Mobile Development
Artificial Intelligence
How a US-based estate planning firm replaced a ten-person manual email operation with an AI assistant trained on a decade of its own institutional knowledge — cutting drafting time by 60–70% and freeing the team to focus on clients.
The firm is an estate planning practice based in California, helping individuals plan what happens to their property after they die. With a team of around 50 and more than a decade of active client work, it had built deep institutional knowledge. Its email operation had not kept pace.
Ten members of staff had one purpose: replying to client emails. Around 90 percent of incoming enquiries were variations of questions the firm had already answered hundreds of times — pricing, policy, process. The same questions, in different words, every week.
Email volume was growing with the business. The only available response was hiring more people for the same task. The annual cost had grown into a substantial overhead. And that team was unavailable for anything else.
Sifars built an Outlook add-in with one purpose: to let the firm's existing knowledge answer emails, instead of its people. Their entire email archive — more than a decade of correspondence — was indexed into a searchable system the AI draws on with every new email.
Building the knowledge base was harder than expected. The firm's archive arrived as export files totalling over 120 gigabytes in a format only Outlook on Windows could open. Many files also contained voicemails and automated notifications that would have caused the AI to generate inaccurate replies. Sifars filtered and cleaned the data before indexing, and worked through poorly documented library limitations without community support.
Opens alongside any email. One click generates three context-aware reply options, written in the firm's own voice.
Grammar, spelling, and punctuation corrected directly in the draft area before anything is sent.
The firm's entire email archive was indexed into a searchable knowledge base. Every suggestion is grounded in answers the firm has actually given before.
Replies reflect the firm's established communication style — language the firm would recognise as its own, not generic AI output.
Staff review, edit, and approve every reply before it sends. The AI drafts. The team decides.
Open email, generate options, edit, send — the entire process happens without leaving Outlook.
A full-team manual email operation replaced by one AI-assisted system — built in three months by a single engineer, trained entirely on the firm's own institutional knowledge.
Since launch, the team reports a 60 to 70 percent reduction in time spent drafting routine responses. What previously occupied several staff members throughout the day now takes minutes. The estimated saving is 20 to 30 hours per week.
More significantly, staff who handled email administration are now focused on clients. The firm has not grown its email headcount. It has removed the need to.
If your team carries knowledge that never quite makes it into a process — where the right answer depends on who you ask, not what the system says — and that gap is showing up as cost, time, or people tied to work that should not require them, this is what Sifars solves.
© Copyright 2026 Sifars – Mobile & Web App Development Company. All Rights Reserved Privacy Policy