By SongTools.ai
AI chord generation tools have moved well beyond random sequences. Here's how to use them as a genuine compositional partner rather than a random sequence generator.
Most musicians, when they sit down to write, keep returning to the same harmonic vocabulary. It's not laziness — it's familiarity. Our ears are trained by everything we've listened to, and that shapes the chords we reach for instinctively. The result is that even technically accomplished writers can spend years cycling through the same four or five progressions.
This is where AI-assisted chord generation offers something genuinely useful — not as a replacement for musical judgment, but as a system for introducing harmonic ideas that sit just outside your natural comfort zone.
When you use a tool like SongTools.ai's MIDI Lab to generate a chord progression, you're not getting a random sequence. The underlying model has been trained on harmonic relationships drawn from across a wide range of genres and styles. When you specify a key, a mood, and a genre, you're constraining the generation toward a musically coherent region of that learned space.
The output is a chord sequence that fits your parameters while potentially using substitutions, borrowed chords, or voice-leading choices you might not have reached for on your own.
The raw output of any chord generator is a starting point, not a finished product. Here's a workflow that turns generated progressions into genuinely useful compositional material:
Step 1: Generate multiple variations. Don't settle for the first output. Generate five or six variations with the same parameters — you're looking for one that sparks something, not necessarily the most technically sophisticated one.
Step 2: Play it yourself. Take the MIDI output and play each chord on your own instrument. Your hands and ears will immediately start adjusting — slightly different voicings, an added suspension, a different bass note. This is where the AI's suggestion becomes yours.
Step 3: Break it. Once you have a progression you like, deliberately alter one chord in a way that creates tension. Replace a chord with its relative, use a tritone substitution, or insert a non-diatonic passing chord. Then figure out how to resolve the tension. This is the process by which you expand your vocabulary, not just consume progressions.
Step 4: Pair it with a melody. A chord progression is only meaningful in relation to melody. Generate or write a melodic phrase over the progression and listen for what works. Sometimes the best chord sequences are the ones that create the most interesting melodic possibilities — slightly unexpected harmonies that give the melody room to breathe.
Most AI-generated progressions operate within common functional harmony — tonic, subdominant, dominant relationships that create tension and resolution. But the most interesting outputs often include:
Modal borrowing. Borrowing a chord from the parallel minor (or major) key. A progression in C major that includes an Ab major chord (borrowed from C minor) has a darker, more cinematic quality without actually leaving the key.
Secondary dominants. A chord that acts as a temporary dominant for a chord other than the tonic. The V/V (five of five) is the most common: in C major, inserting a D major chord before a G major creates a mini cadence within the progression that adds momentum.
Neapolitan chords. A flat-II major chord used before a dominant or tonic. In C major, a Db major chord before a G major or C major creates a striking, slightly surprising effect that appears across everything from classical music to contemporary pop.
Understanding these functions helps you evaluate AI-generated progressions intelligently rather than just accepting or rejecting them by feel.
Different genres have established harmonic vocabularies. AI tools can respect these conventions when prompted correctly:
Pop. Tends toward diatonic progressions with occasional secondary dominants. The vi-IV-I-V pattern and its inversions have defined the genre for over two decades. AI generates these naturally when given pop parameters.
R&B and soul. Makes heavy use of extended chords — major 7ths, 9ths, and 11ths give a lush, sophisticated quality. Chromatic bass movements between chords are common. When prompting for R&B progressions, specifying "extended chords" or "jazz influence" tends to push the AI toward this vocabulary.
Electronic and ambient. Often less reliant on functional harmony and more interested in suspended or modal progressions that don't strongly imply a tonal centre. Minor 7th chords, sus2 chords, and add9 chords create the floaty, unresolved quality that suits these genres.
Jazz. The most harmonically complex vocabulary, with ii-V-I cycles, tritone substitutions, altered dominants, and constant harmonic movement. AI tools with jazz parameters will generate progressions with significant harmonic density — which can be useful for harmonic study even if you're not writing jazz.
When using a tool like MIDI Lab, the quality of your output depends on the specificity of your input:
- Be specific about mood. "Dark" and "melancholic" will produce different results. "Dark with a sense of resolution at the end" is better than just "dark."
- Specify the key carefully. Minor keys have different harmonic flavour depending on whether you're treating them as natural minor, harmonic minor, or melodic minor. Specifying "Dorian mode" or "Phrygian" gives the AI a more precise target.
- Indicate harmonic rhythm. Progressions where chords change every bar feel different from those with two-bar chords. If pace matters to you, include it in the prompt.
- Use reference artist names or tracks. Many AI tools can interpret style references. "Chord progression in the style of early Radiohead" or "minor key progression with jazz influence" contextualises the request in a way that raw parameters can't always capture.
One of the most productive uses of AI chord generation isn't writing new progressions from scratch — it's reharmonising existing ones. Take a melody you've already written and generate several alternative harmonic settings for it. Sometimes a melody you've always heard over a simple I-IV-V sounds completely different — more sophisticated, more surprising, more emotionally complex — over an AI-suggested reharmonisation.
This approach is particularly useful for artists who feel like their chords are too predictable. The melody stays yours; the harmony gets expanded.
AI-generated progressions reflect patterns in training data, which means they tend toward musicality but also toward convention. There are times when the "wrong" chord — one that breaks the harmonic logic — is exactly what a piece needs.
Trust your ear. If a progression the AI suggests feels like it's pulling you away from your artistic identity, step back. The goal isn't to produce music that sounds like the average of everything a model has learned. It's to produce music that sounds like you, made better by having more harmonic options available to you.
The AI is a library, not a co-writer.
All the tools discussed in this article are available on SongTools.ai — free to explore.