AI and agentic workflows
Put an assistant, checker or automated workflow to work on one named job. Each outcome says what the automation does, what a person approves, what it never does alone, and the test you can run yourself to see it work.
311 outcomes in this list.
Outcomes
- A managed back-office finance lane: monthly statement, inbox and payment-run checksFor one small business: monthly supplier statement checks, inbox routing tests, payment-run checks and a month-end draft, plus one bounded extra. Proposed GBP 787 a month, by written proposal.Quoted before we start
- A managed compliance evidence lane: evidence index, access reviews and policy checksFor one control area of one framework: evidence refreshed each quarter, access reviews recorded, policies checked against systems monthly, and one bounded item a month. Proposed GBP 695 a month.Quoted before we start
- A managed security assurance lane: answers, app reviews and a monthly sign-in reviewFor one company: your answer library kept current, a quarterly connected-app review, a monthly sign-in review and one bounded item a month. Proposed GBP 615 a month; timing by written proposal.Quoted before we start
- A managed support lane: your bot and help centre kept inside written limitsFor one help desk and its channels: a monthly bot check, a kept-current help centre and up to two bounded items a month. Proposed GBP 720 a month; timing and capacity by written proposal.Quoted before we start
- A monthly agent-assisted lane for our engineering team: review, backlog and CI careFor one repository: AI reviewer kept tuned, dependency-bot verdicts, flaky-test and backlog reports, agent-PR checks, plus up to two bounded items a month. Proposed £1,100 a month; quote first.Quoted before we start
- A written review policy and rules for what gets held or removed, tested on 60 reviewsA written review policy and a decision table for publish, hold or remove with a reason, tested on 60 labelled reviews. Genuine negative reviews stay, and a person decides every removal.£195
- Add alt text to your product images, checked on a sample you chooseAlt text for up to 200 product images, scored by your reviewer on a 50-image sample you choose, with no prices or claims in the text and text printed in images transcribed or flagged.£149
- Add company size and sector to each new lead, with a source for every valueFor new leads with a business email, named company fields are written with a source and date, or marked unknown. Free-mail addresses and personal data are left alone.£395
- Add data checks to one table and show they would have caught your last bad loadFor one table, up to 15 checks in your tool's syntax, each shown to fire on a planted bad copy and stay quiet on 90 days of good history. The model sees only the schema, counts and ranges.£295
- Add the paragraph of commentary to my monthly report, and prove every number in itYour existing report gains a short commentary where every figure is copied from a computed table and a script proves it. No cause is stated unless you supplied it.£149
- Answer customer questions from your own help pages, with a person taking overA chatbot for any website that answers only from up to 30 help pages you name, cites the page, says it does not know otherwise, and hands listed cases to a person.£295
- Answer customers' invoice and balance questions from your own ledger exportAn automated assistant answers each customer's invoice-status and balance questions from your ledger export, says it is automated, and hands everything else to your named person.£395
- Answer staff policy questions from your own handbook, with a named person taking overAn assistant limited to your current handbook that answers with the section it quotes, or says it is not covered and names who to ask, tested on 100 questions you write.£295
- Answer WhatsApp questions from your own help pages, with a person able to take overA WhatsApp assistant answers only from the help pages you name, says in its first message that it is automated, and hands over to a named person. Customers message first; it never sends first.£395
- API docs whose examples really work: every example run against your test instanceReference pages for up to 30 endpoints, built from your code or OpenAPI file, with every example executed against a test instance and the real response pasted. A drift script fails when they diverge.From £395
- Ask every customer for a review the right way, with no screeningAfter each completed job or order, every eligible customer gets one identical request and at most one reminder. Nobody is screened by satisfaction and no incentive is offered.£295
- Ask up to 20 questions about your codebase each month, with file and line receiptsEach month up to 20 questions about your repository are answered with file and line citations at a named commit, checked by script. "Not in the code" is a valid answer. This is not a chatbot.£145 / month
- Before a payment run, get a list of probable duplicates and changed bank detailsFor each payment batch you get an exceptions list: probable duplicates, changed bank details, first-time payees and unusual amounts, each with its evidence for your approver to hold or release.£195
- Before we run AI over a million rows, tell me what it will cost and cap itA sampled cost estimate with a range, a per-job cap and a usage query that reproduces the billed credits, checked against a 1,000-row trial that you run in your own account.£195
- Build an interview timetable with no clashes from everyone's stated free timesA timetable for one interview round built from the free times people send you as emails, notes or calendar exports, with every clash checked by a script and invite wording drafted for you to send.£99
- Build the evidence pack for one chargeback from your records, in the turnaround we quoteFor one dispute, an evidence pack built only from your records, in the order the reason code asks for, ready for you to submit. We do not submit and do not promise a win.£149
- Build the reporting tables I have specified, and prove them against my control totalsUp to three reporting tables, written as tested dbt models or SQL views from your written spec, each reconciling to control totals you sign off. Reviewed and merged by your team.From £695
- Check AI-written pull requests for the usual agent mistakes before a person reads themA check you own flags weakened or skipped tests, lowered coverage, disabled CI steps, copied helpers and new package names on pull requests you mark as agent-written. It approves nothing.£295
- Check each batch of expense claims against your written policy and get a list of flagsEach claim batch is checked against the limits, missing-receipt, duplicate and weekend rules in your written policy, and you get a flag list with the rule quoted. The approver stays a person.£149
- Check every help article against your current policy and product, with evidence for eachEvery article gets a status (current, outdated, contradicts another or broken link) with the source it conflicts with, against the sources you name. We change nothing.£195
- Check every month for leavers who still have access to your named appsEach month we compare your staff list with your directory and up to five apps and send a list of leavers who still have an active account, seat or token, with the evidence row for each.£95 / month
- Check every month that all your mail sources are set up and nothing odd sends as youEach month a written record shows each sending source authenticated, unknown senders listed, the spam rate against Google's limit and one-click unsubscribe present. Mostly rules and checks.£79 / month
- Check every new package a pull request adds, before anyone merges itA check lists each new or changed package in a pull request with its registry age, maintainers, install scripts, licence and whether it exists. Tested on names your maintainer plants.£195
- Check that your top 25 products show the same price, stock and return terms in four placesFor 25 products, price, stock status, variants, key specs and shipping and return terms are compared across your page, structured data, product feed and Shopify catalogue entry on a stated date.£245
- Check your AI-written product descriptions against your own data before they go liveFor up to 150 products, every number, spec and claim in AI-written copy is traced to your source data or flagged, and a clean import file holds only the fixes you approve.£245
- Check your supplier statements against your bills and get a list of what does not matchFor one month and up to 15 suppliers, every statement line is matched to a bill, credit note or payment in your ledger export, or listed as a difference with its page reference.£195
- Choose an AI pull-request reviewer by testing up to three on your own old PRsUp to three AI reviewers run on the same 30 of your old pull requests, with a ranked table, each tool's published price with the date read and a recommendation that lists what was not tested.£195
- Code your own win-loss interviews into reasons, with the quotes behind eachInterview notes you collected yourself are coded to reasons you approve, with the verbatim quote behind each code. We never contact your buyers.£249
- Collect each interviewer's notes under the same questions, with no scoringOne sheet per candidate placing each interviewer's own written notes under the agreed questions, with gaps shown and no scores, ranking or recommendation.£149
- Contain one account, app grant or mailbox, only after your admin approves that one actionAutomation proposes one containing action with evidence and an undo note, and carries it out only after your named admin approves that single action. Nothing runs without approval.£695
- Cut your help-desk tags to an agreed list, with a definition and two examples eachYour tag list cut to an agreed size, each tag with a one-line definition and two examples, an old-to-new mapping and a retire list, checked by two of your staff on 50 tickets.£149
- Define our key numbers once so everyone, and any AI tool, gets the same answerUp to ten metrics are defined as code on one schema, with a 30-question set that reconciles to figures you sign off. Five planted questions outside the definitions must be refused.From £495
- Document every custom CRM field and see which ones nobody usesYou get a data dictionary of every custom property, with fill rates and where each is used, plus a list of fields that look safe to retire. No changes made.£149
- Draft "where is my order" replies from order facts, with a check that stops stale onesEight templates (four replies, four hand-over notes), filled from named order and tracking fields, with a rule that sends any draft whose value differs from the live order to a person.£149
- Draft help articles from your solved IT tickets, each step traced to a ticketEight to ten help articles drafted from your solved tickets, each step traced to a ticket line, scanned for personal data and followed step by step by a person on your side.£249
Not listed?
Tell us what you need done, what it is for, and when. Please don’t send passwords, code or customer data.