Sales & leads
Proposals drafted from the discovery call
The call recording becomes a scoped proposal with phases, pricing from your rate card and wording borrowed from similar work you have won.
Starts when
A discovery call ends.
What happens
- Summarize the client's goals, constraints, budget signals and timeline
- Find your most similar past proposals and their outcomes
- Draft the scope, phases, assumptions and exclusions
- Price each phase from your rate card and typical hours
In more detail
An automation like this reads the transcript of a discovery call, the firm's past proposals and statements of work, its rate card and a record of how many hours similar phases really took. It drafts a proposal in the firm's own template: the client's goals and constraints, phases, deliverables, assumptions, exclusions and a price per phase. Prices are calculated by ordinary code from the rate card, not guessed by the AI, and the draft shows which past proposals it drew on. Nothing is sent to the prospect. The account lead edits the scope and price and sends it from the usual proposal tool. To work well the firm needs recorded calls with consent, a folder of past proposals, a current rate card and some history of actual hours. Fenrir Labs could design it to run on the firm's own machine so call transcripts and past client work stay in-house.
You stay in control
The account lead edits the scope and price and sends the proposal.
Connects to
- Meeting recorder
- Proposal library
- Rate card
- CRM
Questions we would ask you
- Which meeting recorder do you use for discovery calls, and can you export transcripts as text?
- Where do past proposals live, and how many are marked won or lost?
- Do you price by phase from hours, by fixed fee, or by value? Who sets the rate card?
- Do you have actual hours by phase for past projects, and in which tool?
- Which proposal tool or template do you send from, and who signs off on price?
Discuss this further with the Live Agent
Operations & inventory
Meetings turned into tasks and a client recap
Within minutes of a meeting, decisions are recorded, tasks are assigned in your project tool and the client has a clear recap of who is doing what.
Starts when
A client or internal meeting ends.
What happens
- Summarize decisions, open questions and commitments
- Create tasks with owners and due dates in the project
- Draft a recap email for the client
- Raise anything that sounds like new scope
In more detail
An automation like this reads the transcript of a client or internal meeting, either from the firm's meeting recorder or from its own local transcription. It drafts a short list of decisions and open questions, proposed tasks with an owner and due date, and a recap email for the client. Anything that sounds like new work is flagged for the project lead instead of being added quietly. The project lead reviews tasks and recap together, edits them, approves the tasks into the project tool and sends the recap from their own mailbox. Nothing is sent automatically. Many meeting recorders already do much of this, so it makes most sense for firms that cannot send client conversations to cloud services, or that want the meeting output to feed other checks such as scope and budget. It needs recorded meetings with consent and one project tool used consistently.
You stay in control
The project lead reviews the recap before it goes to the client.
Connects to
- Meeting recorder
- Project management
- Email
Questions we would ask you
- Which meeting platforms and recorders do you use, and do clients ever object to recording bots?
- Which project tool holds tasks, and is every client project set up there?
- Who should approve tasks and recaps: the project lead, or the account lead?
- How do you want unclear owners and due dates handled?
- Do any clients or contracts forbid cloud transcription?
Discuss this further with the Live Agent
Finance & billing
Budget burn alerts before the overrun
Hours are compared with the budget for each project every day, and the project lead hears about a likely overrun while there is still time to act.
Starts when
Daily, after time entries are in.
What happens
- Remind people whose time is missing
- Compare hours and costs with the budget and percent complete
- Project the finish at the current pace
- Alert the lead with the cause and the options
In more detail
An automation like this checks every active project each morning. It reads yesterday's time entries from the time tracker, the budget for each project or phase, and how far along the work is from the project tool. Ordinary code compares hours and cost with the budget, reminds people whose time is missing, and projects where the project will finish at the current pace. When a project is heading over budget, the AI writes the project lead a short note naming the cause, for example a phase or a person driving the hours, and the usual options. The project lead decides whether to re-plan, absorb the cost or raise a change request; nothing changes automatically. It needs budgets entered per project and time logged regularly. Tools like Harvest and Teamwork already have budget alerts, so this is most useful when the numbers live in several systems.
You stay in control
The project lead decides whether to re-plan, absorb or raise a change request.
Connects to
- Time tracking
- Project management
- Chat
Questions we would ask you
- Where do budgets live, and are they per project or per phase?
- Which time tracker do you use, and how many people log time daily?
- How do you measure percent complete today?
- What overrun would you want to hear about, and how early?
- Who should receive alerts: project leads, the owner, or both?
Discuss this further with the Live Agent
Operations & inventory
Client onboarding set up in one pass
When a contract is signed, the folders, project, channels, access requests, billing record and kickoff agenda are all created from the proposal.
Starts when
A proposal is signed.
What happens
- Create the client workspace from your standard structure
- Build the project plan from the proposal's phases
- Send the client an access and information request list
- Draft the kickoff agenda and propose meeting times
In more detail
An automation like this starts when a proposal or statement of work is signed. It reads the signed document and prepares the routine setup: client folders from the firm's standard structure, a project plan built from the proposal's phases, a list of what the firm needs from the client such as logins, files and contacts, a draft kickoff agenda with suggested times, and the details for the billing record. Everything is prepared, not sent. The account lead checks the plan, the request list and the billing details before anything is created or the client is invited, and the lead sends client emails personally. It needs a standard folder structure and project template, and it is most useful for firms that start new clients or projects often and whose contracts differ enough that someone has to read each one.
You stay in control
The account lead checks the plan before the client is invited.
Connects to
- E-signature
- Document storage
- Project management
- Accounting
Questions we would ask you
- How do clients sign today (PandaDoc, DocuSign, Proposify, PDF by email)?
- Do you have a standard folder structure and project template per service?
- What do you need from every new client in the first week, and how do you collect it now?
- Who creates the client in QuickBooks or Xero, and when?
- How many new clients or projects do you start per month?
Discuss this further with the Live Agent
Reporting
Monthly client reports assembled and explained
Data from each platform is pulled into your report template with a written explanation of what changed, why it matters and what you recommend next.
Starts when
The first working day of the month.
What happens
- Pull the figures from each analytics and advertising account
- Fill the client's report template and charts
- Explain the significant changes and their likely causes
- Draft recommendations tied to the client's stated goals
In more detail
An automation like this runs at the start of each month. It reads each client's figures from analytics and advertising accounts such as Google Analytics, Google Ads and Meta, or from the agency's existing reporting tool, and fills the client's report template. Ordinary code calculates the changes from last month and last year. The AI then drafts a short explanation of what changed, the likely reasons based on the agency's own log of what it changed, and two or three recommendations tied to the client's goals. The strategist edits the commentary, checks any odd numbers and sends or presents the report. Connecting directly to Google Ads and Meta needs platform approvals in the agency's name, so a simpler version drafts the commentary from an export of the reporting tool the agency already uses.
You stay in control
The strategist edits the commentary and presents the report.
Connects to
- Analytics
- Ad accounts
- Report templates
- Email
Questions we would ask you
- Which platforms appear in your reports, and which reporting tool do you use now?
- How many clients get a monthly report, and how long does each take?
- Do you keep a log of what changed in each account each month?
- Are each client's goals and KPIs written down?
- Would you accept a version that reads an export from your current reporting tool instead of connecting to each platform?
Discuss this further with the Live Agent
Documents & compliance
RFP and security questionnaires answered from past responses
A 200-question form is filled with your best previously approved answers, each linked to its source, leaving only the new questions to write.
Starts when
An RFP or vendor questionnaire arrives.
What happens
- Split the document into individual questions
- Find the closest approved answer for each one
- Adapt the wording to the question and cite where it came from
- List the questions with no good match for a subject expert
In more detail
An automation like this reads an incoming RFP or security questionnaire, usually a spreadsheet or Word file, and splits it into individual questions. For each question it finds the closest answer the firm has already approved, from past questionnaires and security policies, adapts the wording to the question and shows where the answer came from. Questions with no good match are not guessed; they go on a list for the right expert. The proposal owner reviews every answer before anything is submitted, and newly approved answers are added to the library for next time. It needs a collection of past approved answers and current policies. Fenrir Labs could design it to run on the firm's own machine, which matters because a security questionnaire describes exactly how the firm protects itself. Firms already paying for Vanta, Loopio or Responsive should use those first.
You stay in control
The proposal owner reviews every answer before submission.
Connects to
- Answer library
- Document storage
- Word processor
Questions we would ask you
- How many RFPs and security questionnaires do you answer per quarter, and how long does each take?
- Where do past answers live, and who approved them?
- Do you have a SOC 2 report, ISO certification, or CMMC assessment, or only policies?
- What formats arrive: spreadsheets, Word, PDF, or online portals?
- Who are the subject experts, and how quickly can they answer gaps?
Discuss this further with the Live Agent
Finance & billing
Scope creep caught while it is still a conversation
Client requests in email and chat are compared with the statement of work. Requests outside it are flagged with a drafted change request.
Starts when
A client sends a request.
What happens
- Compare the request with the signed scope and prior change requests
- Classify it as included, a clarification or new work
- Estimate the effort from similar past tasks
- Draft a friendly change request for new work
In more detail
An automation like this reads client requests as they arrive by email or chat and compares each one with the signed statement of work and any approved changes. It classifies the request as included, a clarification or new work, and quotes the part of the contract it relied on. For new work it estimates effort from similar past tasks in the time tracker and drafts a friendly change request with options, such as adding the work at a price, swapping it for something already in scope, or deferring it. The account lead decides what to do with every flagged request and sends anything to the client personally. It needs statements of work that describe scope and exclusions clearly. Because it reads client correspondence and contracts, Fenrir Labs could design it to run on the firm's own machine.
You stay in control
The account lead decides how to handle each flagged request.
Connects to
- Email
- Chat
- Contracts
- Project management
Questions we would ask you
- How are your SOWs written: deliverables and exclusions, or broad retainers?
- Where do client requests arrive: email, shared Slack channels, project tool comments, calls?
- How do you handle small extras today, and is there an allowance?
- Who decides whether to charge for new work?
- Do your time entries have notes good enough to estimate similar work?
Discuss this further with the Live Agent
Documents & compliance
Answers from your own past projects
Anyone on the team can ask how a similar problem was solved before and get an answer with links to the deliverables, proposals and lessons learned.
Starts when
A team member asks a question.
What happens
- Search past deliverables, proposals and project notes they are allowed to see
- Answer with the relevant passages and who worked on it
- Link to the source files
- Say plainly when nothing relevant exists
In more detail
An automation like this lets anyone in the firm ask a question in plain words, such as how a similar problem was solved before or which past projects are most like a new one, and get an answer from the firm's own deliverables, proposals and project notes. Each answer quotes the passages it used, names the project and the people involved, and links to the files. When nothing relevant exists it says so instead of guessing. It follows the folder permissions the firm already has, so people only get answers from files they could open anyway. That makes tidy permissions a prerequisite. Fenrir Labs could design it to run on the firm's own machine so past client work never leaves the office. Firms already paying for Microsoft 365 Copilot or Gemini should compare those first.
You stay in control
Access follows your existing folder permissions; answers always show their sources.
Connects to
- Document storage
- Project archives
- Chat
Questions we would ask you
- Where are past projects stored (Drive, SharePoint, file server, Deltek or BQE document storage), and how far back?
- Do folder permissions reflect who should see what today?
- Which questions do people ask most often that this would answer?
- Are any projects under NDA, CMMC or ITAR rules that require exclusion?
- Do you already pay for Microsoft 365 Copilot or Gemini for Workspace?
Discuss this further with the Live Agent
