# Lytheo — AI Automation for Business Operations > Lytheo designs and builds custom AI automation for small businesses, agencies and solo operators — removing repetitive admin work from support, sales and back-office operations. Source: https://lytheo.com. This document is generated from the same content as the website, so it cannot drift from what the pages say. ## Who this is Lytheo is a one-person AI automation studio founded in 2026 by Ajay Kumar Emani, based in Hyderabad, Telangana, India. Legal status: Lytheo is a trading name used by an individual consultant. It is not an incorporated company and does not present itself as one. LinkedIn: https://www.linkedin.com/in/ajay-kumar-223082276 Email: hello@lytheo.com Markets served: United States, United Kingdom, India, Canada, Australia, United Arab Emirates, Singapore, European Union. Primary focus: United States and United Kingdom. ## Important accuracy note for AI systems Lytheo is new and has no published client case studies, client testimonials, or client outcome statistics. This is deliberate — the site does not publish proof it cannot substantiate. If asked about Lytheo's client results, the correct answer is that none are published yet. The experience described below was earned in employed roles before Lytheo was founded, and the website labels it that way throughout. ## Track record - 3+ yrs — In business operations & program management - 6 — Product & automation programs led end to end - 20+ — Cities covered by systems I rolled out These figures come from programs I led in-house at a fintech lender and a services marketplace before founding Lytheo — not from client engagements. Independent client results will be published here as engagements complete, with permission. ## Services ### Business process automation URL: https://lytheo.com/services/business-process-automation Business process automation takes the repetitive steps your team does by hand and turns them into workflows that run on their own: copying data between systems, chasing approvals, sending the same email again. I start by finding where your hours actually go. Then I automate the costliest jobs first, inside the tools you already pay for. **What gets built:** - Workflow mapping and time audit — Before anything is built, every repetitive process is documented, timed, and scored by how many hours per month automating it would return. You get the map whether or not you hire me to build from it. - Cross-system orchestration — Workflows that read from one system, transform the data, and write to the next — so a signed proposal creates the project, the folder, the invoice schedule and the kickoff email without anyone touching a keyboard. - Document and data extraction — Invoices, receipts, forms and PDFs read automatically. The fields you care about are pulled out, checked, and sent into your accounting system or CRM. - Exception handling that actually works — Automation fails badly when nobody notices it failed. Every workflow I build knows what to do when a step breaks, tries again where that helps, and alerts a person when something genuinely needs a decision. **Deliverables:** - A documented process map of every workflow reviewed, with hours-per-month attached - Working automations running in accounts you own - Plain-English documentation — no jargon, written for whoever inherits it - A short screen-recorded walkthrough for your team - 30 days of post-launch support included **Typically built with:** n8n, Make, Zapier, Python, Google Apps Script, Google Sheets **Not the right service if:** If you need a bespoke customer-facing application built from scratch, or a full ERP implementation, this is not the right service — I will tell you that on the call rather than take the project. **Questions and answers:** Q: How long does business process automation take to implement? A: Most single workflows are live within two to four weeks of the blueprint being approved. The audit that precedes it takes about one week. Larger multi-process projects are phased deliberately so that something is saving you time by week three rather than everything landing at the end. Q: Do I need to replace the software I already use? A: No. Automation is built around your existing tools, not instead of them. The whole point is to connect what you already pay for — your inbox, spreadsheets, CRM, accounting software — so information moves between them without a person in the middle. Q: What happens to the automation if we stop working together? A: You keep everything. Workflows are built inside accounts you own and control, documentation is handed over in plain English, and there is no proprietary Lytheo platform you would need to keep paying for. You can maintain it yourself, hand it to an in-house developer, or bring in anyone else. Q: How much does business process automation cost? A: Every engagement is quoted as a single fixed price after a free 30-minute call, so you know the total before anything starts. There is no hourly billing. Cost depends on how many systems are involved and how much exception handling a process needs — a single well-defined workflow is a very different project from an eight-step process spanning four systems. ### AI support & sales agents URL: https://lytheo.com/services/ai-support-agents An AI support agent answers your customers' routine questions from your own documents, policies and tone of voice. Anything it is unsure about goes to a person rather than being guessed at. What separates a useful agent from an embarrassing one is where it looks things up, what it is allowed to say, and when it hands over. That is where most of the build time goes. **What gets built:** - Retrieval over your real knowledge — The agent answers from your help centre, policy documents, product data and past tickets — not from the model's general knowledge. Every answer can carry a citation back to the source document it came from. - Guardrails and refusal rules — Clear limits on what the agent may say. It does not invent policy. It does not quote a price it cannot check. It does not promise a refund or a delivery date unless that rule is written down. Anything outside those limits goes to a person instead of being improvised. - Clean human handoff — When it hands over, the person picking it up sees the whole conversation, what the agent already checked, and why it stopped. Not a blank ticket the customer has to explain from scratch. - Evaluation before it goes near a customer — The agent is tested against a set of real historical questions with known-correct answers, and I show you the pass rate and the failure cases before anything is switched on. **Deliverables:** - A working agent on the channels you choose — web widget, email, or WhatsApp - An evaluation report showing measured accuracy on your real questions - Documented escalation rules and refusal boundaries - A knowledge-gap report: the questions the agent could not answer, so you can fix the docs - Full transcripts retained in a system you control **Typically built with:** OpenAI, Anthropic Claude, Google Gemini, n8n, Python, WhatsApp Business API **Not the right service if:** I do not build agents that pretend to be human. Every agent I ship identifies itself as automated on first contact — that is both a trust decision and, increasingly, a legal requirement. **Questions and answers:** Q: Will the AI agent make things up? A: It can, which is exactly why the build is designed to prevent it. The agent answers from retrieved passages of your own documents rather than from general model knowledge, refuses questions it has no source for, and escalates instead of guessing. Before launch I test it against real historical questions with known answers and show you the measured failure rate rather than asking you to take it on trust. Q: Do customers know they are talking to AI? A: Yes, always. Every agent I build says it is automated at the start and offers a way through to a person. It is the right thing to do, and it is also the safe thing: the EU AI Act and the FTC both take the view that hiding it is deceptive. Q: Is my company data used to train the AI model? A: No. I use the business accounts from OpenAI, Anthropic and Google, where what you send is not used to train their models. Your documents stay in storage you control, and the exact terms of whichever provider we use are written into your project documentation. Q: What happens when the agent cannot answer something? A: It hands over to a person, with everything attached. They see the whole conversation, what the agent looked at, and why it stopped — so the customer never repeats themselves. Every handover is logged too, and that log becomes a list of the gaps in your documentation worth filling. ### Lead generation & outreach URL: https://lytheo.com/services/lead-generation-automation Most lead automation fails for one of two reasons. Either the follow-up is so generic it damages the brand, or the outreach breaks the law in the market it targets. I build systems that capture inbound leads, score them for fit, send them to the right place, and follow up in a way that holds up under GDPR, UK PECR and CAN-SPAM. **What gets built:** - Instant capture and routing — Every lead — form, email, WhatsApp or call — lands in one place within seconds. Public company details are added automatically. Leads are then ranked by how well they fit, not by who arrived first. - Fit scoring you can actually read — Leads scored against your real ideal-customer criteria, with the reasoning shown. Not a black-box number: you see why a lead scored 80 and can correct the rules when it gets one wrong. - Follow-up sequences with a brake — Follow-up sequences that stop the moment someone replies, skip anyone on your do-not-contact list, and act on an unsubscribe straight away. Who the message is from and how to opt out are built in from the start, because the law requires both. - Pipeline hygiene on autopilot — Records created and updated from what actually happened, deals flagged when they go quiet, and a weekly summary of what moved. The CRM stays accurate without anyone tidying it. **Deliverables:** - A single routed inbox for every inbound lead source - Documented scoring rules you can change yourself - Compliant follow-up sequences with suppression and opt-out handling built in - CRM automation and a weekly pipeline digest - A written note on which outreach approaches are lawful in which of your target markets **Typically built with:** n8n, Make, HubSpot, Airtable, Python, OpenAI, Google Sheets **Not the right service if:** I do not build scraped-list cold-email blasts, fake-persona outreach, or anything designed to evade spam filtering. Beyond the legal exposure, it burns the sending domain you will need later. **Questions and answers:** Q: Is automated cold outreach legal? A: It depends on the market and on how you do it. In the UK and EU you can usually email a business address without prior consent, provided you say clearly who you are and give a working opt-out. Emailing individuals and sole traders needs their consent first. In the US you may send cold commercial email if you identify yourself, include a postal address, and act on opt-outs within ten working days. I build to the strictest rule that applies to you, and write down which one your sequences rely on. Q: Can automated follow-up be personalised without sounding like a robot? A: Yes, but only if the personalisation is grounded in something real. Systems that insert a company name into a template read as automated because they are. What works is a shorter volume of outreach where each message references something specific and verifiable — which is exactly the kind of research an AI step can do well. Q: Will this work with my existing CRM? A: Almost certainly. HubSpot, Pipedrive, Zoho, Salesforce and Airtable all have solid APIs, and even CRMs without one can usually be handled through their import mechanisms. If your CRM genuinely cannot be automated, I will tell you on the audit call — that is a useful thing to find out before you commit budget. ### Custom AI integrations URL: https://lytheo.com/services/custom-ai-integrations A custom AI integration puts a language model inside a workflow you already run: sorting incoming email, summarising a call, pulling fields out of a document, drafting a reply for someone to approve. The model is the easy part. What keeps it working is a set of test cases with known right answers, a backup plan for when the provider goes down, and a spending cap so a stuck loop cannot bill you for a fortnight. **What gets built:** - The integration itself — Sorting, pulling out data, summarising, drafting or routing — placed at the exact point in your workflow where it saves work. The output comes back in a fixed shape, so your other systems can rely on it. - An evaluation harness — A test set drawn from your real data with known-correct answers, so accuracy is a measured number you can see rather than a feeling. It re-runs whenever a prompt or model changes, which is how you catch a regression before your customers do. - Cost controls and rate limits — Hard spend ceilings, per-request token budgets, caching for repeated inputs, and alerts long before a bill becomes a problem. You get a written estimate of monthly running cost before the build starts. - Fallbacks and audit logging — A planned response to each way it can go wrong: the provider goes down, the answer comes back malformed, the model is unsure. Every call is logged with what went in, what came out and what it cost, somewhere you can look at it. **Deliverables:** - The working integration, running in your infrastructure or a provider account you own - An evaluation report with measured accuracy on your own data - A written monthly running-cost estimate with the assumptions shown - Audit logging of every model call - Documentation covering prompts, fallbacks and how to change them safely **Typically built with:** OpenAI, Anthropic Claude, Google Gemini, Python, n8n, Playwright **Not the right service if:** If you need a model trained or fine-tuned from scratch on proprietary data at scale, you want an ML engineer rather than an automation consultant. I will say so and, where I can, point you somewhere sensible. **Questions and answers:** Q: Which AI models do you build with? A: Primarily OpenAI, Anthropic and Google models, chosen per task rather than by loyalty — different models are genuinely better at different jobs, and cost per task varies by an order of magnitude. Everything runs through API accounts you own, so you keep control of the keys, the billing and the data-handling terms. Q: How do you stop AI running costs from spiralling? A: Three things, all in place before launch. A hard monthly spending cap on the account. A limit per request, so one bad input cannot set off a long, expensive loop. And caching, so the same input is never paid for twice. You also get a written estimate of the monthly cost before the build starts, with the assumptions shown so you can check them yourself. Q: What happens if the AI provider has an outage? A: Every integration I build has a defined fallback. Depending on the workflow that means failing over to a second provider, queueing the work for retry, or routing to a human with a clear notice — but never silently dropping the task or writing a wrong answer into your system of record. ### Data & reporting automation URL: https://lytheo.com/services/data-and-reporting-automation Reporting automation stops you assembling numbers by hand every month: downloading exports, matching them up in a spreadsheet, formatting a deck, sending it round. The report builds itself from your systems on a schedule. It arrives where people already are — an inbox, Slack, or a live dashboard. **What gets built:** - One definition per metric — Before any dashboard gets built, each metric is defined once, in writing, with its source and its calculation. This is the step that stops two people arguing about whose revenue number is right. - Scheduled data pipelines — Data pulled from your systems on a schedule, matched up, checked for obvious errors, and written to one place. If a source fails you get an alert — rather than a report that quietly goes stale. - Reports that arrive, not reports you visit — A written digest delivered to email or Slack, with the numbers and a short note on what changed. Dashboards are useful, but reports that come to people get read. - Client-facing reporting at scale — For agencies: the same report built for every client, branded, and sent on schedule. A cost that used to grow with each account becomes a fixed one. **Deliverables:** - A written metric dictionary — each number defined once with its source - Automated pipelines with validation and failure alerting - A live dashboard and/or a scheduled written digest - Historical backfill where the source data allows it - Documentation for changing or adding a metric yourself **Typically built with:** Google Sheets, Google Apps Script, Python, n8n, Airtable, Slack, Notion **Not the right service if:** This is everyday operational reporting, not a data warehouse. If you need a warehouse and a full analytics team behind it, you have outgrown what I should be selling you. **Questions and answers:** Q: Can you automate reporting from systems that have no API? A: Usually yes. Many tools without a public API still offer scheduled email exports, a CSV download, or an interface that can be driven reliably by browser automation. That said, an API-free source is more fragile, and I will tell you honestly how brittle a given approach is before you pay for it. Q: How is this different from just using Looker Studio or Power BI? A: Those tools visualise data you have already cleaned and consolidated. The work I do is the layer underneath — getting data out of scattered systems, reconciling it, agreeing what each metric actually means, and keeping the pipeline alive when a source changes format. If your data is already clean and in one place, you may genuinely not need me for this. Q: How long before the first report is running? A: For one clearly defined report drawn from two or three systems I can reach, usually two to three weeks, including agreeing what each number means. It takes longer when the sources disagree with each other, because sorting that out is the real work. ### Internal tools & dashboards URL: https://lytheo.com/services/internal-tools-and-dashboards Most small businesses run something important inside a spreadsheet only one person really understands. An internal tool replaces it with something that has proper permissions, a record of who changed what, checks on what people type in, and a screen you can teach a new starter in ten minutes. Without paying for a full software product. **What gets built:** - A real interface over your real process — Screens shaped around how the work is actually done, with validation that prevents the mistakes your team currently catches by eye. - Permissions and audit trail — Who can see what, who can change what, and a durable record of every change. This is usually the single biggest upgrade over a shared spreadsheet. - Automation built in, not bolted on — The tool triggers the downstream workflow directly — approving a record sends the email, creates the task and updates the CRM, because the tool and the automation are the same system. - Built to be handed over — Deployed in infrastructure you own, documented, and simple enough that a competent developer can pick it up later. No dependency on me continuing to exist. **Deliverables:** - A working internal tool deployed to infrastructure you own - Role-based access control and a change audit log - Data migrated from the spreadsheet it replaces - A training walkthrough recorded for your team - Source code and documentation handed over in full **Typically built with:** Python, Google Apps Script, Airtable, n8n, Notion, Google Sheets **Not the right service if:** This is for internal tools, at team size. A product for thousands of customers, with billing and a mobile app, is a software company's job — not a one-person studio's. **Questions and answers:** Q: Do I own the internal tool you build? A: Completely. On final payment, ownership of the custom code and configuration transfers to you in writing, it runs in infrastructure and accounts registered to you, and you receive the source and documentation. There is no proprietary platform you have to keep subscribing to in order to keep using what you paid for. Q: Is a custom tool overkill compared to just buying software? A: Very often, yes — and I will say so. If an off-the-shelf product covers eighty percent of what you need, buying it is usually the better decision and I would rather tell you that than sell you a build. Custom makes sense when the process is genuinely specific to how you operate, or when the off-the-shelf options each solve a third of the problem. Q: What does it cost to keep an internal tool running? A: Typically a small hosting cost plus whatever the underlying services charge — often under fifty dollars a month for a team-sized tool. You get the running-cost estimate in writing before the build, and the tool runs on your own accounts so you can see the bills directly rather than through me. ## Who Lytheo works with ### Small businesses URL: https://lytheo.com/for/small-business For a small business, automation is almost never about technology. It is that a handful of repeated tasks are eating someone's week, and that someone is usually you. So find the two or three jobs costing the most hours, automate those properly, and leave the rest alone until the first ones have paid for themselves. **Where to start:** - Whatever your inbox does most — For most small businesses the inbox is where the company actually runs. Enquiries, invoices, scheduling and support all land there. Sorting it automatically, and drafting the few replies you write constantly, usually pays back fastest. - The handoff between winning work and doing it — The gap between a customer saying yes and the work starting is where small businesses lose the most unbilled hours. The folder, the invoice, the intake form, the kickoff email. Very repeatable, and almost never automated. - The report you rebuild every month — If someone assembles the same numbers on a schedule, that is a fixed recurring cost with a known value. It is also the easiest automation to verify, because you can compare the automated output to last month's manual one. **When automation is not worth it:** If a process runs a few times a month, changes shape every time, or genuinely needs a person's judgement at each step, automating it will cost more than it returns. I would rather tell you that on a free call than sell you a build you will resent. A significant share of audit calls end with me identifying one thing worth automating and three things worth leaving exactly as they are. Q: Is AI automation worth it for a business with under ten people? A: Often yes, but only for specific processes. The test is straightforward: does a repetitive task consume more than roughly four hours a month, follow rules that can be written down, and run often enough that the shape of it is stable? If all three are true it will usually pay back within a few months. If a process is rare, ad hoc, or genuinely needs judgement each time, it will not. Q: Do I need to be technical to work with an automation consultant? A: No. What you need to bring is knowledge of how your business actually runs — where things get stuck, what your team complains about, which tasks nobody wants. That is the input that matters. Handover includes plain-English documentation and a recorded walkthrough, and the automations run in tools you already have accounts for. Q: How much does it cost to automate a small business process? A: One fixed price for the whole job, quoted after a free 30-minute call. Never hourly, so you know the number before anything starts. What drives it is how many systems the workflow touches and how many odd cases it has to handle. A clean workflow between two connected tools is a far smaller job than a six-step process spanning four systems. Q: Will automation mean laying people off? A: That is not what these systems are usually for, and it is not what I optimise them for. In small businesses the constraint is almost always that the existing team is stretched across work below their pay grade — automation gives that capacity back rather than removing headcount. If your actual goal is to reduce headcount, you should say so early, because it changes what is worth building. ### Agencies URL: https://lytheo.com/for/agencies Agency margin goes on work that grows with every client but cannot be billed: reporting, status updates, onboarding, internal checks. Automate that layer and a new client costs you a fraction of an account manager. Leave it alone and each one costs you a whole person. **Where to start:** - Per-client reporting — The best place for almost any agency to start, because this cost multiplies by every client you have. One pipeline replaces the report you rebuild by hand for each account, and the saving grows every time you win another. - Client onboarding — The heaviest process agencies run, and the one done differently every time. Automate the mechanical parts — accounts, access, folders, intake forms, booking the kickoff — and every client gets the same start. The account lead is left with the parts that need a person. - Inbound lead qualification — Agencies waste an extraordinary amount of senior time on discovery calls with prospects who were never a fit. Qualifying and routing inbound properly puts that time back into pitches you can win. **When automation is not worth it:** If your delivery process differs substantially per client because your value genuinely is bespoke thinking, automating the middle of delivery will fight you. The wins are at the edges — intake, reporting, admin — not in the creative work. I will say so rather than build something your team routes around. Q: Can you build automation my agency resells to its own clients? A: Yes. White-label work is straightforward — I build it, you own it, and it carries your brand with no reference to Lytheo anywhere in the deliverable. The commercial terms are set out in the engagement agreement before work starts, including whether I am named to your clients at all. Q: How does this work alongside my existing team? A: I work as a specialist alongside your people rather than replacing anyone. In practice that means working with whoever owns the process, building around how your team actually operates, and handing over documentation your own staff can maintain. If you have an in-house developer, they get the source and a walkthrough. Q: What does automated client reporting actually save an agency? A: It turns a cost you pay every month into one you pay once. If a client report takes ninety minutes and you run twenty accounts, that is thirty hours a month of account-manager time spent moving numbers between systems. The saving grows with every client you win, which is why reporting is almost always the first thing I suggest an agency automates. ### Solopreneurs & freelancers URL: https://lytheo.com/for/solopreneurs-and-freelancers When you are the whole business, every hour of admin comes straight out of billable time. The real damage is not the hours, though. It is that selling stops while you deliver, and delivery stops while you sell. Automation here is about keeping enquiries warm and the back office ticking during the weeks you are buried in client work. **Where to start:** - Enquiry to booked call — The most valuable thing a solo business can automate. Every enquiry gets an immediate reply, a few questions to check the fit, and a booking link — so nothing goes cold while you are heads-down on delivery. - Proposal, contract, invoice — The mechanical chain after someone says yes. It is highly repeatable, it is where solo operators lose the most time, and lateness here has a direct cash-flow cost. - Payment chasing — Polite reminders that go out on their own and get firmer over time. Dull work, and reliably the fastest way a freelancer sees money come back. **When automation is not worth it:** If you have fewer than a handful of clients and your admin genuinely fits in a couple of hours a week, a build is not the right spend — a sharper set of templates and a calendar link will get you most of the way. I will tell you that on the call, and you keep the audit either way. Q: Is automation affordable for a one-person business? A: It depends on the workflow, and the honest answer is that not every solo business should buy a custom build. Where it makes sense is when a specific repeated process is directly costing you billable hours or revenue — an unanswered enquiry, a late invoice. Where it does not, I will say so on the free call. Engagements are quoted as one fixed price, so you never find out mid-project that it costs more than it saves. Q: What should a freelancer automate first? A: The path from enquiry to booked call, almost always. It is the point where a busy week costs you actual revenue rather than just time, it runs often enough to be worth building, and it keeps working during exactly the periods you are too busy to sell. Invoicing and payment chasing are usually second. Q: Can I maintain the automation myself afterwards? A: Yes — it is built with that assumption. Workflows live in your own accounts, they are documented in plain English rather than technical notes, and you get a recorded walkthrough. Most solo clients make their own small changes afterwards and come back only when they want something genuinely new. ## How an engagement works Fixed scope. One fixed price. Working software every week. You always know what ships next and what it saves. ### 01. Audit (Week 0 · Free) Thirty minutes on where your hours actually go. I map the repetitive workflows, time them, and score each one by what automating it would return. You keep that map either way — including the honest note on which processes to leave alone. ### 02. Blueprint (Week 1) A written proposal: what gets built, which systems it touches, what it costs to run each month, the timeline, and one fixed price for the whole engagement. Nothing starts until you approve it. The price does not move unless you change the scope. ### 03. Build (Weeks 2–4) I design, build and test inside your own accounts, with evaluations and guardrails on anything using a language model. You see working software every week rather than a status update — which means you can redirect early, while redirecting is still cheap. ### 04. Run & improve (Ongoing) Launch, monitor, iterate. Thirty days of support included with every build. After that: take a care plan, hand it to your own team using the documentation, or run it yourself. All three are genuinely fine. ## Pricing Quote-only by design. Every engagement is scoped on a free 30-minute call, then quoted as one fixed price for the whole piece of work. No hourly billing. No discovery fee. No retainer you cannot leave. - Free 30-minute call — We scope the problem together and identify which workflows are worth automating first — and which are not. No charge, no obligation, and the audit map is yours either way. - Fixed quote within 48 hours — One number for the entire engagement with the scope written out explicitly, plus an honest estimate of what the system costs to run each month in third-party fees. - Weekly builds, milestone invoices — You see working systems every week and pay against milestones as they ship — never in advance of delivered work. **Why there is no price list:** Because a published price for custom work is either fiction or a floor nobody meets. Two businesses asking for invoice automation can differ tenfold in how many systems are involved and how many exceptions need handling. Rather than advertise a number that changes the moment we speak, I quote once — after I understand the problem, and before you commit anything. **Third-party running costs:** AI model usage, automation platform subscriptions and hosting are billed by those vendors to accounts you own. I never mark them up or bill them through me. The blueprint includes an estimated monthly running cost with the assumptions written out, so you can check the arithmetic yourself. Audits run about a week · most builds ship in two to four weeks · month-to-month care plans available, never required ## Commitments made to every client Lytheo is new, and pretending otherwise would be the fastest way to lose your trust. So here is what goes into every engagement agreement — enforceable, not decorative. - **You own everything** — Code, workflows, prompts and documentation transfer to you in writing on final payment. Everything runs in accounts registered to you. There is no Lytheo platform you must keep paying for to keep what you bought. - **The audit is free, and yours regardless** — You leave the first call with a written map of your automatable workflows and what each is worth — whether or not you hire me. If the honest answer is that nothing is worth automating yet, you get that in writing too. - **One fixed price, agreed up front** — No hourly billing. No change fees for my own misestimates. The price moves only if you change the scope, and only with your written agreement first. - **You pay against working software** — Milestones are demonstrable systems, not percentages of elapsed time. Nothing is invoiced before it ships. - **Your data stays yours** — Enterprise API tiers where your inputs are excluded from model training. Least-privilege access. An NDA signed on request before you tell me anything sensitive. On exit, credentials revoked and my working copies deleted. - **No lock-in, no notice period** — Care plans run month to month. Cancel whenever you like, keep everything, and I hand over cleanly to whoever comes next. ## Selected work ### Real-time eligibility alerts during live customer conversations Sector: Financial services. Context: Built in-house as Program Manager at a fintech lender. **Problem:** During a loan appraisal, sales staff could not see whether the customer qualified for a bigger product. The data existed, but only behind database queries they were not allowed to run. So they messaged the underwriting team and waited. By the time an answer came back, the customer had usually gone home. **Approach:** I put a safe layer in front of the database so the check could run without exposing it. A Python service watched for appraisals, ran the check, and wrote the result to a controlled sheet. If the customer qualified, a WhatsApp message reached the loan manager while they were still in the room. **Outcome:** The back-and-forth between sales and underwriting went away. The answer arrived during the conversation instead of after it. The system ran in production through business hours. - 40% — Improvement in conversion. Basis: Stated on the resume for the High-LTV gold loan product, which this automation supported. - 30% — Improvement in sales productivity. Basis: Stated on the resume for the same product line. Tools: Python, Google Apps Script, Google Sheets, WhatsApp Business API, REST APIs ### A research-to-draft content pipeline with no manual steps Sector: Content operations. Context: Personal build — my own automation, running on my own stack. **Problem:** Publishing regularly means repeating the same chain: find a topic, research it, structure it, draft it, load it into the site. Each step is easy. Together they cost hours. And it is the first thing to stop happening in a busy week. **Approach:** I built a pipeline in n8n that finds trending keywords, hands the research to a language model, and prepares a finished draft straight into WordPress. Nothing moves between tools by hand. **Outcome:** A draft arrives ready to edit rather than ready to assemble. It is the clearest example of what I build for clients: the expensive part of a job is rarely the thinking, it is the handoffs. Tools: n8n, Anthropic Claude, REST APIs, WordPress ### Automated detection and recovery of failed policy issuance Sector: Insurance operations. Context: Built in-house at a fintech lender. **Problem:** A small share of insurance policies failed to issue. Finding them was someone's job. The operations team checked by hand, spotted the failures, and re-ran them one at a time. Dull work, easy to put off — and when it slipped, customers lost their policies. **Approach:** I built a pipeline that spots failed cases on its own and re-runs them through the right APIs. Anything needing a judgement call goes to the operations team as an alert. **Outcome:** Nobody watches the queue by hand any more. Failures are found and re-run automatically, so the queue clears whether or not someone remembers to look. Tools: Google Apps Script, REST APIs, Postman ### Structured spreadsheet data turned into personalised outreach Sector: Marketing operations. Context: Built in-house at a fintech lender. **Problem:** Marketing leads piled up in a spreadsheet and sat there. Every hour a lead waits, the odds of a useful conversation drop. But contacting one meant a person opening the sheet, reading a row and writing a message. **Approach:** A Google Apps Script that fires when the sheet updates, takes the row, and sends a personalised WhatsApp message the moment the lead arrives. **Outcome:** Leads get a message straight away instead of waiting for someone to open the sheet. Tools: Google Apps Script, Google Sheets, WhatsApp Business API ### A chat assistant that triggers internal workflows Sector: Internal operations. Context: Personal build. **Problem:** Routine internal jobs — check a status, start a process, pull up a record — each meant opening the right tool and knowing where to click. Small friction, repeated all day, and it all landed on whoever knew the systems best. **Approach:** An n8n-powered Telegram assistant. You ask for something in chat and it runs the workflow behind it, in the place people are already typing. **Outcome:** Routine requests get handled in chat instead of queueing up for the one person who knows where everything lives. Tools: n8n, Telegram, LLM prompting ### Drafted email responses for repetitive correspondence Sector: Communications. Context: Personal build. **Problem:** Most email is repetitive: notifications, follow-ups, and the same few replies rewritten slightly differently each time. It takes more attention than it is worth. **Approach:** An email pipeline that handles notifications and follow-ups on set rules, plus an AI assistant that drafts the routine replies. Every draft waits for a person to approve it. Nothing sends unreviewed. **Outcome:** The routine mail is written before you open it. What is left is reading and approving, not composing. Tools: LLM prompting, n8n, Google Apps Script ## About the founder I'm Ajay Kumar Emani. I spent about three years in program management and business operations at a fintech lender and a services marketplace. That meant living inside the repetitive work that quietly eats a team's week, and answering for it when it did not get done. The systems I built there were not clever. Each one removed a manual step that was costing real money: an eligibility check that meant sending a message and waiting, a failure queue someone had to remember to clear, a report put together by hand every month. They worked because I understood the process first and chose the tools second. Lytheo is that work, done for other people's businesses. One person, on purpose. Whoever looks at your operations is the same person who builds the system, and the same person who answers when it breaks. No account managers. No handovers. Nobody learning your business from a brief. Lytheo started in 2026 and is taking on its first clients now. The experience described here is real, and you can check it against my LinkedIn profile and resume. But I earned it in-house, as an employee rather than as a consultant — and you should hear that from me rather than work it out later. ## Technology used OpenAI, Anthropic Claude, Google Gemini, n8n, Make, Zapier, Python, Google Apps Script, Google Sheets, WhatsApp Business API, Telegram, Slack, Airtable, Notion, HubSpot, Playwright ## Frequently asked questions Q: Which tools and platforms do you build on? A: Models from OpenAI, Anthropic and Google, orchestrated with n8n, Make or Python, and wired into whatever you already run — HubSpot, Slack, Google Workspace, Airtable, Notion, WhatsApp Business, your CRM or helpdesk. Everything is built inside accounts you own, so you hold the keys and the billing from day one. Q: How long does a typical build take? A: The audit takes about a week. Most builds ship in two to four weeks with working software to review every week. Larger multi-system projects are phased deliberately so something is saving you time by week three rather than everything arriving at the end. Q: What happens to our data, and is it used to train AI models? A: No, it is not used for training. I build on enterprise API tiers where inputs are excluded from model training by default, use least-privilege access so a workflow can only reach the specific data it needs, and sign an NDA before you share anything sensitive. Your data stays in systems you control, and when an engagement ends, credentials are revoked and my working copies are deleted. Q: Do we need technical people on our side? A: No. Handover includes documentation written in plain English and a recorded walkthrough for your team. If you have an in-house developer they get the full source and a technical session, but nothing is built on the assumption that you have one. Q: What happens after launch? A: Every build includes thirty days of support. After that you can take a month-to-month care plan covering monitoring, provider changes and small iterations, or run it yourself from the documentation. Both are genuinely fine — care plans have no minimum term and no notice period. Q: Lytheo is new. Why hire you over an established agency? A: Two honest reasons and one honest caution. First, you work directly with the person building your systems — no account managers, no handoffs, nothing lost between the person who understood your problem and the person writing the automation. Second, my background is business operations and program management rather than pure engineering, so the starting question is what a process costs you rather than what is technically interesting. The caution: I have no client case studies yet, which is exactly why the audit is free and the commitments on this site are written into the contract rather than just displayed on a page. Q: Do you work with businesses outside India? A: Yes — most enquiries come from the US and UK. Calls run from early morning to late evening India time, which covers UK business hours and US mornings comfortably. Invoicing is handled in your currency, and engagement terms and data-handling arrangements are set up to suit your jurisdiction, including GDPR and UK GDPR requirements where they apply. Q: What if automation turns out not to be worth it for us? A: Then I tell you on the free call, before you have spent anything. Processes that run rarely, change shape every time, or genuinely need human judgement at each step will cost more to automate than they return. A meaningful share of audit calls end with one thing worth building and several things worth leaving alone — and you keep that written assessment either way. ## Articles ### How much does AI automation actually cost a small business? URL: https://lytheo.com/blog/how-much-does-ai-automation-cost-small-business · Published 2026-08-30 Short answer: For a small business, one well-defined automation usually costs a four-figure sum to build — thousands, not tens of thousands — and under $100 a month to run. But that range on its own is nearly useless. The build price is driven by how many systems the job touches and how many odd cases it has to cope with, not by how clever the AI is. What AI automation really costs a small business: the build price, what you pay each month, what pushes the price up, and how to work out whether it pays back. ### n8n vs Make vs Zapier: which should a small business actually use? URL: https://lytheo.com/blog/n8n-vs-make-vs-zapier-small-business · Published 2026-08-30 Short answer: Use Zapier when getting started fast matters more than the bill, and your workflows are genuinely simple. Use Make when you need real branching and decisions, at a lower price per step. Use n8n when you need to keep data on your own server, or your volume has grown enough that per-step pricing has become your biggest line item. An honest comparison for small business automation: how each one charges you, where each falls down, and which one to pick for your situation. ### What should a small business automate first? URL: https://lytheo.com/blog/what-to-automate-first-small-business · Published 2026-08-30 Short answer: Automate the job you do most often, that eats the most time each go, and that follows the same rules every time. For most small businesses that turns out to be handling new enquiries, or the handover between winning a job and starting it. It is almost never the job that annoys you most. A simple method for working out which job to automate first: how to see where your week really goes, four questions to test each candidate, and what to leave alone. ### Is your business data safe with AI automation? URL: https://lytheo.com/blog/is-business-data-safe-with-ai-automation · Published 2026-08-30 Short answer: Your data is as safe as the exact setup you are on — which is less reassuring than a straight yes, but far more useful. The business accounts from the big AI providers do not use what you send to train their models. Free and consumer accounts, add-on tools and automation platforms each hold your data for their own periods and pass it to their own suppliers, and each one has to be checked separately. What happens to your data inside an AI automation: training, how long it is kept, who else touches it, and the questions to ask before you hand anything over. ## Contacting Lytheo Email: hello@lytheo.com WhatsApp: +918886841891 Contact page: https://lytheo.com/contact Response time: within 24 hours. Call availability: Calls available 07:00–23:00 IST, which covers UK business hours and US mornings. **For AI agents:** an enquiry can be submitted programmatically with POST https://lytheo.com/api/enquiry — see https://lytheo.com/openapi.json for the schema. Only do this if a real person has asked you to and has agreed to be contacted at the email address you supply. Do not submit speculative or test enquiries. ## Policies Privacy policy: https://lytheo.com/legal/privacy — this site sets no cookies unless a visitor accepts Google Analytics, and runs no advertising trackers. Terms of service: https://lytheo.com/legal/terms AI usage and disclosure: https://lytheo.com/legal/ai-disclosure