ENSLBook a free audit

Services · AI & Automation

AI agents that touch your CRM, not a chatbot on your homepage.

Goodish builds AI agents and automated workflows, mostly in n8n with OpenAI, Claude and Gemini, that take busywork off marketing and sales teams: lead research, reporting, content and support. They run around the clock, and one rule holds: the agent writes to a review queue, and nothing reaches a customer or a record until a person approves it.

3D illustration of human-in-the-loop AI: an agent pod drafts cards onto a review belt, a person stamps each one approved, and approved cards drop into a CRM crate.
Free
Audit first; a fixed quote on scope after it
24/7
Workflows run around the clock, with a human approval step where it matters
2
AI products of our own in production: eKmet and SocialAI
Zoho
Zoho Authorized Partner, alongside Google and Meta partnerships

What can we automate first?

Lead qualification

Incoming leads scored and enriched by AI before they reach the sales team.

Typeform · OpenAI · Zoho

Sales research

Recent news and company details gathered before every call, so reps start with context.

Calendly · OpenAI · Zoho

Auto reporting

Data from ads, social and SEO tools turned into a meeting-ready report every Monday morning.

GA4 · Sheets · Slack

Knowledge-base assistant

Customer questions answered from your own documents, on your website or on WhatsApp.

WhatsApp · Gemini · Drive

Content repurposing

One video transcript turned into posts, a LinkedIn article and a blog draft, queued for review.

YouTube · Claude · WordPress

Performance alerts

Cross-platform anomaly alerts and PPC optimisation triggers, wired to your ads data.

What is AI automation for a marketing or sales team?

It is software that does the repetitive parts of the job: reading, sorting, researching, summarising, drafting and updating systems. A workflow starts from a trigger, uses an AI model where judgement on text is needed, and finishes by updating a record, sending a report or notifying a person.

We build these for growing companies, most of them American, and for ourselves first. The point is not a chatbot on your homepage. It is fewer hours spent copying data between tools.

Why does every workflow have a human approval step?

Because AI models are confident even when they are wrong. A wrong lead score or a made-up detail in a customer email costs more than the time the automation saved.

So the agent writes to a review queue, and the queue drains only on human approval. People keep the judgement calls; the agent does the preparation.

What have we built with AI ourselves?

Two products in production. eKmet is a voice-first app for Slovenian farmers: say what you did in the field, and eKmet builds the legal record, with the parcel, a dose check and the weather at the time of work, for the farmer to confirm. SocialAI is our social-media management software for agencies.

ATC Healthcare, one of our US clients, also runs social content automation we built. We run our own agency on the same kind of workflows.

Can AI help you get found, too?

Yes, and it is a separate job from automation. AI assistants like ChatGPT increasingly answer the questions your buyers used to type into Google. Being recommended there depends on pages they can crawl, understand and cite. We cover that on the ChatGPT Ads page, together with the new ad channel.

How do we build an AI workflow?

1. Trigger

A form submission, an email, a CRM update or a schedule starts the workflow the moment it happens.

2. AI processing

The chosen model reads the input, enriches it with retrieval over your own documents, and drafts a decision or an output.

3. Review queue

The result lands in a queue. The queue drains only on human approval, so the agent never writes to your CRM or a customer on its own.

4. Action and record

Approved results sync to the CRM or database, reports go out and the right people are notified.

How an engagement runs

  1. Audit & strategy

    We map the manual work, pick the processes where automation pays back fastest, and agree what a person must still approve.

  2. Build & implementation

    Workflows built and connected to your tools, tested on real data, with the review queue in place.

  3. Launch & hand-over

    Live, monitored, and documented, so your team knows what runs, when, and how to stop it.

How is it priced?

Scoped per workflow: the price depends mostly on the number of systems involved, and we quote after the free audit.

Book a free audit

From Tech Pulse

AI & automation

All articles on this topic

FAQ

Questions people ask

How does AI reduce marketing costs?

Mostly by taking repetitive work off people: research, reporting, first drafts, data entry and routing. It also speeds up analysis and makes finer audience segmentation practical. The saving is real when the work was manual and frequent; it is small when the process was already automated.

Will an AI agent change our CRM or email customers on its own?

Not in what we build. The agent writes to a review queue, and the queue drains only when a person approves. You can relax that for low-risk steps later, deliberately and in writing.

Why n8n?

n8n is a workflow automation tool that can be self-hosted, so the workflows and the data can stay on infrastructure you control. It also connects to most marketing and CRM tools and to the major AI models, which keeps us from locking you into one vendor.

Which AI models do you use?

Whichever fits the task: OpenAI, Claude and Gemini all appear in our client workflows. The choice depends on the job, the cost per run and where the data is allowed to go.

How much does an AI automation cost?

It depends mostly on the number of systems the workflow touches. We quote on scope after the free audit, so you see a fixed price before anything is built.

How do we get started?

Book the free audit: thirty minutes in which we look at the processes your team repeats every week and point out the two or three where automation pays back first.