Summary

A fill-in email template for AI-assisted client follow-up in Dave by voolama (hellodave.ai), plus a 10-point reviewer checklist to run before approving the draft. The template covers what was discussed, what was agreed, who does what by when, and the next meeting. The checklist covers accuracy, tone, commitment language, and confidentiality. Grounded in the Federal Plain Language Guidelines (plainlanguage.gov/guidelines/) and the Dave by voolama tasks and agents help documentation (DEV v0.11.0, confirmed 2026-08-30).

The fill-in template: paste this into your agent's systemPrompt

Copy the text below. Fill in every bracketed field for your business. Paste the completed text as the value of the systemPrompt key in your agent's Config JSON in Dave by voolama (hellodave.ai). The seven-section structure follows the Federal Plain Language Guidelines (plainlanguage.gov/guidelines/): active voice, named responsible parties, and concrete deadlines rather than vague commitments. Sourced from the agents help documentation, DEV v0.11.0, confirmed 2026-08-30.

You are a professional client communications assistant for [BUSINESS NAME], a [TYPE OF BUSINESS, e.g., marketing agency / accounting firm / consulting practice] based in [CITY, STATE].

A staff member has submitted meeting notes from a client meeting. Write a professional follow-up email using only the information in the meeting notes. Do not add, invent, or infer any commitment, agreement, or action item that is not explicitly stated in the meeting notes. If a piece of information is missing (for example, the next meeting date is not in the notes), use [NOT IN NOTES] as a placeholder.

SECTION 1: SUBJECT LINE
Write a subject line in this format: Follow-up: [MEETING TOPIC] - [DATE OF MEETING]

SECTION 2: OPENING
Address the email to the specific client contact named in the meeting notes. Use their first name if the notes include it. Open with one sentence confirming the meeting: 'Thank you for meeting with us on [DATE].'

SECTION 3: WHAT WE DISCUSSED
Write one to three sentences summarizing the main topics covered in the meeting. Use only the topics listed in the meeting notes. Do not add context or background not in the notes.

SECTION 4: WHAT WE AGREED
List each agreement or decision reached in the meeting as a numbered item. For each item, state: (a) what was agreed, (b) who is responsible (name the specific person, not 'we' or 'the team'), and (c) the deadline if one was stated. If no deadline was stated, write [DEADLINE NOT DISCUSSED] rather than inventing one.

SECTION 5: OPEN QUESTIONS
List any questions that need a client answer before the next step can proceed. If there are no open questions, omit this section.

SECTION 6: NEXT MEETING
State the date, time, and format of the next meeting if it was scheduled in the notes. If no next meeting was scheduled, write [NEXT MEETING NOT SCHEDULED] rather than suggesting one.

SECTION 7: SIGN-OFF
Sign off as: [SIGN-OFF NAME OR ROLE, e.g., 'The [BUSINESS NAME] Team' or '[FIRST NAME], [ROLE]']

FORMAT: Return the email as plain text. No markdown. Use numbered lists only in Section 4. Keep the total email under 300 words.

Set the full Config JSON as follows in the Config (JSON) editor at Agents in the sidebar:

{"systemPrompt": "[paste completed template text here]", "model": "gpt-4o", "temperature": 0.3, "maxTokens": 400}

Temperature 0.3 produces consistent, professional output with low variation across different meeting types. The agents help documentation states temperature controls randomness in responses (0.0 = deterministic, 1.0 = creative). Sourced from the agents help documentation, DEV v0.11.0, confirmed 2026-08-30.

10-point reviewer checklist: run this before clicking Approve

Run all 10 checks in the Payload section of the Human Review task before clicking Approve. If any check fails, click Request Changes and add a Comment explaining what needs to be corrected. Navigate to Tasks in the sidebar to open the Task Inbox. The 10 checks below are grounded in the four failure modes the Federal Plain Language Guidelines (plainlanguage.gov/guidelines/) identify as most common in written commitments: vague actions, unnamed responsible parties, missing deadlines, and language stronger than the underlying agreement. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.

Accuracy checks (4 checks)

  1. Every action item in Section 4 was actually discussed in the meeting. Read each numbered item in the AI draft against the meeting notes submitted at the User Interaction task. If the draft includes a commitment not in the notes, click Request Changes and remove it. The Federal Plain Language Guidelines require that written commitments reflect the actual agreement: a commitment that sounds plausible but was not discussed is a false record.
  2. Every deadline in Section 4 was stated in the meeting, not invented. If the draft states a deadline, confirm it appears in the meeting notes. If the notes do not include a deadline for that item, the draft should read [DEADLINE NOT DISCUSSED]. If it does not, click Request Changes.
  3. The email is addressed to the correct client contact. Confirm the name in the opening matches the client contact named in the meeting notes. A draft addressed to 'the team' or 'Hi there' instead of the specific contact named in the notes requires a Request Changes decision.
  4. The meeting date in the subject line and opening is correct. Confirm the date in the subject line and the opening sentence matches the meeting date submitted at the User Interaction task. A transposed date in a client follow-up creates a record dispute.

Commitment language checks (3 checks)

  1. No commitment uses language stronger than what was discussed. The Federal Plain Language Guidelines (plainlanguage.gov/guidelines/) require that written commitments reflect the actual agreement, not a stronger or weaker version of it. A commitment to 'explore options' is not the same as a commitment to 'deliver a proposal.' Read each action item and confirm the language matches the strength of the agreement in the meeting notes.
  2. No commitment is made on behalf of the client without their agreement. The follow-up email records what both parties agreed. If the draft assigns an action item to the client that was not agreed in the meeting, click Request Changes and remove it.
  3. No pricing, fee, or cost figure appears unless it was stated in the meeting notes. An AI draft that invents a price or fee creates a contractual expectation. If any figure appears in the draft, confirm it is in the meeting notes before approving.

Confidentiality and tone checks (3 checks)

  1. No confidential information from a different client appears in the draft. Confirm the draft contains only information from the meeting notes submitted for this specific client. If the draft includes any information that appears to be from a different client or project, click Reject immediately and notify the workflow owner.
  2. The tone matches the business's standard client communication style. Read the draft as the client will read it. If the tone is significantly more formal or informal than the business's normal style, click Request Changes and add a Comment describing the adjustment needed.
  3. The next meeting details in Section 6 are correct, or the placeholder is present. If a next meeting was scheduled, confirm the date, time, and format in the draft match the meeting notes. If no next meeting was scheduled, confirm the draft reads [NEXT MEETING NOT SCHEDULED] rather than a suggested date the AI invented.

After completing all 10 checks, add a Comment to the Human Review task summarizing what you checked. Example: 'Checked all 10 items. Removed one action item not in notes. Corrected client name. Approved.' The tasks help documentation states: 'Use comments to document your decisions. Even a brief note explaining why you rejected or requested changes helps workflow owners and other reviewers understand the history.' Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.

What this template does not cover

This template and checklist cover the AI agent systemPrompt configuration and the Human Review task review process for client follow-up emails in Dave by voolama (hellodave.ai). Four things this resource does not address:

  • Dave by voolama does not send emails. The platform generates the draft and records the approval. Sending the email to the client is a separate step outside the platform, using the business's normal email system. The tasks and agents help documentation does not describe any integration with named email platforms.
  • This template does not cover legally sensitive communications. If a client meeting produced an agreement with legal effect (a contract, a settlement, a regulatory commitment), the follow-up email documenting that agreement should be reviewed by a qualified attorney before it is sent. This template is for routine meeting follow-up, not for legally binding correspondence.
  • The template does not adapt automatically to industry-specific terminology. A hypothetical accounting firm's follow-up email uses different terminology than a hypothetical marketing agency's. The bracketed fields in Sections 1 and 7 are the primary customization points. If the agent consistently produces terminology that does not match the business's industry, revise the systemPrompt's first paragraph to name the specific service type and client relationship.
  • Building the workflow step by step is a separate topic. Connecting nodes, configuring the Human Review node, and setting the workflow to active are covered in the Build Your First Workflow quick-reference. This resource assumes the workflow is already built and the agent is already created.

Frequently asked questions

Where do I paste this template in Dave by voolama?

Into the systemPrompt field inside the Config JSON editor. Navigate to Agents in the sidebar, open your follow-up email agent, click Create Version, and paste the completed template text as the value of the systemPrompt key. The full Config JSON should look like: {"systemPrompt": "[your completed template text]", "model": "gpt-4o", "temperature": 0.3, "maxTokens": 400}. Validate the JSON before clicking Create Version: paste it into a browser console and run JSON.parse(). A SyntaxError means a formatting problem to fix first. Sourced from the agents help documentation, DEV v0.11.0, confirmed 2026-08-30.

What does the reviewer see in the Human Review task for a client follow-up email?

The task detail page in Dave by voolama (hellodave.ai) shows: the task type (human_review) as the page heading, a status badge (pending, amber), three metadata cards (Instance, Assignee, Due), and the Payload section containing the AI-drafted email as JSON. Three decision buttons appear: Approve (green), Reject (red), and Request Changes (outlined). An optional Comment field lets the reviewer note what to change. The decision is recorded in the audit log with the reviewer's user ID and a timestamp. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.

How do I stop a team member from changing the follow-up email template when they build a workflow?

Save the completed template as a Prompt Library entry and lock it to the agent. Navigate to Agents in the sidebar, switch to the Prompt Library tab, click New Prompt, select Category: Content Generation, paste the completed template, and save. Then open the agent, select the library prompt, and toggle the Lock Prompt switch. The agents help documentation (DEV v0.11.0, confirmed 2026-08-30) states: 'When locked, workflow designers cannot override it, the agent always uses its locked prompt.' Only Admin and Create roles can create agents and lock prompts. NIST SP 800-53 Rev 5 AC-6 section 6a (csrc.nist.gov/pubs/sp/800/53/r5/upd1/final) requires least-privilege role assignment.

Which Dave by voolama roles can approve a client follow-up email draft?

Admin and Use. Only Admin and Use roles can view the Task Inbox and submit decisions (Approve, Reject, Request Changes) on Human Review tasks. Dave by voolama has six default roles: Admin, Create, Curate, Use, Reporting, and Financial. Roles are editable. A staff member who only reviews and approves follow-up email drafts needs the Use role. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.

Is there an audit record of who approved each client follow-up email in Dave by voolama?

Yes. Every Human Review task decision is recorded in the audit log with the acting user's email and user ID, the decision (Approve, Reject, or Request Changes), the optional comment, and the timestamp. The audit log is at Admin, then Settings, then Audit Log. Only the Admin role can view it. Task completion events support SOC 2 CC6.1 (logical access controls, aicpa-cima.com) and GDPR Article 30 (processing records) when those standards are active. Sourced from the tasks help documentation, DEV v0.11.0, confirmed 2026-08-30.

Sources

  1. Federal Plain Language Guidelines (plainlanguage.gov): Writing for your reader
  2. NIST SP 800-53 Rev 5: AC-6 section 6a Least Privilege (NIST CSRC)
  3. SOC 2 Trust Services Criteria: CC6.1 Logical Access Controls (AICPA-CIMA)
  4. EU AI Act (Regulation EU 2024/1689): Article 14(4) human oversight requirements (EUR-Lex)

Last reviewed